void (*destroy_inode)(struct inode *);
void (*dirty_inode) (struct inode *);
int (*write_inode) (struct inode *, int);
- void (*put_inode) (struct inode *);
void (*drop_inode) (struct inode *);
void (*delete_inode) (struct inode *);
void (*put_super) (struct super_block *);
destroy_inode: no
dirty_inode: no (must not sleep)
write_inode: no
-put_inode: no
drop_inode: no !!!inode_lock!!!
delete_inode: no
put_super: yes yes no
void (*dirty_inode) (struct inode *);
int (*write_inode) (struct inode *, int);
- void (*put_inode) (struct inode *);
void (*drop_inode) (struct inode *);
void (*delete_inode) (struct inode *);
void (*put_super) (struct super_block *);
inode to disc. The second parameter indicates whether the write
should be synchronous or not, not all filesystems check this flag.
- put_inode: called when the VFS inode is removed from the inode
- cache.
-
drop_inode: called when the last access to the inode is dropped,
with the inode_lock spinlock held.
mac5380= [HW,SCSI] Format:
<can_queue>,<cmd_per_lun>,<sg_tablesize>,<hostid>,<use_tags>
- mac53c9x= [HW,SCSI] Format:
- <num_esps>,<disconnect>,<nosync>,<can_queue>,<cmd_per_lun>,<sg_tablesize>,<hostid>,<use_tags>
-
machvec= [IA64] Force the use of a particular machine-vector
(machvec) in a generic kernel.
Example: machvec=hpzx1_swiotlb
This is normally done in pci_enable_device(),
so this option is a temporary workaround
for broken drivers that don't call it.
+ skip_isa_align [X86] do not align io start addr, so can
+ handle more pci cards
firmware [ARM] Do not re-enumerate the bus but instead
just use the configuration from the
bootloader. This is currently used on
+++ /dev/null
- Goals, Design and Implementation of the
- new ultra-scalable O(1) scheduler
-
-
- This is an edited version of an email Ingo Molnar sent to
- lkml on 4 Jan 2002. It describes the goals, design, and
- implementation of Ingo's new ultra-scalable O(1) scheduler.
- Last Updated: 18 April 2002.
-
-
-Goal
-====
-
-The main goal of the new scheduler is to keep all the good things we know
-and love about the current Linux scheduler:
-
- - good interactive performance even during high load: if the user
- types or clicks then the system must react instantly and must execute
- the user tasks smoothly, even during considerable background load.
-
- - good scheduling/wakeup performance with 1-2 runnable processes.
-
- - fairness: no process should stay without any timeslice for any
- unreasonable amount of time. No process should get an unjustly high
- amount of CPU time.
-
- - priorities: less important tasks can be started with lower priority,
- more important tasks with higher priority.
-
- - SMP efficiency: no CPU should stay idle if there is work to do.
-
- - SMP affinity: processes which run on one CPU should stay affine to
- that CPU. Processes should not bounce between CPUs too frequently.
-
- - plus additional scheduler features: RT scheduling, CPU binding.
-
-and the goal is also to add a few new things:
-
- - fully O(1) scheduling. Are you tired of the recalculation loop
- blowing the L1 cache away every now and then? Do you think the goodness
- loop is taking a bit too long to finish if there are lots of runnable
- processes? This new scheduler takes no prisoners: wakeup(), schedule(),
- the timer interrupt are all O(1) algorithms. There is no recalculation
- loop. There is no goodness loop either.
-
- - 'perfect' SMP scalability. With the new scheduler there is no 'big'
- runqueue_lock anymore - it's all per-CPU runqueues and locks - two
- tasks on two separate CPUs can wake up, schedule and context-switch
- completely in parallel, without any interlocking. All
- scheduling-relevant data is structured for maximum scalability.
-
- - better SMP affinity. The old scheduler has a particular weakness that
- causes the random bouncing of tasks between CPUs if/when higher
- priority/interactive tasks, this was observed and reported by many
- people. The reason is that the timeslice recalculation loop first needs
- every currently running task to consume its timeslice. But when this
- happens on eg. an 8-way system, then this property starves an
- increasing number of CPUs from executing any process. Once the last
- task that has a timeslice left has finished using up that timeslice,
- the recalculation loop is triggered and other CPUs can start executing
- tasks again - after having idled around for a number of timer ticks.
- The more CPUs, the worse this effect.
-
- Furthermore, this same effect causes the bouncing effect as well:
- whenever there is such a 'timeslice squeeze' of the global runqueue,
- idle processors start executing tasks which are not affine to that CPU.
- (because the affine tasks have finished off their timeslices already.)
-
- The new scheduler solves this problem by distributing timeslices on a
- per-CPU basis, without having any global synchronization or
- recalculation.
-
- - batch scheduling. A significant proportion of computing-intensive tasks
- benefit from batch-scheduling, where timeslices are long and processes
- are roundrobin scheduled. The new scheduler does such batch-scheduling
- of the lowest priority tasks - so nice +19 jobs will get
- 'batch-scheduled' automatically. With this scheduler, nice +19 jobs are
- in essence SCHED_IDLE, from an interactiveness point of view.
-
- - handle extreme loads more smoothly, without breakdown and scheduling
- storms.
-
- - O(1) RT scheduling. For those RT folks who are paranoid about the
- O(nr_running) property of the goodness loop and the recalculation loop.
-
- - run fork()ed children before the parent. Andrea has pointed out the
- advantages of this a few months ago, but patches for this feature
- do not work with the old scheduler as well as they should,
- because idle processes often steal the new child before the fork()ing
- CPU gets to execute it.
-
-
-Design
-======
-
-The core of the new scheduler contains the following mechanisms:
-
- - *two* priority-ordered 'priority arrays' per CPU. There is an 'active'
- array and an 'expired' array. The active array contains all tasks that
- are affine to this CPU and have timeslices left. The expired array
- contains all tasks which have used up their timeslices - but this array
- is kept sorted as well. The active and expired array is not accessed
- directly, it's accessed through two pointers in the per-CPU runqueue
- structure. If all active tasks are used up then we 'switch' the two
- pointers and from now on the ready-to-go (former-) expired array is the
- active array - and the empty active array serves as the new collector
- for expired tasks.
-
- - there is a 64-bit bitmap cache for array indices. Finding the highest
- priority task is thus a matter of two x86 BSFL bit-search instructions.
-
-the split-array solution enables us to have an arbitrary number of active
-and expired tasks, and the recalculation of timeslices can be done
-immediately when the timeslice expires. Because the arrays are always
-access through the pointers in the runqueue, switching the two arrays can
-be done very quickly.
-
-this is a hybride priority-list approach coupled with roundrobin
-scheduling and the array-switch method of distributing timeslices.
-
- - there is a per-task 'load estimator'.
-
-one of the toughest things to get right is good interactive feel during
-heavy system load. While playing with various scheduler variants i found
-that the best interactive feel is achieved not by 'boosting' interactive
-tasks, but by 'punishing' tasks that want to use more CPU time than there
-is available. This method is also much easier to do in an O(1) fashion.
-
-to establish the actual 'load' the task contributes to the system, a
-complex-looking but pretty accurate method is used: there is a 4-entry
-'history' ringbuffer of the task's activities during the last 4 seconds.
-This ringbuffer is operated without much overhead. The entries tell the
-scheduler a pretty accurate load-history of the task: has it used up more
-CPU time or less during the past N seconds. [the size '4' and the interval
-of 4x 1 seconds was found by lots of experimentation - this part is
-flexible and can be changed in both directions.]
-
-the penalty a task gets for generating more load than the CPU can handle
-is a priority decrease - there is a maximum amount to this penalty
-relative to their static priority, so even fully CPU-bound tasks will
-observe each other's priorities, and will share the CPU accordingly.
-
-the SMP load-balancer can be extended/switched with additional parallel
-computing and cache hierarchy concepts: NUMA scheduling, multi-core CPUs
-can be supported easily by changing the load-balancer. Right now it's
-tuned for my SMP systems.
-
-i skipped the prev->mm == next->mm advantage - no workload i know of shows
-any sensitivity to this. It can be added back by sacrificing O(1)
-schedule() [the current and one-lower priority list can be searched for a
-that->mm == current->mm condition], but costs a fair number of cycles
-during a number of important workloads, so i wanted to avoid this as much
-as possible.
-
-- the SMP idle-task startup code was still racy and the new scheduler
-triggered this. So i streamlined the idle-setup code a bit. We do not call
-into schedule() before all processors have started up fully and all idle
-threads are in place.
-
-- the patch also cleans up a number of aspects of sched.c - moves code
-into other areas of the kernel where it's appropriate, and simplifies
-certain code paths and data constructs. As a result, the new scheduler's
-code is smaller than the old one.
-
- Ingo
S: Maintained
INTEL ETHERNET DRIVERS (e100/e1000/e1000e/igb/ixgb/ixgbe)
-P: Auke Kok
-M: auke-jan.h.kok@intel.com
-P: Jesse Brandeburg
-M: jesse.brandeburg@intel.com
P: Jeff Kirsher
M: jeffrey.t.kirsher@intel.com
+P: Jesse Brandeburg
+M: jesse.brandeburg@intel.com
P: Bruce Allan
M: bruce.w.allan@intel.com
P: John Ronciak
* (if do_page_fault didn't fix the mapping,
* the writeback won't do good)
*/
+disable_wb:
#ifdef DEBUG
printk(".. disabling wb2\n");
#endif
if (fp->un.fmt7.wb2a == fp->un.fmt7.faddr)
fp->un.fmt7.wb2s &= ~WBV_040;
+ if (fp->un.fmt7.wb3a == fp->un.fmt7.faddr)
+ fp->un.fmt7.wb3s &= ~WBV_040;
}
- } else if (send_fault_sig(&fp->ptregs) > 0) {
- printk("68040 access error, ssw=%x\n", ssw);
- trap_c(fp);
+ } else {
+ /* In case of a bus error we either kill the process or expect
+ * the kernel to catch the fault, which then is also responsible
+ * for cleaning up the mess.
+ */
+ current->thread.signo = SIGBUS;
+ current->thread.faddr = fp->un.fmt7.faddr;
+ if (send_fault_sig(&fp->ptregs) >= 0)
+ printk("68040 bus error (ssw=%x, faddr=%lx)\n", ssw,
+ fp->un.fmt7.faddr);
+ goto disable_wb;
}
do_040writebacks(fp);
struct mac_booter_data mac_bi_data;
int mac_bisize = sizeof mac_bi_data;
-struct mac_hw_present mac_hw_present;
-EXPORT_SYMBOL(mac_hw_present);
-
/* New m68k bootinfo stuff and videobase */
extern int m68k_num_memory;
m68k_ramdisk.addr, m68k_ramdisk.size);
#endif
- /*
- * TODO: set the various fields in macintosh_config->hw_present here!
- */
- switch (macintosh_config->scsi_type) {
- case MAC_SCSI_OLD:
- MACHW_SET(MAC_SCSI_80);
- break;
- case MAC_SCSI_QUADRA:
- case MAC_SCSI_QUADRA2:
- case MAC_SCSI_QUADRA3:
- MACHW_SET(MAC_SCSI_96);
- if ((macintosh_config->ident == MAC_MODEL_Q900) ||
- (macintosh_config->ident == MAC_MODEL_Q950))
- MACHW_SET(MAC_SCSI_96_2);
- break;
- default:
- printk(KERN_WARNING "config.c: wtf: unknown scsi, using 53c80\n");
- MACHW_SET(MAC_SCSI_80);
- break;
- }
-
iop_init();
via_init();
oss_init();
endif
obj-$(CONFIG_PPC_LIB_RHEAP) += rheap.o
+obj-$(CONFIG_HAS_IOMEM) += devres.o
--- /dev/null
+/*
+ * Copyright (C) 2008 Freescale Semiconductor, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/device.h> /* devres_*(), devm_ioremap_release() */
+#include <linux/io.h> /* ioremap_flags() */
+#include <linux/module.h> /* EXPORT_SYMBOL() */
+
+/**
+ * devm_ioremap_prot - Managed ioremap_flags()
+ * @dev: Generic device to remap IO address for
+ * @offset: BUS offset to map
+ * @size: Size of map
+ * @flags: Page flags
+ *
+ * Managed ioremap_prot(). Map is automatically unmapped on driver
+ * detach.
+ */
+void __iomem *devm_ioremap_prot(struct device *dev, resource_size_t offset,
+ size_t size, unsigned long flags)
+{
+ void __iomem **ptr, *addr;
+
+ ptr = devres_alloc(devm_ioremap_release, sizeof(*ptr), GFP_KERNEL);
+ if (!ptr)
+ return NULL;
+
+ addr = ioremap_flags(offset, size, flags);
+ if (addr) {
+ *ptr = addr;
+ devres_add(dev, ptr);
+ } else
+ devres_free(ptr);
+
+ return addr;
+}
+EXPORT_SYMBOL(devm_ioremap_prot);
dp = PDE(inode);
data = (unsigned int *)dp->data;
- if (!data) {
- printk(KERN_ERR "scanlog: read failed no data\n");
- return -EIO;
- }
-
if (count > RTAS_DATA_BUF_SIZE)
count = RTAS_DATA_BUF_SIZE;
struct proc_dir_entry *dp = PDE(inode);
unsigned int *data = (unsigned int *)dp->data;
- if (!data) {
- printk(KERN_ERR "scanlog: open failed no data\n");
- return -EIO;
- }
-
if (data[0] != 0) {
/* This imperfect test stops a second copy of the
* data (or a reset while data is being copied)
struct proc_dir_entry *dp = PDE(inode);
unsigned int *data = (unsigned int *)dp->data;
- if (!data) {
- printk(KERN_ERR "scanlog: release failed no data\n");
- return -EIO;
- }
data[0] = 0;
return 0;
if (!data)
goto err;
- ent = proc_create("ppc64/rtas/scan-log-dump", S_IRUSR, NULL,
- &scanlog_fops);
+ ent = proc_create_data("ppc64/rtas/scan-log-dump", S_IRUSR, NULL,
+ &scanlog_fops, data);
if (!ent)
goto err;
- ent->data = data;
proc_ppc64_scan_log_dump = ent;
return 0;
unsigned long stack_size)
{
unsigned long parent_tid_ptr, child_tid_ptr;
+ unsigned long orig_i1 = regs->u_regs[UREG_I1];
+ long ret;
parent_tid_ptr = regs->u_regs[UREG_I2];
child_tid_ptr = regs->u_regs[UREG_I4];
- return do_fork(clone_flags, stack_start,
- regs, stack_size,
- (int __user *) parent_tid_ptr,
- (int __user *) child_tid_ptr);
+ ret = do_fork(clone_flags, stack_start,
+ regs, stack_size,
+ (int __user *) parent_tid_ptr,
+ (int __user *) child_tid_ptr);
+
+ /* If we get an error and potentially restart the system
+ * call, we're screwed because copy_thread() clobbered
+ * the parent's %o1. So detect that case and restore it
+ * here.
+ */
+ if ((unsigned long)ret >= -ERESTART_RESTARTBLOCK)
+ regs->u_regs[UREG_I1] = orig_i1;
+
+ return ret;
}
/* Copy a Sparc thread. The fork() return value conventions
{
if (ARCH_SUN4C_SUN4 &&
(len > 0x20000000 ||
- ((flags & MAP_FIXED) &&
- addr < 0xe0000000 && addr + len > 0x20000000)))
+ (addr < 0xe0000000 && addr + len > 0x20000000)))
return -EINVAL;
/* See asm-sparc/uaccess.h */
unsigned long stack_size)
{
int __user *parent_tid_ptr, *child_tid_ptr;
+ unsigned long orig_i1 = regs->u_regs[UREG_I1];
+ long ret;
#ifdef CONFIG_COMPAT
if (test_thread_flag(TIF_32BIT)) {
child_tid_ptr = (int __user *) regs->u_regs[UREG_I4];
}
- return do_fork(clone_flags, stack_start,
- regs, stack_size,
- parent_tid_ptr, child_tid_ptr);
+ ret = do_fork(clone_flags, stack_start,
+ regs, stack_size,
+ parent_tid_ptr, child_tid_ptr);
+
+ /* If we get an error and potentially restart the system
+ * call, we're screwed because copy_thread() clobbered
+ * the parent's %o1. So detect that case and restore it
+ * here.
+ */
+ if ((unsigned long)ret >= -ERESTART_RESTARTBLOCK)
+ regs->u_regs[UREG_I1] = orig_i1;
+
+ return ret;
}
/* Copy a Sparc thread. The fork() return value conventions
void *info = call_data->info;
clear_softint(1 << irq);
-
- irq_enter();
-
- if (!call_data->wait) {
- /* let initiator proceed after getting data */
- atomic_inc(&call_data->finished);
- }
-
- func(info);
-
- irq_exit();
-
if (call_data->wait) {
/* let initiator proceed only after completion */
+ func(info);
atomic_inc(&call_data->finished);
+ } else {
+ /* let initiator proceed after getting data */
+ atomic_inc(&call_data->finished);
+ func(info);
}
}
void smp_receive_signal_client(int irq, struct pt_regs *regs)
{
- irq_enter();
clear_softint(1 << irq);
- irq_exit();
}
void smp_new_mmu_context_version_client(int irq, struct pt_regs *regs)
struct mm_struct *mm;
unsigned long flags;
- irq_enter();
-
clear_softint(1 << irq);
/* See if we need to allocate a new TLB context because
load_secondary_context(mm);
__flush_tlb_mm(CTX_HWBITS(mm->context),
SECONDARY_CONTEXT);
-
- irq_exit();
}
void smp_new_mmu_context_version(void)
{
clear_softint(1 << irq);
- irq_enter();
-
preempt_disable();
__asm__ __volatile__("flushw");
prom_world(0);
preempt_enable();
-
- irq_exit();
}
/* /proc/profile writes can call this, don't __init it please. */
if (len >= STACK_TOP32)
return -EINVAL;
- if ((flags & MAP_FIXED) && addr > STACK_TOP32 - len)
+ if (addr > STACK_TOP32 - len)
return -EINVAL;
} else {
if (len >= VA_EXCLUDE_START)
return -EINVAL;
- if ((flags & MAP_FIXED) && invalid_64bit_range(addr, len))
+ if (invalid_64bit_range(addr, len))
return -EINVAL;
}
/* 32-bit timeval and related flotsam. */
-static long get_tv32(struct timeval *o, struct compat_timeval __user *i)
-{
- return (!access_ok(VERIFY_READ, i, sizeof(*i)) ||
- (__get_user(o->tv_sec, &i->tv_sec) |
- __get_user(o->tv_usec, &i->tv_usec)));
-}
-
static inline long put_tv32(struct compat_timeval __user *o, struct timeval *i)
{
return (!access_ok(VERIFY_WRITE, o, sizeof(*o)) ||
return do_sys_settimeofday(tv ? &kts : NULL, tz ? &ktz : NULL);
}
-asmlinkage long sys32_utimes(char __user *filename,
- struct compat_timeval __user *tvs)
-{
- struct timespec tv[2];
-
- if (tvs) {
- struct timeval ktvs[2];
- if (get_tv32(&ktvs[0], tvs) ||
- get_tv32(&ktvs[1], 1+tvs))
- return -EFAULT;
-
- if (ktvs[0].tv_usec < 0 || ktvs[0].tv_usec >= 1000000 ||
- ktvs[1].tv_usec < 0 || ktvs[1].tv_usec >= 1000000)
- return -EINVAL;
-
- tv[0].tv_sec = ktvs[0].tv_sec;
- tv[0].tv_nsec = 1000 * ktvs[0].tv_usec;
- tv[1].tv_sec = ktvs[1].tv_sec;
- tv[1].tv_nsec = 1000 * ktvs[1].tv_usec;
- }
-
- return do_utimes(AT_FDCWD, filename, tvs ? tv : NULL, 0);
-}
-
/* These are here just in case some old sparc32 binary calls it. */
asmlinkage long sys32_pause(void)
{
/*120*/ .word compat_sys_readv, compat_sys_writev, sys32_settimeofday, sys32_fchown16, sys_fchmod
.word sys_nis_syscall, sys32_setreuid16, sys32_setregid16, sys_rename, sys_truncate
/*130*/ .word sys_ftruncate, sys_flock, compat_sys_lstat64, sys_nis_syscall, sys_nis_syscall
- .word sys_nis_syscall, sys32_mkdir, sys_rmdir, sys32_utimes, compat_sys_stat64
+ .word sys_nis_syscall, sys32_mkdir, sys_rmdir, compat_sys_utimes, compat_sys_stat64
/*140*/ .word sys32_sendfile64, sys_nis_syscall, sys32_futex, sys_gettid, compat_sys_getrlimit
.word compat_sys_setrlimit, sys_pivot_root, sys32_prctl, sys_pciconfig_read, sys_pciconfig_write
/*150*/ .word sys_nis_syscall, sys_inotify_init, sys_inotify_add_watch, sys_poll, sys_getdents64
initrd_end = ramdisk_image + sparc_ramdisk_size;
lmb_reserve(initrd_start, initrd_end);
+
+ initrd_start += PAGE_OFFSET;
+ initrd_end += PAGE_OFFSET;
}
#endif
}
__asm__ __volatile__("wrpr %0, 0, %%pstate"
: : "r" (pstate));
}
-
-#ifdef CONFIG_MEMORY_HOTPLUG
-
-void online_page(struct page *page)
-{
- ClearPageReserved(page);
- init_page_count(page);
- __free_page(page);
- totalram_pages++;
- num_physpages++;
-}
-
-#endif /* CONFIG_MEMORY_HOTPLUG */
KERNEL_DEFINES = $(strip -Derrno=kernel_errno -Dsigprocmask=kernel_sigprocmask \
-Dmktime=kernel_mktime $(ARCH_KERNEL_DEFINES))
KBUILD_CFLAGS += $(KERNEL_DEFINES)
-# Disable unit-at-a-time mode on pre-gcc-4.0 compilers, it makes gcc use
-# a lot more stack due to the lack of sharing of stacklots:
-KBUILD_CFLAGS += $(shell if [ $(call cc-version) -lt 0400 ] ; then \
- echo $(call cc-option,-fno-unit-at-a-time); fi ;)
+KBUILD_CFLAGS += $(call cc-option,-fno-unit-at-a-time,)
PHONY += linux
### Arch settings
config X86
def_bool y
+ select HAVE_UNSTABLE_SCHED_CLOCK
select HAVE_IDE
select HAVE_OPROFILE
select HAVE_KPROBES
__FIXADDR_TOP = -reserve - PAGE_SIZE;
__VMALLOC_RESERVE += reserve;
}
-
-int pmd_bad(pmd_t pmd)
-{
- WARN_ON_ONCE(pmd_bad_v1(pmd) != pmd_bad_v2(pmd));
-
- return pmd_bad_v1(pmd);
-}
#include <asm/numa.h>
#include "pci.h"
-static int __devinit can_skip_ioresource_align(const struct dmi_system_id *d)
-{
- pci_probe |= PCI_CAN_SKIP_ISA_ALIGN;
- printk(KERN_INFO "PCI: %s detected, can skip ISA alignment\n", d->ident);
- return 0;
-}
-
-static struct dmi_system_id acpi_pciprobe_dmi_table[] __devinitdata = {
-/*
- * Systems where PCI IO resource ISA alignment can be skipped
- * when the ISA enable bit in the bridge control is not set
- */
- {
- .callback = can_skip_ioresource_align,
- .ident = "IBM System x3800",
- .matches = {
- DMI_MATCH(DMI_SYS_VENDOR, "IBM"),
- DMI_MATCH(DMI_PRODUCT_NAME, "x3800"),
- },
- },
- {
- .callback = can_skip_ioresource_align,
- .ident = "IBM System x3850",
- .matches = {
- DMI_MATCH(DMI_SYS_VENDOR, "IBM"),
- DMI_MATCH(DMI_PRODUCT_NAME, "x3850"),
- },
- },
- {
- .callback = can_skip_ioresource_align,
- .ident = "IBM System x3950",
- .matches = {
- DMI_MATCH(DMI_SYS_VENDOR, "IBM"),
- DMI_MATCH(DMI_PRODUCT_NAME, "x3950"),
- },
- },
- {}
-};
-
struct pci_root_info {
char *name;
unsigned int res_num;
int pxm;
#endif
- dmi_check_system(acpi_pciprobe_dmi_table);
-
if (domain && !pci_domains_supported) {
printk(KERN_WARNING "PCI: Multiple domains not supported "
"(dom %d, bus %d)\n", domain, busnum);
rom_r->start = rom_r->end = rom_r->flags = 0;
}
+static int __devinit can_skip_ioresource_align(const struct dmi_system_id *d)
+{
+ pci_probe |= PCI_CAN_SKIP_ISA_ALIGN;
+ printk(KERN_INFO "PCI: %s detected, can skip ISA alignment\n", d->ident);
+ return 0;
+}
+
+static struct dmi_system_id can_skip_pciprobe_dmi_table[] __devinitdata = {
+/*
+ * Systems where PCI IO resource ISA alignment can be skipped
+ * when the ISA enable bit in the bridge control is not set
+ */
+ {
+ .callback = can_skip_ioresource_align,
+ .ident = "IBM System x3800",
+ .matches = {
+ DMI_MATCH(DMI_SYS_VENDOR, "IBM"),
+ DMI_MATCH(DMI_PRODUCT_NAME, "x3800"),
+ },
+ },
+ {
+ .callback = can_skip_ioresource_align,
+ .ident = "IBM System x3850",
+ .matches = {
+ DMI_MATCH(DMI_SYS_VENDOR, "IBM"),
+ DMI_MATCH(DMI_PRODUCT_NAME, "x3850"),
+ },
+ },
+ {
+ .callback = can_skip_ioresource_align,
+ .ident = "IBM System x3950",
+ .matches = {
+ DMI_MATCH(DMI_SYS_VENDOR, "IBM"),
+ DMI_MATCH(DMI_PRODUCT_NAME, "x3950"),
+ },
+ },
+ {}
+};
+
+void __init dmi_check_skip_isa_align(void)
+{
+ dmi_check_system(can_skip_pciprobe_dmi_table);
+}
+
/*
* Called after each bus is probed, but before its children
* are examined.
{}
};
+void __init dmi_check_pciprobe(void)
+{
+ dmi_check_system(pciprobe_dmi_table);
+}
+
struct pci_bus * __devinit pcibios_scan_root(int busnum)
{
struct pci_bus *bus = NULL;
struct pci_sysdata *sd;
- dmi_check_system(pciprobe_dmi_table);
-
while ((bus = pci_find_next_bus(bus)) != NULL) {
if (bus->number == busnum) {
/* Already scanned */
} else if (!strcmp(str, "routeirq")) {
pci_routeirq = 1;
return NULL;
+ } else if (!strcmp(str, "skip_isa_align")) {
+ pci_probe |= PCI_CAN_SKIP_ISA_ALIGN;
+ return NULL;
}
return str;
}
pcibios_disable_irq(dev);
}
-struct pci_bus *pci_scan_bus_on_node(int busno, struct pci_ops *ops, int node)
+struct pci_bus * __devinit pci_scan_bus_on_node(int busno, struct pci_ops *ops, int node)
{
struct pci_bus *bus = NULL;
struct pci_sysdata *sd;
return bus;
}
-struct pci_bus *pci_scan_bus_with_sysdata(int busno)
+struct pci_bus * __devinit pci_scan_bus_with_sysdata(int busno)
{
return pci_scan_bus_on_node(busno, &pci_root_ops, -1);
}
*/
static void fam10h_pci_cfg_space_size(struct pci_dev *dev)
{
- dev->cfg_size = pci_cfg_space_size_ext(dev, 0);
+ dev->cfg_size = pci_cfg_space_size_ext(dev);
}
DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1200, fam10h_pci_cfg_space_size);
printk(KERN_ERR
"PCI: Fatal: No config space access function found\n");
+ dmi_check_pciprobe();
+
+ dmi_check_skip_isa_align();
+
return 0;
}
arch_initcall(pci_access_init);
pci_dmi_bf,
};
+extern void __init dmi_check_pciprobe(void);
+extern void __init dmi_check_skip_isa_align(void);
+
/* pci-i386.c */
extern unsigned int pcibios_max_latency;
If unsure, say N.
config SATA_INIC162X
- tristate "Initio 162x SATA support (HIGHLY EXPERIMENTAL)"
- depends on PCI && EXPERIMENTAL
+ tristate "Initio 162x SATA support"
+ depends on PCI
help
This option enables support for Initio 162x Serial ATA.
If unsure, say N.
+config PATA_SCH
+ tristate "Intel SCH PATA support"
+ depends on PCI
+ help
+ This option enables support for Intel SCH PATA on the Intel
+ SCH (US15W, US15L, UL11L) series host controllers.
+
+ If unsure, say N.
+
config PATA_BF54X
tristate "Blackfin 54x ATAPI support"
depends on BF542 || BF548 || BF549
obj-$(CONFIG_PATA_TRIFLEX) += pata_triflex.o
obj-$(CONFIG_PATA_IXP4XX_CF) += pata_ixp4xx_cf.o
obj-$(CONFIG_PATA_SCC) += pata_scc.o
+obj-$(CONFIG_PATA_SCH) += pata_sch.o
obj-$(CONFIG_PATA_BF54X) += pata_bf54x.o
obj-$(CONFIG_PATA_PLATFORM) += pata_platform.o
obj-$(CONFIG_PATA_OF_PLATFORM) += pata_of_platform.o
void __iomem *port_mmio = ahci_port_base(link->ap);
u8 status = readl(port_mmio + PORT_TFDATA) & 0xFF;
- if (!(status & ATA_BUSY))
- return 1;
- return 0;
+ return ata_check_ready(status);
}
static int ahci_softreset(struct ata_link *link, unsigned int *class,
if (dev->vendor == PCI_VENDOR_ID_AL)
ata_pci_bmdma_clear_simplex(dev);
+ if (dev->vendor == PCI_VENDOR_ID_ATI) {
+ int rc = pcim_enable_device(dev);
+ if (rc < 0)
+ return rc;
+ pcim_pin_device(dev);
+ }
return ata_pci_sff_init_one(dev, ppi, &generic_sht, NULL);
}
{
struct pci_dev *pdev = to_pci_dev(host->dev);
struct piix_host_priv *hpriv = host->private_data;
+ struct ata_device *dev0 = &host->ports[0]->link.device[0];
+ u32 scontrol;
int i;
/* check for availability */
return;
hpriv->sidpr = pcim_iomap_table(pdev)[PIIX_SIDPR_BAR];
+
+ /* SCR access via SIDPR doesn't work on some configurations.
+ * Give it a test drive by inhibiting power save modes which
+ * we'll do anyway.
+ */
+ scontrol = piix_sidpr_read(dev0, SCR_CONTROL);
+
+ /* if IPM is already 3, SCR access is probably working. Don't
+ * un-inhibit power save modes as BIOS might have inhibited
+ * them for a reason.
+ */
+ if ((scontrol & 0xf00) != 0x300) {
+ scontrol |= 0x300;
+ piix_sidpr_write(dev0, SCR_CONTROL, scontrol);
+ scontrol = piix_sidpr_read(dev0, SCR_CONTROL);
+
+ if ((scontrol & 0xf00) != 0x300) {
+ dev_printk(KERN_INFO, host->dev, "SCR access via "
+ "SIDPR is available but doesn't work\n");
+ return;
+ }
+ }
+
host->ports[0]->ops = &piix_sidpr_sata_ops;
host->ports[1]->ops = &piix_sidpr_sata_ops;
}
EXPORT_SYMBOL_GPL(ata_eh_thaw_port);
EXPORT_SYMBOL_GPL(ata_eh_qc_complete);
EXPORT_SYMBOL_GPL(ata_eh_qc_retry);
+EXPORT_SYMBOL_GPL(ata_eh_analyze_ncq_error);
EXPORT_SYMBOL_GPL(ata_do_eh);
EXPORT_SYMBOL_GPL(ata_std_error_handler);
* LOCKING:
* Kernel thread context (may sleep).
*/
-static void ata_eh_analyze_ncq_error(struct ata_link *link)
+void ata_eh_analyze_ncq_error(struct ata_link *link)
{
struct ata_port *ap = link->ap;
struct ata_eh_context *ehc = &link->eh_context;
{
u8 status = link->ap->ops->sff_check_status(link->ap);
- if (!(status & ATA_BUSY))
- return 1;
- if (status == 0xff)
- return -ENODEV;
- return 0;
+ return ata_check_ready(status);
}
/**
.port_ops = &pacpi_ops,
};
const struct ata_port_info *ppi[] = { &info, NULL };
+ if (pdev->vendor == PCI_VENDOR_ID_ATI) {
+ int rc = pcim_enable_device(pdev);
+ if (rc < 0)
+ return rc;
+ pcim_pin_device(pdev);
+ }
return ata_pci_sff_init_one(pdev, ppi, &pacpi_sht, NULL);
}
--- /dev/null
+/*
+ * pata_sch.c - Intel SCH PATA controllers
+ *
+ * Copyright (c) 2008 Alek Du <alek.du@intel.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License 2 as published
+ * by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; see the file COPYING. If not, write to
+ * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+/*
+ * Supports:
+ * Intel SCH (AF82US15W, AF82US15L, AF82UL11L) chipsets -- see spec at:
+ * http://download.intel.com/design/chipsets/embedded/datashts/319537.pdf
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/init.h>
+#include <linux/blkdev.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <scsi/scsi_host.h>
+#include <linux/libata.h>
+#include <linux/dmi.h>
+
+#define DRV_NAME "pata_sch"
+#define DRV_VERSION "0.2"
+
+/* see SCH datasheet page 351 */
+enum {
+ D0TIM = 0x80, /* Device 0 Timing Register */
+ D1TIM = 0x84, /* Device 1 Timing Register */
+ PM = 0x07, /* PIO Mode Bit Mask */
+ MDM = (0x03 << 8), /* Multi-word DMA Mode Bit Mask */
+ UDM = (0x07 << 16), /* Ultra DMA Mode Bit Mask */
+ PPE = (1 << 30), /* Prefetch/Post Enable */
+ USD = (1 << 31), /* Use Synchronous DMA */
+};
+
+static int sch_init_one(struct pci_dev *pdev,
+ const struct pci_device_id *ent);
+static void sch_set_piomode(struct ata_port *ap, struct ata_device *adev);
+static void sch_set_dmamode(struct ata_port *ap, struct ata_device *adev);
+
+static const struct pci_device_id sch_pci_tbl[] = {
+ /* Intel SCH PATA Controller */
+ { PCI_VDEVICE(INTEL, PCI_DEVICE_ID_INTEL_SCH_IDE), 0 },
+ { } /* terminate list */
+};
+
+static struct pci_driver sch_pci_driver = {
+ .name = DRV_NAME,
+ .id_table = sch_pci_tbl,
+ .probe = sch_init_one,
+ .remove = ata_pci_remove_one,
+#ifdef CONFIG_PM
+ .suspend = ata_pci_device_suspend,
+ .resume = ata_pci_device_resume,
+#endif
+};
+
+static struct scsi_host_template sch_sht = {
+ ATA_BMDMA_SHT(DRV_NAME),
+};
+
+static struct ata_port_operations sch_pata_ops = {
+ .inherits = &ata_bmdma_port_ops,
+ .cable_detect = ata_cable_unknown,
+ .set_piomode = sch_set_piomode,
+ .set_dmamode = sch_set_dmamode,
+};
+
+static struct ata_port_info sch_port_info = {
+ .flags = 0,
+ .pio_mask = ATA_PIO4, /* pio0-4 */
+ .mwdma_mask = ATA_MWDMA2, /* mwdma0-2 */
+ .udma_mask = ATA_UDMA5, /* udma0-5 */
+ .port_ops = &sch_pata_ops,
+};
+
+MODULE_AUTHOR("Alek Du <alek.du@intel.com>");
+MODULE_DESCRIPTION("SCSI low-level driver for Intel SCH PATA controllers");
+MODULE_LICENSE("GPL");
+MODULE_DEVICE_TABLE(pci, sch_pci_tbl);
+MODULE_VERSION(DRV_VERSION);
+
+/**
+ * sch_set_piomode - Initialize host controller PATA PIO timings
+ * @ap: Port whose timings we are configuring
+ * @adev: ATA device
+ *
+ * Set PIO mode for device, in host controller PCI config space.
+ *
+ * LOCKING:
+ * None (inherited from caller).
+ */
+
+static void sch_set_piomode(struct ata_port *ap, struct ata_device *adev)
+{
+ unsigned int pio = adev->pio_mode - XFER_PIO_0;
+ struct pci_dev *dev = to_pci_dev(ap->host->dev);
+ unsigned int port = adev->devno ? D1TIM : D0TIM;
+ unsigned int data;
+
+ pci_read_config_dword(dev, port, &data);
+ /* see SCH datasheet page 351 */
+ /* set PIO mode */
+ data &= ~(PM | PPE);
+ data |= pio;
+ /* enable PPE for block device */
+ if (adev->class == ATA_DEV_ATA)
+ data |= PPE;
+ pci_write_config_dword(dev, port, data);
+}
+
+/**
+ * sch_set_dmamode - Initialize host controller PATA DMA timings
+ * @ap: Port whose timings we are configuring
+ * @adev: ATA device
+ *
+ * Set MW/UDMA mode for device, in host controller PCI config space.
+ *
+ * LOCKING:
+ * None (inherited from caller).
+ */
+
+static void sch_set_dmamode(struct ata_port *ap, struct ata_device *adev)
+{
+ unsigned int dma_mode = adev->dma_mode;
+ struct pci_dev *dev = to_pci_dev(ap->host->dev);
+ unsigned int port = adev->devno ? D1TIM : D0TIM;
+ unsigned int data;
+
+ pci_read_config_dword(dev, port, &data);
+ /* see SCH datasheet page 351 */
+ if (dma_mode >= XFER_UDMA_0) {
+ /* enable Synchronous DMA mode */
+ data |= USD;
+ data &= ~UDM;
+ data |= (dma_mode - XFER_UDMA_0) << 16;
+ } else { /* must be MWDMA mode, since we masked SWDMA already */
+ data &= ~(USD | MDM);
+ data |= (dma_mode - XFER_MW_DMA_0) << 8;
+ }
+ pci_write_config_dword(dev, port, data);
+}
+
+/**
+ * sch_init_one - Register SCH ATA PCI device with kernel services
+ * @pdev: PCI device to register
+ * @ent: Entry in sch_pci_tbl matching with @pdev
+ *
+ * LOCKING:
+ * Inherited from PCI layer (may sleep).
+ *
+ * RETURNS:
+ * Zero on success, or -ERRNO value.
+ */
+
+static int __devinit sch_init_one(struct pci_dev *pdev,
+ const struct pci_device_id *ent)
+{
+ static int printed_version;
+ const struct ata_port_info *ppi[] = { &sch_port_info, NULL };
+ struct ata_host *host;
+ int rc;
+
+ if (!printed_version++)
+ dev_printk(KERN_DEBUG, &pdev->dev,
+ "version " DRV_VERSION "\n");
+
+ /* enable device and prepare host */
+ rc = pcim_enable_device(pdev);
+ if (rc)
+ return rc;
+ rc = ata_pci_sff_prepare_host(pdev, ppi, &host);
+ if (rc)
+ return rc;
+ pci_set_master(pdev);
+ return ata_pci_sff_activate_host(host, ata_sff_interrupt, &sch_sht);
+}
+
+static int __init sch_init(void)
+{
+ return pci_register_driver(&sch_pci_driver);
+}
+
+static void __exit sch_exit(void)
+{
+ pci_unregister_driver(&sch_pci_driver);
+}
+
+module_init(sch_init);
+module_exit(sch_exit);
* right. Documentation is available at initio's website but it only
* documents registers (not programming model).
*
- * - ATA disks work.
- * - Hotplug works.
- * - ATAPI read works but burning doesn't. This thing is really
- * peculiar about ATAPI and I couldn't figure out how ATAPI PIO and
- * ATAPI DMA WRITE should be programmed. If you've got a clue, be
- * my guest.
- * - Both STR and STD work.
+ * This driver has interesting history. The first version was written
+ * from the documentation and a 2.4 IDE driver posted on a Taiwan
+ * company, which didn't use any IDMA features and couldn't handle
+ * LBA48. The resulting driver couldn't handle LBA48 devices either
+ * making it pretty useless.
+ *
+ * After a while, initio picked the driver up, renamed it to
+ * sata_initio162x, updated it to use IDMA for ATA DMA commands and
+ * posted it on their website. It only used ATA_PROT_DMA for IDMA and
+ * attaching both devices and issuing IDMA and !IDMA commands
+ * simultaneously broke it due to PIRQ masking interaction but it did
+ * show how to use the IDMA (ADMA + some initio specific twists)
+ * engine.
+ *
+ * Then, I picked up their changes again and here's the usable driver
+ * which uses IDMA for everything. Everything works now including
+ * LBA48, CD/DVD burning, suspend/resume and hotplug. There are some
+ * issues tho. Result Tf is not resported properly, NCQ isn't
+ * supported yet and CD/DVD writing works with DMA assisted PIO
+ * protocol (which, for native SATA devices, shouldn't cause any
+ * noticeable difference).
+ *
+ * Anyways, so, here's finally a working driver for inic162x. Enjoy!
+ *
+ * initio: If you guys wanna improve the driver regarding result TF
+ * access and other stuff, please feel free to contact me. I'll be
+ * happy to assist.
*/
#include <linux/kernel.h>
#include <scsi/scsi_device.h>
#define DRV_NAME "sata_inic162x"
-#define DRV_VERSION "0.3"
+#define DRV_VERSION "0.4"
enum {
- MMIO_BAR = 5,
+ MMIO_BAR_PCI = 5,
+ MMIO_BAR_CARDBUS = 1,
NR_PORTS = 2,
+ IDMA_CPB_TBL_SIZE = 4 * 32,
+
+ INIC_DMA_BOUNDARY = 0xffffff,
+
+ HOST_ACTRL = 0x08,
HOST_CTL = 0x7c,
HOST_STAT = 0x7e,
HOST_IRQ_STAT = 0xbc,
PORT_SIZE = 0x40,
/* registers for ATA TF operation */
- PORT_TF = 0x00,
- PORT_ALT_STAT = 0x08,
+ PORT_TF_DATA = 0x00,
+ PORT_TF_FEATURE = 0x01,
+ PORT_TF_NSECT = 0x02,
+ PORT_TF_LBAL = 0x03,
+ PORT_TF_LBAM = 0x04,
+ PORT_TF_LBAH = 0x05,
+ PORT_TF_DEVICE = 0x06,
+ PORT_TF_COMMAND = 0x07,
+ PORT_TF_ALT_STAT = 0x08,
PORT_IRQ_STAT = 0x09,
PORT_IRQ_MASK = 0x0a,
PORT_PRD_CTL = 0x0b,
PORT_PRD_ADDR = 0x0c,
PORT_PRD_XFERLEN = 0x10,
+ PORT_CPB_CPBLAR = 0x18,
+ PORT_CPB_PTQFIFO = 0x1c,
/* IDMA register */
PORT_IDMA_CTL = 0x14,
+ PORT_IDMA_STAT = 0x16,
+
+ PORT_RPQ_FIFO = 0x1e,
+ PORT_RPQ_CNT = 0x1f,
PORT_SCR = 0x20,
/* HOST_CTL bits */
HCTL_IRQOFF = (1 << 8), /* global IRQ off */
- HCTL_PWRDWN = (1 << 13), /* power down PHYs */
+ HCTL_FTHD0 = (1 << 10), /* fifo threshold 0 */
+ HCTL_FTHD1 = (1 << 11), /* fifo threshold 1*/
+ HCTL_PWRDWN = (1 << 12), /* power down PHYs */
HCTL_SOFTRST = (1 << 13), /* global reset (no phy reset) */
HCTL_RPGSEL = (1 << 15), /* register page select */
PIRQ_PENDING = (1 << 7), /* port IRQ pending (STAT only) */
PIRQ_ERR = PIRQ_OFFLINE | PIRQ_ONLINE | PIRQ_FATAL,
-
- PIRQ_MASK_DMA_READ = PIRQ_REPLY | PIRQ_ATA,
- PIRQ_MASK_OTHER = PIRQ_REPLY | PIRQ_COMPLETE,
+ PIRQ_MASK_DEFAULT = PIRQ_REPLY | PIRQ_ATA,
PIRQ_MASK_FREEZE = 0xff,
/* PORT_PRD_CTL bits */
IDMA_CTL_RST_IDMA = (1 << 5), /* reset IDMA machinary */
IDMA_CTL_GO = (1 << 7), /* IDMA mode go */
IDMA_CTL_ATA_NIEN = (1 << 8), /* ATA IRQ disable */
+
+ /* PORT_IDMA_STAT bits */
+ IDMA_STAT_PERR = (1 << 0), /* PCI ERROR MODE */
+ IDMA_STAT_CPBERR = (1 << 1), /* ADMA CPB error */
+ IDMA_STAT_LGCY = (1 << 3), /* ADMA legacy */
+ IDMA_STAT_UIRQ = (1 << 4), /* ADMA unsolicited irq */
+ IDMA_STAT_STPD = (1 << 5), /* ADMA stopped */
+ IDMA_STAT_PSD = (1 << 6), /* ADMA pause */
+ IDMA_STAT_DONE = (1 << 7), /* ADMA done */
+
+ IDMA_STAT_ERR = IDMA_STAT_PERR | IDMA_STAT_CPBERR,
+
+ /* CPB Control Flags*/
+ CPB_CTL_VALID = (1 << 0), /* CPB valid */
+ CPB_CTL_QUEUED = (1 << 1), /* queued command */
+ CPB_CTL_DATA = (1 << 2), /* data, rsvd in datasheet */
+ CPB_CTL_IEN = (1 << 3), /* PCI interrupt enable */
+ CPB_CTL_DEVDIR = (1 << 4), /* device direction control */
+
+ /* CPB Response Flags */
+ CPB_RESP_DONE = (1 << 0), /* ATA command complete */
+ CPB_RESP_REL = (1 << 1), /* ATA release */
+ CPB_RESP_IGNORED = (1 << 2), /* CPB ignored */
+ CPB_RESP_ATA_ERR = (1 << 3), /* ATA command error */
+ CPB_RESP_SPURIOUS = (1 << 4), /* ATA spurious interrupt error */
+ CPB_RESP_UNDERFLOW = (1 << 5), /* APRD deficiency length error */
+ CPB_RESP_OVERFLOW = (1 << 6), /* APRD exccess length error */
+ CPB_RESP_CPB_ERR = (1 << 7), /* CPB error flag */
+
+ /* PRD Control Flags */
+ PRD_DRAIN = (1 << 1), /* ignore data excess */
+ PRD_CDB = (1 << 2), /* atapi packet command pointer */
+ PRD_DIRECT_INTR = (1 << 3), /* direct interrupt */
+ PRD_DMA = (1 << 4), /* data transfer method */
+ PRD_WRITE = (1 << 5), /* data dir, rsvd in datasheet */
+ PRD_IOM = (1 << 6), /* io/memory transfer */
+ PRD_END = (1 << 7), /* APRD chain end */
};
+/* Comman Parameter Block */
+struct inic_cpb {
+ u8 resp_flags; /* Response Flags */
+ u8 error; /* ATA Error */
+ u8 status; /* ATA Status */
+ u8 ctl_flags; /* Control Flags */
+ __le32 len; /* Total Transfer Length */
+ __le32 prd; /* First PRD pointer */
+ u8 rsvd[4];
+ /* 16 bytes */
+ u8 feature; /* ATA Feature */
+ u8 hob_feature; /* ATA Ex. Feature */
+ u8 device; /* ATA Device/Head */
+ u8 mirctl; /* Mirror Control */
+ u8 nsect; /* ATA Sector Count */
+ u8 hob_nsect; /* ATA Ex. Sector Count */
+ u8 lbal; /* ATA Sector Number */
+ u8 hob_lbal; /* ATA Ex. Sector Number */
+ u8 lbam; /* ATA Cylinder Low */
+ u8 hob_lbam; /* ATA Ex. Cylinder Low */
+ u8 lbah; /* ATA Cylinder High */
+ u8 hob_lbah; /* ATA Ex. Cylinder High */
+ u8 command; /* ATA Command */
+ u8 ctl; /* ATA Control */
+ u8 slave_error; /* Slave ATA Error */
+ u8 slave_status; /* Slave ATA Status */
+ /* 32 bytes */
+} __packed;
+
+/* Physical Region Descriptor */
+struct inic_prd {
+ __le32 mad; /* Physical Memory Address */
+ __le16 len; /* Transfer Length */
+ u8 rsvd;
+ u8 flags; /* Control Flags */
+} __packed;
+
+struct inic_pkt {
+ struct inic_cpb cpb;
+ struct inic_prd prd[LIBATA_MAX_PRD + 1]; /* + 1 for cdb */
+ u8 cdb[ATAPI_CDB_LEN];
+} __packed;
+
struct inic_host_priv {
- u16 cached_hctl;
+ void __iomem *mmio_base;
+ u16 cached_hctl;
};
struct inic_port_priv {
- u8 dfl_prdctl;
- u8 cached_prdctl;
- u8 cached_pirq_mask;
+ struct inic_pkt *pkt;
+ dma_addr_t pkt_dma;
+ u32 *cpb_tbl;
+ dma_addr_t cpb_tbl_dma;
};
static struct scsi_host_template inic_sht = {
- ATA_BMDMA_SHT(DRV_NAME),
+ ATA_BASE_SHT(DRV_NAME),
+ .sg_tablesize = LIBATA_MAX_PRD, /* maybe it can be larger? */
+ .dma_boundary = INIC_DMA_BOUNDARY,
};
static const int scr_map[] = {
static void __iomem *inic_port_base(struct ata_port *ap)
{
- return ap->host->iomap[MMIO_BAR] + ap->port_no * PORT_SIZE;
-}
-
-static void __inic_set_pirq_mask(struct ata_port *ap, u8 mask)
-{
- void __iomem *port_base = inic_port_base(ap);
- struct inic_port_priv *pp = ap->private_data;
+ struct inic_host_priv *hpriv = ap->host->private_data;
- writeb(mask, port_base + PORT_IRQ_MASK);
- pp->cached_pirq_mask = mask;
-}
-
-static void inic_set_pirq_mask(struct ata_port *ap, u8 mask)
-{
- struct inic_port_priv *pp = ap->private_data;
-
- if (pp->cached_pirq_mask != mask)
- __inic_set_pirq_mask(ap, mask);
+ return hpriv->mmio_base + ap->port_no * PORT_SIZE;
}
static void inic_reset_port(void __iomem *port_base)
{
void __iomem *idma_ctl = port_base + PORT_IDMA_CTL;
- u16 ctl;
- ctl = readw(idma_ctl);
- ctl &= ~(IDMA_CTL_RST_IDMA | IDMA_CTL_ATA_NIEN | IDMA_CTL_GO);
+ /* stop IDMA engine */
+ readw(idma_ctl); /* flush */
+ msleep(1);
/* mask IRQ and assert reset */
- writew(ctl | IDMA_CTL_RST_IDMA | IDMA_CTL_ATA_NIEN, idma_ctl);
+ writew(IDMA_CTL_RST_IDMA, idma_ctl);
readw(idma_ctl); /* flush */
-
- /* give it some time */
msleep(1);
/* release reset */
- writew(ctl | IDMA_CTL_ATA_NIEN, idma_ctl);
+ writew(0, idma_ctl);
/* clear irq */
writeb(0xff, port_base + PORT_IRQ_STAT);
-
- /* reenable ATA IRQ, turn off IDMA mode */
- writew(ctl, idma_ctl);
}
static int inic_scr_read(struct ata_port *ap, unsigned sc_reg, u32 *val)
{
- void __iomem *scr_addr = ap->ioaddr.scr_addr;
+ void __iomem *scr_addr = inic_port_base(ap) + PORT_SCR;
void __iomem *addr;
if (unlikely(sc_reg >= ARRAY_SIZE(scr_map)))
static int inic_scr_write(struct ata_port *ap, unsigned sc_reg, u32 val)
{
- void __iomem *scr_addr = ap->ioaddr.scr_addr;
- void __iomem *addr;
+ void __iomem *scr_addr = inic_port_base(ap) + PORT_SCR;
if (unlikely(sc_reg >= ARRAY_SIZE(scr_map)))
return -EINVAL;
- addr = scr_addr + scr_map[sc_reg] * 4;
writel(val, scr_addr + scr_map[sc_reg] * 4);
return 0;
}
-/*
- * In TF mode, inic162x is very similar to SFF device. TF registers
- * function the same. DMA engine behaves similary using the same PRD
- * format as BMDMA but different command register, interrupt and event
- * notification methods are used. The following inic_bmdma_*()
- * functions do the impedance matching.
- */
-static void inic_bmdma_setup(struct ata_queued_cmd *qc)
+static void inic_stop_idma(struct ata_port *ap)
{
- struct ata_port *ap = qc->ap;
- struct inic_port_priv *pp = ap->private_data;
void __iomem *port_base = inic_port_base(ap);
- int rw = qc->tf.flags & ATA_TFLAG_WRITE;
-
- /* make sure device sees PRD table writes */
- wmb();
-
- /* load transfer length */
- writel(qc->nbytes, port_base + PORT_PRD_XFERLEN);
-
- /* turn on DMA and specify data direction */
- pp->cached_prdctl = pp->dfl_prdctl | PRD_CTL_DMAEN;
- if (!rw)
- pp->cached_prdctl |= PRD_CTL_WR;
- writeb(pp->cached_prdctl, port_base + PORT_PRD_CTL);
- /* issue r/w command */
- ap->ops->sff_exec_command(ap, &qc->tf);
+ readb(port_base + PORT_RPQ_FIFO);
+ readb(port_base + PORT_RPQ_CNT);
+ writew(0, port_base + PORT_IDMA_CTL);
}
-static void inic_bmdma_start(struct ata_queued_cmd *qc)
+static void inic_host_err_intr(struct ata_port *ap, u8 irq_stat, u16 idma_stat)
{
- struct ata_port *ap = qc->ap;
+ struct ata_eh_info *ehi = &ap->link.eh_info;
struct inic_port_priv *pp = ap->private_data;
- void __iomem *port_base = inic_port_base(ap);
+ struct inic_cpb *cpb = &pp->pkt->cpb;
+ bool freeze = false;
- /* start host DMA transaction */
- pp->cached_prdctl |= PRD_CTL_START;
- writeb(pp->cached_prdctl, port_base + PORT_PRD_CTL);
-}
+ ata_ehi_clear_desc(ehi);
+ ata_ehi_push_desc(ehi, "irq_stat=0x%x idma_stat=0x%x",
+ irq_stat, idma_stat);
-static void inic_bmdma_stop(struct ata_queued_cmd *qc)
-{
- struct ata_port *ap = qc->ap;
- struct inic_port_priv *pp = ap->private_data;
- void __iomem *port_base = inic_port_base(ap);
+ inic_stop_idma(ap);
- /* stop DMA engine */
- writeb(pp->dfl_prdctl, port_base + PORT_PRD_CTL);
-}
+ if (irq_stat & (PIRQ_OFFLINE | PIRQ_ONLINE)) {
+ ata_ehi_push_desc(ehi, "hotplug");
+ ata_ehi_hotplugged(ehi);
+ freeze = true;
+ }
-static u8 inic_bmdma_status(struct ata_port *ap)
-{
- /* event is already verified by the interrupt handler */
- return ATA_DMA_INTR;
+ if (idma_stat & IDMA_STAT_PERR) {
+ ata_ehi_push_desc(ehi, "PCI error");
+ freeze = true;
+ }
+
+ if (idma_stat & IDMA_STAT_CPBERR) {
+ ata_ehi_push_desc(ehi, "CPB error");
+
+ if (cpb->resp_flags & CPB_RESP_IGNORED) {
+ __ata_ehi_push_desc(ehi, " ignored");
+ ehi->err_mask |= AC_ERR_INVALID;
+ freeze = true;
+ }
+
+ if (cpb->resp_flags & CPB_RESP_ATA_ERR)
+ ehi->err_mask |= AC_ERR_DEV;
+
+ if (cpb->resp_flags & CPB_RESP_SPURIOUS) {
+ __ata_ehi_push_desc(ehi, " spurious-intr");
+ ehi->err_mask |= AC_ERR_HSM;
+ freeze = true;
+ }
+
+ if (cpb->resp_flags &
+ (CPB_RESP_UNDERFLOW | CPB_RESP_OVERFLOW)) {
+ __ata_ehi_push_desc(ehi, " data-over/underflow");
+ ehi->err_mask |= AC_ERR_HSM;
+ freeze = true;
+ }
+ }
+
+ if (freeze)
+ ata_port_freeze(ap);
+ else
+ ata_port_abort(ap);
}
static void inic_host_intr(struct ata_port *ap)
{
void __iomem *port_base = inic_port_base(ap);
- struct ata_eh_info *ehi = &ap->link.eh_info;
+ struct ata_queued_cmd *qc = ata_qc_from_tag(ap, ap->link.active_tag);
u8 irq_stat;
+ u16 idma_stat;
- /* fetch and clear irq */
+ /* read and clear IRQ status */
irq_stat = readb(port_base + PORT_IRQ_STAT);
writeb(irq_stat, port_base + PORT_IRQ_STAT);
+ idma_stat = readw(port_base + PORT_IDMA_STAT);
- if (likely(!(irq_stat & PIRQ_ERR))) {
- struct ata_queued_cmd *qc =
- ata_qc_from_tag(ap, ap->link.active_tag);
+ if (unlikely((irq_stat & PIRQ_ERR) || (idma_stat & IDMA_STAT_ERR)))
+ inic_host_err_intr(ap, irq_stat, idma_stat);
- if (unlikely(!qc || (qc->tf.flags & ATA_TFLAG_POLLING))) {
- ap->ops->sff_check_status(ap); /* clear ATA interrupt */
- return;
- }
+ if (unlikely(!qc))
+ goto spurious;
- if (likely(ata_sff_host_intr(ap, qc)))
- return;
+ if (likely(idma_stat & IDMA_STAT_DONE)) {
+ inic_stop_idma(ap);
- ap->ops->sff_check_status(ap); /* clear ATA interrupt */
- ata_port_printk(ap, KERN_WARNING, "unhandled "
- "interrupt, irq_stat=%x\n", irq_stat);
+ /* Depending on circumstances, device error
+ * isn't reported by IDMA, check it explicitly.
+ */
+ if (unlikely(readb(port_base + PORT_TF_COMMAND) &
+ (ATA_DF | ATA_ERR)))
+ qc->err_mask |= AC_ERR_DEV;
+
+ ata_qc_complete(qc);
return;
}
- /* error */
- ata_ehi_push_desc(ehi, "irq_stat=0x%x", irq_stat);
-
- if (irq_stat & (PIRQ_OFFLINE | PIRQ_ONLINE)) {
- ata_ehi_hotplugged(ehi);
- ata_port_freeze(ap);
- } else
- ata_port_abort(ap);
+ spurious:
+ ata_port_printk(ap, KERN_WARNING, "unhandled interrupt: "
+ "cmd=0x%x irq_stat=0x%x idma_stat=0x%x\n",
+ qc ? qc->tf.command : 0xff, irq_stat, idma_stat);
}
static irqreturn_t inic_interrupt(int irq, void *dev_instance)
{
struct ata_host *host = dev_instance;
- void __iomem *mmio_base = host->iomap[MMIO_BAR];
+ struct inic_host_priv *hpriv = host->private_data;
u16 host_irq_stat;
int i, handled = 0;;
- host_irq_stat = readw(mmio_base + HOST_IRQ_STAT);
+ host_irq_stat = readw(hpriv->mmio_base + HOST_IRQ_STAT);
if (unlikely(!(host_irq_stat & HIRQ_GLOBAL)))
goto out;
return IRQ_RETVAL(handled);
}
+static int inic_check_atapi_dma(struct ata_queued_cmd *qc)
+{
+ /* For some reason ATAPI_PROT_DMA doesn't work for some
+ * commands including writes and other misc ops. Use PIO
+ * protocol instead, which BTW is driven by the DMA engine
+ * anyway, so it shouldn't make much difference for native
+ * SATA devices.
+ */
+ if (atapi_cmd_type(qc->cdb[0]) == READ)
+ return 0;
+ return 1;
+}
+
+static void inic_fill_sg(struct inic_prd *prd, struct ata_queued_cmd *qc)
+{
+ struct scatterlist *sg;
+ unsigned int si;
+ u8 flags = 0;
+
+ if (qc->tf.flags & ATA_TFLAG_WRITE)
+ flags |= PRD_WRITE;
+
+ if (ata_is_dma(qc->tf.protocol))
+ flags |= PRD_DMA;
+
+ for_each_sg(qc->sg, sg, qc->n_elem, si) {
+ prd->mad = cpu_to_le32(sg_dma_address(sg));
+ prd->len = cpu_to_le16(sg_dma_len(sg));
+ prd->flags = flags;
+ prd++;
+ }
+
+ WARN_ON(!si);
+ prd[-1].flags |= PRD_END;
+}
+
+static void inic_qc_prep(struct ata_queued_cmd *qc)
+{
+ struct inic_port_priv *pp = qc->ap->private_data;
+ struct inic_pkt *pkt = pp->pkt;
+ struct inic_cpb *cpb = &pkt->cpb;
+ struct inic_prd *prd = pkt->prd;
+ bool is_atapi = ata_is_atapi(qc->tf.protocol);
+ bool is_data = ata_is_data(qc->tf.protocol);
+ unsigned int cdb_len = 0;
+
+ VPRINTK("ENTER\n");
+
+ if (is_atapi)
+ cdb_len = qc->dev->cdb_len;
+
+ /* prepare packet, based on initio driver */
+ memset(pkt, 0, sizeof(struct inic_pkt));
+
+ cpb->ctl_flags = CPB_CTL_VALID | CPB_CTL_IEN;
+ if (is_atapi || is_data)
+ cpb->ctl_flags |= CPB_CTL_DATA;
+
+ cpb->len = cpu_to_le32(qc->nbytes + cdb_len);
+ cpb->prd = cpu_to_le32(pp->pkt_dma + offsetof(struct inic_pkt, prd));
+
+ cpb->device = qc->tf.device;
+ cpb->feature = qc->tf.feature;
+ cpb->nsect = qc->tf.nsect;
+ cpb->lbal = qc->tf.lbal;
+ cpb->lbam = qc->tf.lbam;
+ cpb->lbah = qc->tf.lbah;
+
+ if (qc->tf.flags & ATA_TFLAG_LBA48) {
+ cpb->hob_feature = qc->tf.hob_feature;
+ cpb->hob_nsect = qc->tf.hob_nsect;
+ cpb->hob_lbal = qc->tf.hob_lbal;
+ cpb->hob_lbam = qc->tf.hob_lbam;
+ cpb->hob_lbah = qc->tf.hob_lbah;
+ }
+
+ cpb->command = qc->tf.command;
+ /* don't load ctl - dunno why. it's like that in the initio driver */
+
+ /* setup PRD for CDB */
+ if (is_atapi) {
+ memcpy(pkt->cdb, qc->cdb, ATAPI_CDB_LEN);
+ prd->mad = cpu_to_le32(pp->pkt_dma +
+ offsetof(struct inic_pkt, cdb));
+ prd->len = cpu_to_le16(cdb_len);
+ prd->flags = PRD_CDB | PRD_WRITE;
+ if (!is_data)
+ prd->flags |= PRD_END;
+ prd++;
+ }
+
+ /* setup sg table */
+ if (is_data)
+ inic_fill_sg(prd, qc);
+
+ pp->cpb_tbl[0] = pp->pkt_dma;
+}
+
static unsigned int inic_qc_issue(struct ata_queued_cmd *qc)
{
struct ata_port *ap = qc->ap;
+ void __iomem *port_base = inic_port_base(ap);
- /* ATA IRQ doesn't wait for DMA transfer completion and vice
- * versa. Mask IRQ selectively to detect command completion.
- * Without it, ATA DMA read command can cause data corruption.
- *
- * Something similar might be needed for ATAPI writes. I
- * tried a lot of combinations but couldn't find the solution.
- */
- if (qc->tf.protocol == ATA_PROT_DMA &&
- !(qc->tf.flags & ATA_TFLAG_WRITE))
- inic_set_pirq_mask(ap, PIRQ_MASK_DMA_READ);
- else
- inic_set_pirq_mask(ap, PIRQ_MASK_OTHER);
+ /* fire up the ADMA engine */
+ writew(HCTL_FTHD0, port_base + HOST_CTL);
+ writew(IDMA_CTL_GO, port_base + PORT_IDMA_CTL);
+ writeb(0, port_base + PORT_CPB_PTQFIFO);
+
+ return 0;
+}
+
+static void inic_tf_read(struct ata_port *ap, struct ata_taskfile *tf)
+{
+ void __iomem *port_base = inic_port_base(ap);
+
+ tf->feature = readb(port_base + PORT_TF_FEATURE);
+ tf->nsect = readb(port_base + PORT_TF_NSECT);
+ tf->lbal = readb(port_base + PORT_TF_LBAL);
+ tf->lbam = readb(port_base + PORT_TF_LBAM);
+ tf->lbah = readb(port_base + PORT_TF_LBAH);
+ tf->device = readb(port_base + PORT_TF_DEVICE);
+ tf->command = readb(port_base + PORT_TF_COMMAND);
+}
- /* Issuing a command to yet uninitialized port locks up the
- * controller. Most of the time, this happens for the first
- * command after reset which are ATA and ATAPI IDENTIFYs.
- * Fast fail if stat is 0x7f or 0xff for those commands.
+static bool inic_qc_fill_rtf(struct ata_queued_cmd *qc)
+{
+ struct ata_taskfile *rtf = &qc->result_tf;
+ struct ata_taskfile tf;
+
+ /* FIXME: Except for status and error, result TF access
+ * doesn't work. I tried reading from BAR0/2, CPB and BAR5.
+ * None works regardless of which command interface is used.
+ * For now return true iff status indicates device error.
+ * This means that we're reporting bogus sector for RW
+ * failures. Eeekk....
*/
- if (unlikely(qc->tf.command == ATA_CMD_ID_ATA ||
- qc->tf.command == ATA_CMD_ID_ATAPI)) {
- u8 stat = ap->ops->sff_check_status(ap);
- if (stat == 0x7f || stat == 0xff)
- return AC_ERR_HSM;
- }
+ inic_tf_read(qc->ap, &tf);
- return ata_sff_qc_issue(qc);
+ if (!(tf.command & ATA_ERR))
+ return false;
+
+ rtf->command = tf.command;
+ rtf->feature = tf.feature;
+ return true;
}
static void inic_freeze(struct ata_port *ap)
{
void __iomem *port_base = inic_port_base(ap);
- __inic_set_pirq_mask(ap, PIRQ_MASK_FREEZE);
-
- ap->ops->sff_check_status(ap);
+ writeb(PIRQ_MASK_FREEZE, port_base + PORT_IRQ_MASK);
writeb(0xff, port_base + PORT_IRQ_STAT);
-
- readb(port_base + PORT_IRQ_STAT); /* flush */
}
static void inic_thaw(struct ata_port *ap)
{
void __iomem *port_base = inic_port_base(ap);
- ap->ops->sff_check_status(ap);
writeb(0xff, port_base + PORT_IRQ_STAT);
+ writeb(PIRQ_MASK_DEFAULT, port_base + PORT_IRQ_MASK);
+}
- __inic_set_pirq_mask(ap, PIRQ_MASK_OTHER);
+static int inic_check_ready(struct ata_link *link)
+{
+ void __iomem *port_base = inic_port_base(link->ap);
- readb(port_base + PORT_IRQ_STAT); /* flush */
+ return ata_check_ready(readb(port_base + PORT_TF_COMMAND));
}
/*
void __iomem *port_base = inic_port_base(ap);
void __iomem *idma_ctl = port_base + PORT_IDMA_CTL;
const unsigned long *timing = sata_ehc_deb_timing(&link->eh_context);
- u16 val;
int rc;
/* hammer it into sane state */
inic_reset_port(port_base);
- val = readw(idma_ctl);
- writew(val | IDMA_CTL_RST_ATA, idma_ctl);
+ writew(IDMA_CTL_RST_ATA, idma_ctl);
readw(idma_ctl); /* flush */
msleep(1);
- writew(val & ~IDMA_CTL_RST_ATA, idma_ctl);
+ writew(0, idma_ctl);
rc = sata_link_resume(link, timing, deadline);
if (rc) {
struct ata_taskfile tf;
/* wait for link to become ready */
- rc = ata_sff_wait_after_reset(link, 1, deadline);
+ rc = ata_wait_after_reset(link, deadline, inic_check_ready);
/* link occupied, -ENODEV too is an error */
if (rc) {
ata_link_printk(link, KERN_WARNING, "device not ready "
return rc;
}
- ata_sff_tf_read(ap, &tf);
+ inic_tf_read(ap, &tf);
*class = ata_dev_classify(&tf);
}
static void inic_error_handler(struct ata_port *ap)
{
void __iomem *port_base = inic_port_base(ap);
- struct inic_port_priv *pp = ap->private_data;
- unsigned long flags;
- /* reset PIO HSM and stop DMA engine */
inic_reset_port(port_base);
-
- spin_lock_irqsave(ap->lock, flags);
- ap->hsm_task_state = HSM_ST_IDLE;
- writeb(pp->dfl_prdctl, port_base + PORT_PRD_CTL);
- spin_unlock_irqrestore(ap->lock, flags);
-
- /* PIO and DMA engines have been stopped, perform recovery */
ata_std_error_handler(ap);
}
inic_reset_port(inic_port_base(qc->ap));
}
-static void inic_dev_config(struct ata_device *dev)
-{
- /* inic can only handle upto LBA28 max sectors */
- if (dev->max_sectors > ATA_MAX_SECTORS)
- dev->max_sectors = ATA_MAX_SECTORS;
-
- if (dev->n_sectors >= 1 << 28) {
- ata_dev_printk(dev, KERN_ERR,
- "ERROR: This driver doesn't support LBA48 yet and may cause\n"
- " data corruption on such devices. Disabling.\n");
- ata_dev_disable(dev);
- }
-}
-
static void init_port(struct ata_port *ap)
{
void __iomem *port_base = inic_port_base(ap);
+ struct inic_port_priv *pp = ap->private_data;
- /* Setup PRD address */
+ /* clear packet and CPB table */
+ memset(pp->pkt, 0, sizeof(struct inic_pkt));
+ memset(pp->cpb_tbl, 0, IDMA_CPB_TBL_SIZE);
+
+ /* setup PRD and CPB lookup table addresses */
writel(ap->prd_dma, port_base + PORT_PRD_ADDR);
+ writel(pp->cpb_tbl_dma, port_base + PORT_CPB_CPBLAR);
}
static int inic_port_resume(struct ata_port *ap)
static int inic_port_start(struct ata_port *ap)
{
- void __iomem *port_base = inic_port_base(ap);
+ struct device *dev = ap->host->dev;
struct inic_port_priv *pp;
- u8 tmp;
int rc;
/* alloc and initialize private data */
- pp = devm_kzalloc(ap->host->dev, sizeof(*pp), GFP_KERNEL);
+ pp = devm_kzalloc(dev, sizeof(*pp), GFP_KERNEL);
if (!pp)
return -ENOMEM;
ap->private_data = pp;
- /* default PRD_CTL value, DMAEN, WR and START off */
- tmp = readb(port_base + PORT_PRD_CTL);
- tmp &= ~(PRD_CTL_DMAEN | PRD_CTL_WR | PRD_CTL_START);
- pp->dfl_prdctl = tmp;
-
/* Alloc resources */
rc = ata_port_start(ap);
- if (rc) {
- kfree(pp);
+ if (rc)
return rc;
- }
+
+ pp->pkt = dmam_alloc_coherent(dev, sizeof(struct inic_pkt),
+ &pp->pkt_dma, GFP_KERNEL);
+ if (!pp->pkt)
+ return -ENOMEM;
+
+ pp->cpb_tbl = dmam_alloc_coherent(dev, IDMA_CPB_TBL_SIZE,
+ &pp->cpb_tbl_dma, GFP_KERNEL);
+ if (!pp->cpb_tbl)
+ return -ENOMEM;
init_port(ap);
}
static struct ata_port_operations inic_port_ops = {
- .inherits = &ata_sff_port_ops,
+ .inherits = &sata_port_ops,
- .bmdma_setup = inic_bmdma_setup,
- .bmdma_start = inic_bmdma_start,
- .bmdma_stop = inic_bmdma_stop,
- .bmdma_status = inic_bmdma_status,
+ .check_atapi_dma = inic_check_atapi_dma,
+ .qc_prep = inic_qc_prep,
.qc_issue = inic_qc_issue,
+ .qc_fill_rtf = inic_qc_fill_rtf,
.freeze = inic_freeze,
.thaw = inic_thaw,
- .softreset = ATA_OP_NULL, /* softreset is broken */
.hardreset = inic_hardreset,
.error_handler = inic_error_handler,
.post_internal_cmd = inic_post_internal_cmd,
- .dev_config = inic_dev_config,
.scr_read = inic_scr_read,
.scr_write = inic_scr_write,
};
static struct ata_port_info inic_port_info = {
- /* For some reason, ATAPI_PROT_PIO is broken on this
- * controller, and no, PIO_POLLING does't fix it. It somehow
- * manages to report the wrong ireason and ignoring ireason
- * results in machine lock up. Tell libata to always prefer
- * DMA.
- */
.flags = ATA_FLAG_SATA | ATA_FLAG_PIO_DMA,
.pio_mask = 0x1f, /* pio0-4 */
.mwdma_mask = 0x07, /* mwdma0-2 */
{
struct ata_host *host = dev_get_drvdata(&pdev->dev);
struct inic_host_priv *hpriv = host->private_data;
- void __iomem *mmio_base = host->iomap[MMIO_BAR];
int rc;
rc = ata_pci_device_do_resume(pdev);
return rc;
if (pdev->dev.power.power_state.event == PM_EVENT_SUSPEND) {
- rc = init_controller(mmio_base, hpriv->cached_hctl);
+ rc = init_controller(hpriv->mmio_base, hpriv->cached_hctl);
if (rc)
return rc;
}
struct ata_host *host;
struct inic_host_priv *hpriv;
void __iomem * const *iomap;
+ int mmio_bar;
int i, rc;
if (!printed_version++)
host->private_data = hpriv;
- /* acquire resources and fill host */
+ /* Acquire resources and fill host. Note that PCI and cardbus
+ * use different BARs.
+ */
rc = pcim_enable_device(pdev);
if (rc)
return rc;
- rc = pcim_iomap_regions(pdev, 0x3f, DRV_NAME);
+ if (pci_resource_flags(pdev, MMIO_BAR_PCI) & IORESOURCE_MEM)
+ mmio_bar = MMIO_BAR_PCI;
+ else
+ mmio_bar = MMIO_BAR_CARDBUS;
+
+ rc = pcim_iomap_regions(pdev, 1 << mmio_bar, DRV_NAME);
if (rc)
return rc;
host->iomap = iomap = pcim_iomap_table(pdev);
+ hpriv->mmio_base = iomap[mmio_bar];
+ hpriv->cached_hctl = readw(hpriv->mmio_base + HOST_CTL);
for (i = 0; i < NR_PORTS; i++) {
struct ata_port *ap = host->ports[i];
- struct ata_ioports *port = &ap->ioaddr;
- unsigned int offset = i * PORT_SIZE;
-
- port->cmd_addr = iomap[2 * i];
- port->altstatus_addr =
- port->ctl_addr = (void __iomem *)
- ((unsigned long)iomap[2 * i + 1] | ATA_PCI_CTL_OFS);
- port->scr_addr = iomap[MMIO_BAR] + offset + PORT_SCR;
-
- ata_sff_std_ports(port);
-
- ata_port_pbar_desc(ap, MMIO_BAR, -1, "mmio");
- ata_port_pbar_desc(ap, MMIO_BAR, offset, "port");
- ata_port_desc(ap, "cmd 0x%llx ctl 0x%llx",
- (unsigned long long)pci_resource_start(pdev, 2 * i),
- (unsigned long long)pci_resource_start(pdev, (2 * i + 1)) |
- ATA_PCI_CTL_OFS);
- }
- hpriv->cached_hctl = readw(iomap[MMIO_BAR] + HOST_CTL);
+ ata_port_pbar_desc(ap, mmio_bar, -1, "mmio");
+ ata_port_pbar_desc(ap, mmio_bar, i * PORT_SIZE, "port");
+ }
/* Set dma_mask. This devices doesn't support 64bit addressing. */
rc = pci_set_dma_mask(pdev, DMA_32BIT_MASK);
return rc;
}
- rc = init_controller(iomap[MMIO_BAR], hpriv->cached_hctl);
+ rc = init_controller(hpriv->mmio_base, hpriv->cached_hctl);
if (rc) {
dev_printk(KERN_ERR, &pdev->dev,
"failed to initialize controller\n");
#include <linux/platform_device.h>
#include <linux/ata_platform.h>
#include <linux/mbus.h>
+#include <linux/bitops.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_cmnd.h>
#include <scsi/scsi_device.h>
MV_IRQ_COAL_TIME_THRESHOLD = (MV_IRQ_COAL_REG_BASE + 0xd0),
MV_SATAHC0_REG_BASE = 0x20000,
- MV_FLASH_CTL = 0x1046c,
- MV_GPIO_PORT_CTL = 0x104f0,
- MV_RESET_CFG = 0x180d8,
+ MV_FLASH_CTL_OFS = 0x1046c,
+ MV_GPIO_PORT_CTL_OFS = 0x104f0,
+ MV_RESET_CFG_OFS = 0x180d8,
MV_PCI_REG_SZ = MV_MAJOR_REG_AREA_SZ,
MV_SATAHC_REG_SZ = MV_MAJOR_REG_AREA_SZ,
/* PCI interface registers */
PCI_COMMAND_OFS = 0xc00,
+ PCI_COMMAND_MRDTRIG = (1 << 7), /* PCI Master Read Trigger */
PCI_MAIN_CMD_STS_OFS = 0xd30,
STOP_PCI_MASTER = (1 << 2),
PCI_MASTER_EMPTY = (1 << 3),
GLOB_SFT_RST = (1 << 4),
- MV_PCI_MODE = 0xd00,
+ MV_PCI_MODE_OFS = 0xd00,
+ MV_PCI_MODE_MASK = 0x30,
+
MV_PCI_EXP_ROM_BAR_CTL = 0xd2c,
MV_PCI_DISC_TIMER = 0xd04,
MV_PCI_MSI_TRIGGER = 0xc38,
MV_PCI_SERR_MASK = 0xc28,
- MV_PCI_XBAR_TMOUT = 0x1d04,
+ MV_PCI_XBAR_TMOUT_OFS = 0x1d04,
MV_PCI_ERR_LOW_ADDRESS = 0x1d40,
MV_PCI_ERR_HIGH_ADDRESS = 0x1d44,
MV_PCI_ERR_ATTRIBUTE = 0x1d48,
PHY_MODE4 = 0x314,
PHY_MODE2 = 0x330,
SATA_IFCTL_OFS = 0x344,
+ SATA_TESTCTL_OFS = 0x348,
SATA_IFSTAT_OFS = 0x34c,
VENDOR_UNIQUE_FIS_OFS = 0x35c,
- FIS_CFG_OFS = 0x360,
- FIS_CFG_SINGLE_SYNC = (1 << 16), /* SYNC on DMA activation */
+ FISCFG_OFS = 0x360,
+ FISCFG_WAIT_DEV_ERR = (1 << 8), /* wait for host on DevErr */
+ FISCFG_SINGLE_SYNC = (1 << 16), /* SYNC on DMA activation */
MV5_PHY_MODE = 0x74,
- MV5_LT_MODE = 0x30,
- MV5_PHY_CTL = 0x0C,
- SATA_INTERFACE_CFG = 0x050,
+ MV5_LTMODE_OFS = 0x30,
+ MV5_PHY_CTL_OFS = 0x0C,
+ SATA_INTERFACE_CFG_OFS = 0x050,
MV_M2_PREAMP_MASK = 0x7e0,
EDMA_CMD_OFS = 0x28, /* EDMA command register */
EDMA_EN = (1 << 0), /* enable EDMA */
EDMA_DS = (1 << 1), /* disable EDMA; self-negated */
- ATA_RST = (1 << 2), /* reset trans/link/phy */
+ EDMA_RESET = (1 << 2), /* reset eng/trans/link/phy */
+
+ EDMA_STATUS_OFS = 0x30, /* EDMA engine status */
+ EDMA_STATUS_CACHE_EMPTY = (1 << 6), /* GenIIe command cache empty */
+ EDMA_STATUS_IDLE = (1 << 7), /* GenIIe EDMA enabled/idle */
- EDMA_IORDY_TMOUT = 0x34,
- EDMA_ARB_CFG = 0x38,
+ EDMA_IORDY_TMOUT_OFS = 0x34,
+ EDMA_ARB_CFG_OFS = 0x38,
+
+ EDMA_HALTCOND_OFS = 0x60, /* GenIIe halt conditions */
GEN_II_NCQ_MAX_SECTORS = 256, /* max sects/io on Gen2 w/NCQ */
MV_HP_GEN_II = (1 << 7), /* Generation II: 60xx */
MV_HP_GEN_IIE = (1 << 8), /* Generation IIE: 6042/7042 */
MV_HP_PCIE = (1 << 9), /* PCIe bus/regs: 7042 */
+ MV_HP_CUT_THROUGH = (1 << 10), /* can use EDMA cut-through */
/* Port private flags (pp_flags) */
MV_PP_FLAG_EDMA_EN = (1 << 0), /* is EDMA engine enabled? */
MV_PP_FLAG_NCQ_EN = (1 << 1), /* is EDMA set up for NCQ? */
+ MV_PP_FLAG_FBS_EN = (1 << 2), /* is EDMA set up for FBS? */
+ MV_PP_FLAG_DELAYED_EH = (1 << 3), /* delayed dev err handling */
};
#define IS_GEN_I(hpriv) ((hpriv)->hp_flags & MV_HP_GEN_I)
#define IS_GEN_II(hpriv) ((hpriv)->hp_flags & MV_HP_GEN_II)
#define IS_GEN_IIE(hpriv) ((hpriv)->hp_flags & MV_HP_GEN_IIE)
+#define IS_PCIE(hpriv) ((hpriv)->hp_flags & MV_HP_PCIE)
#define HAS_PCI(host) (!((host)->ports[0]->flags & MV_FLAG_SOC))
#define WINDOW_CTRL(i) (0x20030 + ((i) << 4))
unsigned int resp_idx;
u32 pp_flags;
+ unsigned int delayed_eh_pmp_map;
};
struct mv_port_signal {
static int mv5_scr_write(struct ata_port *ap, unsigned int sc_reg_in, u32 val);
static int mv_port_start(struct ata_port *ap);
static void mv_port_stop(struct ata_port *ap);
+static int mv_qc_defer(struct ata_queued_cmd *qc);
static void mv_qc_prep(struct ata_queued_cmd *qc);
static void mv_qc_prep_iie(struct ata_queued_cmd *qc);
static unsigned int mv_qc_issue(struct ata_queued_cmd *qc);
unsigned long deadline);
static int mv_softreset(struct ata_link *link, unsigned int *class,
unsigned long deadline);
+static void mv_pmp_error_handler(struct ata_port *ap);
+static void mv_process_crpb_entries(struct ata_port *ap,
+ struct mv_port_priv *pp);
/* .sg_tablesize is (MV_MAX_SG_CT / 2) in the structures below
* because we have to allow room for worst case splitting of
static struct ata_port_operations mv5_ops = {
.inherits = &ata_sff_port_ops,
+ .qc_defer = mv_qc_defer,
.qc_prep = mv_qc_prep,
.qc_issue = mv_qc_issue,
static struct ata_port_operations mv6_ops = {
.inherits = &mv5_ops,
- .qc_defer = sata_pmp_qc_defer_cmd_switch,
.dev_config = mv6_dev_config,
.scr_read = mv_scr_read,
.scr_write = mv_scr_write,
.pmp_hardreset = mv_pmp_hardreset,
.pmp_softreset = mv_softreset,
.softreset = mv_softreset,
- .error_handler = sata_pmp_error_handler,
+ .error_handler = mv_pmp_error_handler,
};
static struct ata_port_operations mv_iie_ops = {
.inherits = &mv6_ops,
- .qc_defer = ata_std_qc_defer, /* FIS-based switching */
.dev_config = ATA_OP_NULL,
.qc_prep = mv_qc_prep_iie,
};
}
}
+static void mv_wait_for_edma_empty_idle(struct ata_port *ap)
+{
+ void __iomem *port_mmio = mv_ap_base(ap);
+ const u32 empty_idle = (EDMA_STATUS_CACHE_EMPTY | EDMA_STATUS_IDLE);
+ const int per_loop = 5, timeout = (15 * 1000 / per_loop);
+ int i;
+
+ /*
+ * Wait for the EDMA engine to finish transactions in progress.
+ * No idea what a good "timeout" value might be, but measurements
+ * indicate that it often requires hundreds of microseconds
+ * with two drives in-use. So we use the 15msec value above
+ * as a rough guess at what even more drives might require.
+ */
+ for (i = 0; i < timeout; ++i) {
+ u32 edma_stat = readl(port_mmio + EDMA_STATUS_OFS);
+ if ((edma_stat & empty_idle) == empty_idle)
+ break;
+ udelay(per_loop);
+ }
+ /* ata_port_printk(ap, KERN_INFO, "%s: %u+ usecs\n", __func__, i); */
+}
+
/**
* mv_stop_edma_engine - Disable eDMA engine
* @port_mmio: io base address
if (!(pp->pp_flags & MV_PP_FLAG_EDMA_EN))
return 0;
pp->pp_flags &= ~MV_PP_FLAG_EDMA_EN;
+ mv_wait_for_edma_empty_idle(ap);
if (mv_stop_edma_engine(port_mmio)) {
ata_port_printk(ap, KERN_ERR, "Unable to stop eDMA\n");
return -EIO;
}
}
-static void mv_config_fbs(void __iomem *port_mmio, int enable_fbs)
+static int mv_qc_defer(struct ata_queued_cmd *qc)
{
- u32 old_fcfg, new_fcfg, old_ltmode, new_ltmode;
+ struct ata_link *link = qc->dev->link;
+ struct ata_port *ap = link->ap;
+ struct mv_port_priv *pp = ap->private_data;
+
+ /*
+ * Don't allow new commands if we're in a delayed EH state
+ * for NCQ and/or FIS-based switching.
+ */
+ if (pp->pp_flags & MV_PP_FLAG_DELAYED_EH)
+ return ATA_DEFER_PORT;
/*
- * Various bit settings required for operation
- * in FIS-based switching (fbs) mode on GenIIe:
+ * If the port is completely idle, then allow the new qc.
*/
- old_fcfg = readl(port_mmio + FIS_CFG_OFS);
- old_ltmode = readl(port_mmio + LTMODE_OFS);
- if (enable_fbs) {
- new_fcfg = old_fcfg | FIS_CFG_SINGLE_SYNC;
- new_ltmode = old_ltmode | LTMODE_BIT8;
- } else { /* disable fbs */
- new_fcfg = old_fcfg & ~FIS_CFG_SINGLE_SYNC;
- new_ltmode = old_ltmode & ~LTMODE_BIT8;
- }
- if (new_fcfg != old_fcfg)
- writelfl(new_fcfg, port_mmio + FIS_CFG_OFS);
+ if (ap->nr_active_links == 0)
+ return 0;
+
+ if (pp->pp_flags & MV_PP_FLAG_EDMA_EN) {
+ /*
+ * The port is operating in host queuing mode (EDMA).
+ * It can accomodate a new qc if the qc protocol
+ * is compatible with the current host queue mode.
+ */
+ if (pp->pp_flags & MV_PP_FLAG_NCQ_EN) {
+ /*
+ * The host queue (EDMA) is in NCQ mode.
+ * If the new qc is also an NCQ command,
+ * then allow the new qc.
+ */
+ if (qc->tf.protocol == ATA_PROT_NCQ)
+ return 0;
+ } else {
+ /*
+ * The host queue (EDMA) is in non-NCQ, DMA mode.
+ * If the new qc is also a non-NCQ, DMA command,
+ * then allow the new qc.
+ */
+ if (qc->tf.protocol == ATA_PROT_DMA)
+ return 0;
+ }
+ }
+ return ATA_DEFER_PORT;
+}
+
+static void mv_config_fbs(void __iomem *port_mmio, int want_ncq, int want_fbs)
+{
+ u32 new_fiscfg, old_fiscfg;
+ u32 new_ltmode, old_ltmode;
+ u32 new_haltcond, old_haltcond;
+
+ old_fiscfg = readl(port_mmio + FISCFG_OFS);
+ old_ltmode = readl(port_mmio + LTMODE_OFS);
+ old_haltcond = readl(port_mmio + EDMA_HALTCOND_OFS);
+
+ new_fiscfg = old_fiscfg & ~(FISCFG_SINGLE_SYNC | FISCFG_WAIT_DEV_ERR);
+ new_ltmode = old_ltmode & ~LTMODE_BIT8;
+ new_haltcond = old_haltcond | EDMA_ERR_DEV;
+
+ if (want_fbs) {
+ new_fiscfg = old_fiscfg | FISCFG_SINGLE_SYNC;
+ new_ltmode = old_ltmode | LTMODE_BIT8;
+ if (want_ncq)
+ new_haltcond &= ~EDMA_ERR_DEV;
+ else
+ new_fiscfg |= FISCFG_WAIT_DEV_ERR;
+ }
+
+ if (new_fiscfg != old_fiscfg)
+ writelfl(new_fiscfg, port_mmio + FISCFG_OFS);
if (new_ltmode != old_ltmode)
writelfl(new_ltmode, port_mmio + LTMODE_OFS);
+ if (new_haltcond != old_haltcond)
+ writelfl(new_haltcond, port_mmio + EDMA_HALTCOND_OFS);
+}
+
+static void mv_60x1_errata_sata25(struct ata_port *ap, int want_ncq)
+{
+ struct mv_host_priv *hpriv = ap->host->private_data;
+ u32 old, new;
+
+ /* workaround for 88SX60x1 FEr SATA#25 (part 1) */
+ old = readl(hpriv->base + MV_GPIO_PORT_CTL_OFS);
+ if (want_ncq)
+ new = old | (1 << 22);
+ else
+ new = old & ~(1 << 22);
+ if (new != old)
+ writel(new, hpriv->base + MV_GPIO_PORT_CTL_OFS);
}
static void mv_edma_cfg(struct ata_port *ap, int want_ncq)
/* set up non-NCQ EDMA configuration */
cfg = EDMA_CFG_Q_DEPTH; /* always 0x1f for *all* chips */
+ pp->pp_flags &= ~MV_PP_FLAG_FBS_EN;
if (IS_GEN_I(hpriv))
cfg |= (1 << 8); /* enab config burst size mask */
- else if (IS_GEN_II(hpriv))
+ else if (IS_GEN_II(hpriv)) {
cfg |= EDMA_CFG_RD_BRST_EXT | EDMA_CFG_WR_BUFF_LEN;
+ mv_60x1_errata_sata25(ap, want_ncq);
- else if (IS_GEN_IIE(hpriv)) {
- cfg |= (1 << 23); /* do not mask PM field in rx'd FIS */
- cfg |= (1 << 22); /* enab 4-entry host queue cache */
- cfg |= (1 << 18); /* enab early completion */
- cfg |= (1 << 17); /* enab cut-through (dis stor&forwrd) */
+ } else if (IS_GEN_IIE(hpriv)) {
+ int want_fbs = sata_pmp_attached(ap);
+ /*
+ * Possible future enhancement:
+ *
+ * The chip can use FBS with non-NCQ, if we allow it,
+ * But first we need to have the error handling in place
+ * for this mode (datasheet section 7.3.15.4.2.3).
+ * So disallow non-NCQ FBS for now.
+ */
+ want_fbs &= want_ncq;
+
+ mv_config_fbs(port_mmio, want_ncq, want_fbs);
- if (want_ncq && sata_pmp_attached(ap)) {
+ if (want_fbs) {
+ pp->pp_flags |= MV_PP_FLAG_FBS_EN;
cfg |= EDMA_CFG_EDMA_FBS; /* FIS-based switching */
- mv_config_fbs(port_mmio, 1);
- } else {
- mv_config_fbs(port_mmio, 0);
}
+
+ cfg |= (1 << 23); /* do not mask PM field in rx'd FIS */
+ cfg |= (1 << 22); /* enab 4-entry host queue cache */
+ if (HAS_PCI(ap->host))
+ cfg |= (1 << 18); /* enab early completion */
+ if (hpriv->hp_flags & MV_HP_CUT_THROUGH)
+ cfg |= (1 << 17); /* enab cut-thru (dis stor&forwrd) */
}
if (want_ncq) {
return qc;
}
-static void mv_unexpected_intr(struct ata_port *ap)
+static void mv_pmp_error_handler(struct ata_port *ap)
{
+ unsigned int pmp, pmp_map;
struct mv_port_priv *pp = ap->private_data;
- struct ata_eh_info *ehi = &ap->link.eh_info;
- char *when = "";
+ if (pp->pp_flags & MV_PP_FLAG_DELAYED_EH) {
+ /*
+ * Perform NCQ error analysis on failed PMPs
+ * before we freeze the port entirely.
+ *
+ * The failed PMPs are marked earlier by mv_pmp_eh_prep().
+ */
+ pmp_map = pp->delayed_eh_pmp_map;
+ pp->pp_flags &= ~MV_PP_FLAG_DELAYED_EH;
+ for (pmp = 0; pmp_map != 0; pmp++) {
+ unsigned int this_pmp = (1 << pmp);
+ if (pmp_map & this_pmp) {
+ struct ata_link *link = &ap->pmp_link[pmp];
+ pmp_map &= ~this_pmp;
+ ata_eh_analyze_ncq_error(link);
+ }
+ }
+ ata_port_freeze(ap);
+ }
+ sata_pmp_error_handler(ap);
+}
+
+static unsigned int mv_get_err_pmp_map(struct ata_port *ap)
+{
+ void __iomem *port_mmio = mv_ap_base(ap);
+
+ return readl(port_mmio + SATA_TESTCTL_OFS) >> 16;
+}
+
+static void mv_pmp_eh_prep(struct ata_port *ap, unsigned int pmp_map)
+{
+ struct ata_eh_info *ehi;
+ unsigned int pmp;
+
+ /*
+ * Initialize EH info for PMPs which saw device errors
+ */
+ ehi = &ap->link.eh_info;
+ for (pmp = 0; pmp_map != 0; pmp++) {
+ unsigned int this_pmp = (1 << pmp);
+ if (pmp_map & this_pmp) {
+ struct ata_link *link = &ap->pmp_link[pmp];
+
+ pmp_map &= ~this_pmp;
+ ehi = &link->eh_info;
+ ata_ehi_clear_desc(ehi);
+ ata_ehi_push_desc(ehi, "dev err");
+ ehi->err_mask |= AC_ERR_DEV;
+ ehi->action |= ATA_EH_RESET;
+ ata_link_abort(link);
+ }
+ }
+}
+
+static int mv_handle_fbs_ncq_dev_err(struct ata_port *ap)
+{
+ struct mv_port_priv *pp = ap->private_data;
+ int failed_links;
+ unsigned int old_map, new_map;
+
+ /*
+ * Device error during FBS+NCQ operation:
+ *
+ * Set a port flag to prevent further I/O being enqueued.
+ * Leave the EDMA running to drain outstanding commands from this port.
+ * Perform the post-mortem/EH only when all responses are complete.
+ * Follow recovery sequence from 6042/7042 datasheet (7.3.15.4.2.2).
+ */
+ if (!(pp->pp_flags & MV_PP_FLAG_DELAYED_EH)) {
+ pp->pp_flags |= MV_PP_FLAG_DELAYED_EH;
+ pp->delayed_eh_pmp_map = 0;
+ }
+ old_map = pp->delayed_eh_pmp_map;
+ new_map = old_map | mv_get_err_pmp_map(ap);
+
+ if (old_map != new_map) {
+ pp->delayed_eh_pmp_map = new_map;
+ mv_pmp_eh_prep(ap, new_map & ~old_map);
+ }
+ failed_links = hweight16(new_map);
+
+ ata_port_printk(ap, KERN_INFO, "%s: pmp_map=%04x qc_map=%04x "
+ "failed_links=%d nr_active_links=%d\n",
+ __func__, pp->delayed_eh_pmp_map,
+ ap->qc_active, failed_links,
+ ap->nr_active_links);
+
+ if (ap->nr_active_links <= failed_links) {
+ mv_process_crpb_entries(ap, pp);
+ mv_stop_edma(ap);
+ mv_eh_freeze(ap);
+ ata_port_printk(ap, KERN_INFO, "%s: done\n", __func__);
+ return 1; /* handled */
+ }
+ ata_port_printk(ap, KERN_INFO, "%s: waiting\n", __func__);
+ return 1; /* handled */
+}
+
+static int mv_handle_fbs_non_ncq_dev_err(struct ata_port *ap)
+{
/*
- * We got a device interrupt from something that
- * was supposed to be using EDMA or polling.
+ * Possible future enhancement:
+ *
+ * FBS+non-NCQ operation is not yet implemented.
+ * See related notes in mv_edma_cfg().
+ *
+ * Device error during FBS+non-NCQ operation:
+ *
+ * We need to snapshot the shadow registers for each failed command.
+ * Follow recovery sequence from 6042/7042 datasheet (7.3.15.4.2.3).
*/
+ return 0; /* not handled */
+}
+
+static int mv_handle_dev_err(struct ata_port *ap, u32 edma_err_cause)
+{
+ struct mv_port_priv *pp = ap->private_data;
+
+ if (!(pp->pp_flags & MV_PP_FLAG_EDMA_EN))
+ return 0; /* EDMA was not active: not handled */
+ if (!(pp->pp_flags & MV_PP_FLAG_FBS_EN))
+ return 0; /* FBS was not active: not handled */
+
+ if (!(edma_err_cause & EDMA_ERR_DEV))
+ return 0; /* non DEV error: not handled */
+ edma_err_cause &= ~EDMA_ERR_IRQ_TRANSIENT;
+ if (edma_err_cause & ~(EDMA_ERR_DEV | EDMA_ERR_SELF_DIS))
+ return 0; /* other problems: not handled */
+
+ if (pp->pp_flags & MV_PP_FLAG_NCQ_EN) {
+ /*
+ * EDMA should NOT have self-disabled for this case.
+ * If it did, then something is wrong elsewhere,
+ * and we cannot handle it here.
+ */
+ if (edma_err_cause & EDMA_ERR_SELF_DIS) {
+ ata_port_printk(ap, KERN_WARNING,
+ "%s: err_cause=0x%x pp_flags=0x%x\n",
+ __func__, edma_err_cause, pp->pp_flags);
+ return 0; /* not handled */
+ }
+ return mv_handle_fbs_ncq_dev_err(ap);
+ } else {
+ /*
+ * EDMA should have self-disabled for this case.
+ * If it did not, then something is wrong elsewhere,
+ * and we cannot handle it here.
+ */
+ if (!(edma_err_cause & EDMA_ERR_SELF_DIS)) {
+ ata_port_printk(ap, KERN_WARNING,
+ "%s: err_cause=0x%x pp_flags=0x%x\n",
+ __func__, edma_err_cause, pp->pp_flags);
+ return 0; /* not handled */
+ }
+ return mv_handle_fbs_non_ncq_dev_err(ap);
+ }
+ return 0; /* not handled */
+}
+
+static void mv_unexpected_intr(struct ata_port *ap, int edma_was_enabled)
+{
+ struct ata_eh_info *ehi = &ap->link.eh_info;
+ char *when = "idle";
+
ata_ehi_clear_desc(ehi);
- if (pp->pp_flags & MV_PP_FLAG_EDMA_EN) {
- when = " while EDMA enabled";
+ if (!ap || (ap->flags & ATA_FLAG_DISABLED)) {
+ when = "disabled";
+ } else if (edma_was_enabled) {
+ when = "EDMA enabled";
} else {
struct ata_queued_cmd *qc = ata_qc_from_tag(ap, ap->link.active_tag);
if (qc && (qc->tf.flags & ATA_TFLAG_POLLING))
- when = " while polling";
+ when = "polling";
}
- ata_ehi_push_desc(ehi, "unexpected device interrupt%s", when);
+ ata_ehi_push_desc(ehi, "unexpected device interrupt while %s", when);
ehi->err_mask |= AC_ERR_OTHER;
ehi->action |= ATA_EH_RESET;
ata_port_freeze(ap);
* LOCKING:
* Inherited from caller.
*/
-static void mv_err_intr(struct ata_port *ap, struct ata_queued_cmd *qc)
+static void mv_err_intr(struct ata_port *ap)
{
void __iomem *port_mmio = mv_ap_base(ap);
u32 edma_err_cause, eh_freeze_mask, serr = 0;
struct mv_host_priv *hpriv = ap->host->private_data;
unsigned int action = 0, err_mask = 0;
struct ata_eh_info *ehi = &ap->link.eh_info;
-
- ata_ehi_clear_desc(ehi);
+ struct ata_queued_cmd *qc;
+ int abort = 0;
/*
- * Read and clear the err_cause bits. This won't actually
- * clear for some errors (eg. SError), but we will be doing
- * a hard reset in those cases regardless, which *will* clear it.
+ * Read and clear the SError and err_cause bits.
*/
+ sata_scr_read(&ap->link, SCR_ERROR, &serr);
+ sata_scr_write_flush(&ap->link, SCR_ERROR, serr);
+
edma_err_cause = readl(port_mmio + EDMA_ERR_IRQ_CAUSE_OFS);
writelfl(~edma_err_cause, port_mmio + EDMA_ERR_IRQ_CAUSE_OFS);
- ata_ehi_push_desc(ehi, "edma_err_cause=%08x", edma_err_cause);
+ ata_port_printk(ap, KERN_INFO, "%s: err_cause=%08x pp_flags=0x%x\n",
+ __func__, edma_err_cause, pp->pp_flags);
+
+ if (edma_err_cause & EDMA_ERR_DEV) {
+ /*
+ * Device errors during FIS-based switching operation
+ * require special handling.
+ */
+ if (mv_handle_dev_err(ap, edma_err_cause))
+ return;
+ }
+ qc = mv_get_active_qc(ap);
+ ata_ehi_clear_desc(ehi);
+ ata_ehi_push_desc(ehi, "edma_err_cause=%08x pp_flags=%08x",
+ edma_err_cause, pp->pp_flags);
/*
* All generations share these EDMA error cause bits:
*/
- if (edma_err_cause & EDMA_ERR_DEV)
+ if (edma_err_cause & EDMA_ERR_DEV) {
err_mask |= AC_ERR_DEV;
+ action |= ATA_EH_RESET;
+ ata_ehi_push_desc(ehi, "dev error");
+ }
if (edma_err_cause & (EDMA_ERR_D_PAR | EDMA_ERR_PRD_PAR |
EDMA_ERR_CRQB_PAR | EDMA_ERR_CRPB_PAR |
EDMA_ERR_INTRL_PAR)) {
ata_ehi_push_desc(ehi, "EDMA self-disable");
}
if (edma_err_cause & EDMA_ERR_SERR) {
- /*
- * Ensure that we read our own SCR, not a pmp link SCR:
- */
- ap->ops->scr_read(ap, SCR_ERROR, &serr);
- /*
- * Don't clear SError here; leave it for libata-eh:
- */
ata_ehi_push_desc(ehi, "SError=%08x", serr);
err_mask |= AC_ERR_ATA_BUS;
action |= ATA_EH_RESET;
else
ehi->err_mask |= err_mask;
- if (edma_err_cause & eh_freeze_mask)
+ if (err_mask == AC_ERR_DEV) {
+ /*
+ * Cannot do ata_port_freeze() here,
+ * because it would kill PIO access,
+ * which is needed for further diagnosis.
+ */
+ mv_eh_freeze(ap);
+ abort = 1;
+ } else if (edma_err_cause & eh_freeze_mask) {
+ /*
+ * Note to self: ata_port_freeze() calls ata_port_abort()
+ */
ata_port_freeze(ap);
- else
- ata_port_abort(ap);
+ } else {
+ abort = 1;
+ }
+
+ if (abort) {
+ if (qc)
+ ata_link_abort(qc->dev->link);
+ else
+ ata_port_abort(ap);
+ }
}
static void mv_process_crpb_response(struct ata_port *ap,
}
}
ata_status = edma_status >> CRPB_FLAG_STATUS_SHIFT;
- qc->err_mask |= ac_err_mask(ata_status);
- ata_qc_complete(qc);
+ if (!ac_err_mask(ata_status))
+ ata_qc_complete(qc);
+ /* else: leave it for mv_err_intr() */
} else {
ata_port_printk(ap, KERN_ERR, "%s: no qc for tag=%d\n",
__func__, tag);
port_mmio + EDMA_RSP_Q_OUT_PTR_OFS);
}
+static void mv_port_intr(struct ata_port *ap, u32 port_cause)
+{
+ struct mv_port_priv *pp;
+ int edma_was_enabled;
+
+ if (!ap || (ap->flags & ATA_FLAG_DISABLED)) {
+ mv_unexpected_intr(ap, 0);
+ return;
+ }
+ /*
+ * Grab a snapshot of the EDMA_EN flag setting,
+ * so that we have a consistent view for this port,
+ * even if something we call of our routines changes it.
+ */
+ pp = ap->private_data;
+ edma_was_enabled = (pp->pp_flags & MV_PP_FLAG_EDMA_EN);
+ /*
+ * Process completed CRPB response(s) before other events.
+ */
+ if (edma_was_enabled && (port_cause & DONE_IRQ)) {
+ mv_process_crpb_entries(ap, pp);
+ if (pp->pp_flags & MV_PP_FLAG_DELAYED_EH)
+ mv_handle_fbs_ncq_dev_err(ap);
+ }
+ /*
+ * Handle chip-reported errors, or continue on to handle PIO.
+ */
+ if (unlikely(port_cause & ERR_IRQ)) {
+ mv_err_intr(ap);
+ } else if (!edma_was_enabled) {
+ struct ata_queued_cmd *qc = mv_get_active_qc(ap);
+ if (qc)
+ ata_sff_host_intr(ap, qc);
+ else
+ mv_unexpected_intr(ap, edma_was_enabled);
+ }
+}
+
/**
* mv_host_intr - Handle all interrupts on the given host controller
* @host: host specific structure
static int mv_host_intr(struct ata_host *host, u32 main_irq_cause)
{
struct mv_host_priv *hpriv = host->private_data;
- void __iomem *mmio = hpriv->base, *hc_mmio = NULL;
- u32 hc_irq_cause = 0;
+ void __iomem *mmio = hpriv->base, *hc_mmio;
unsigned int handled = 0, port;
for (port = 0; port < hpriv->n_ports; port++) {
struct ata_port *ap = host->ports[port];
- struct mv_port_priv *pp;
- unsigned int shift, hardport, port_cause;
- /*
- * When we move to the second hc, flag our cached
- * copies of hc_mmio (and hc_irq_cause) as invalid again.
- */
- if (port == MV_PORTS_PER_HC)
- hc_mmio = NULL;
- /*
- * Do nothing if port is not interrupting or is disabled:
- */
+ unsigned int p, shift, hardport, port_cause;
+
MV_PORT_TO_SHIFT_AND_HARDPORT(port, shift, hardport);
- port_cause = (main_irq_cause >> shift) & (DONE_IRQ | ERR_IRQ);
- if (!port_cause || !ap || (ap->flags & ATA_FLAG_DISABLED))
- continue;
/*
- * Each hc within the host has its own hc_irq_cause register.
- * We defer reading it until we know we need it, right now:
- *
- * FIXME later: we don't really need to read this register
- * (some logic changes required below if we go that way),
- * because it doesn't tell us anything new. But we do need
- * to write to it, outside the top of this loop,
- * to reset the interrupt triggers for next time.
+ * Each hc within the host has its own hc_irq_cause register,
+ * where the interrupting ports bits get ack'd.
*/
- if (!hc_mmio) {
+ if (hardport == 0) { /* first port on this hc ? */
+ u32 hc_cause = (main_irq_cause >> shift) & HC0_IRQ_PEND;
+ u32 port_mask, ack_irqs;
+ /*
+ * Skip this entire hc if nothing pending for any ports
+ */
+ if (!hc_cause) {
+ port += MV_PORTS_PER_HC - 1;
+ continue;
+ }
+ /*
+ * We don't need/want to read the hc_irq_cause register,
+ * because doing so hurts performance, and
+ * main_irq_cause already gives us everything we need.
+ *
+ * But we do have to *write* to the hc_irq_cause to ack
+ * the ports that we are handling this time through.
+ *
+ * This requires that we create a bitmap for those
+ * ports which interrupted us, and use that bitmap
+ * to ack (only) those ports via hc_irq_cause.
+ */
+ ack_irqs = 0;
+ for (p = 0; p < MV_PORTS_PER_HC; ++p) {
+ if ((port + p) >= hpriv->n_ports)
+ break;
+ port_mask = (DONE_IRQ | ERR_IRQ) << (p * 2);
+ if (hc_cause & port_mask)
+ ack_irqs |= (DMA_IRQ | DEV_IRQ) << p;
+ }
hc_mmio = mv_hc_base_from_port(mmio, port);
- hc_irq_cause = readl(hc_mmio + HC_IRQ_CAUSE_OFS);
- writelfl(~hc_irq_cause, hc_mmio + HC_IRQ_CAUSE_OFS);
+ writelfl(~ack_irqs, hc_mmio + HC_IRQ_CAUSE_OFS);
handled = 1;
}
/*
- * Process completed CRPB response(s) before other events.
- */
- pp = ap->private_data;
- if (hc_irq_cause & (DMA_IRQ << hardport)) {
- if (pp->pp_flags & MV_PP_FLAG_EDMA_EN)
- mv_process_crpb_entries(ap, pp);
- }
- /*
- * Handle chip-reported errors, or continue on to handle PIO.
+ * Handle interrupts signalled for this port:
*/
- if (unlikely(port_cause & ERR_IRQ)) {
- mv_err_intr(ap, mv_get_active_qc(ap));
- } else if (hc_irq_cause & (DEV_IRQ << hardport)) {
- if (!(pp->pp_flags & MV_PP_FLAG_EDMA_EN)) {
- struct ata_queued_cmd *qc = mv_get_active_qc(ap);
- if (qc) {
- ata_sff_host_intr(ap, qc);
- continue;
- }
- }
- mv_unexpected_intr(ap);
- }
+ port_cause = (main_irq_cause >> shift) & (DONE_IRQ | ERR_IRQ);
+ if (port_cause)
+ mv_port_intr(ap, port_cause);
}
return handled;
}
static void mv5_reset_flash(struct mv_host_priv *hpriv, void __iomem *mmio)
{
- writel(0x0fcfffff, mmio + MV_FLASH_CTL);
+ writel(0x0fcfffff, mmio + MV_FLASH_CTL_OFS);
}
static void mv5_read_preamp(struct mv_host_priv *hpriv, int idx,
{
u32 tmp;
- writel(0, mmio + MV_GPIO_PORT_CTL);
+ writel(0, mmio + MV_GPIO_PORT_CTL_OFS);
/* FIXME: handle MV_HP_ERRATA_50XXB2 errata */
int fix_apm_sq = (hpriv->hp_flags & MV_HP_ERRATA_50XXB0);
if (fix_apm_sq) {
- tmp = readl(phy_mmio + MV5_LT_MODE);
+ tmp = readl(phy_mmio + MV5_LTMODE_OFS);
tmp |= (1 << 19);
- writel(tmp, phy_mmio + MV5_LT_MODE);
+ writel(tmp, phy_mmio + MV5_LTMODE_OFS);
- tmp = readl(phy_mmio + MV5_PHY_CTL);
+ tmp = readl(phy_mmio + MV5_PHY_CTL_OFS);
tmp &= ~0x3;
tmp |= 0x1;
- writel(tmp, phy_mmio + MV5_PHY_CTL);
+ writel(tmp, phy_mmio + MV5_PHY_CTL_OFS);
}
tmp = readl(phy_mmio + MV5_PHY_MODE);
{
void __iomem *port_mmio = mv_port_base(mmio, port);
- /*
- * The datasheet warns against setting ATA_RST when EDMA is active
- * (but doesn't say what the problem might be). So we first try
- * to disable the EDMA engine before doing the ATA_RST operation.
- */
mv_reset_channel(hpriv, mmio, port);
ZERO(0x028); /* command */
ZERO(0x024); /* respq outp */
ZERO(0x020); /* respq inp */
ZERO(0x02c); /* test control */
- writel(0xbc, port_mmio + EDMA_IORDY_TMOUT);
+ writel(0xbc, port_mmio + EDMA_IORDY_TMOUT_OFS);
}
#undef ZERO
struct mv_host_priv *hpriv = host->private_data;
u32 tmp;
- tmp = readl(mmio + MV_PCI_MODE);
+ tmp = readl(mmio + MV_PCI_MODE_OFS);
tmp &= 0xff00ffff;
- writel(tmp, mmio + MV_PCI_MODE);
+ writel(tmp, mmio + MV_PCI_MODE_OFS);
ZERO(MV_PCI_DISC_TIMER);
ZERO(MV_PCI_MSI_TRIGGER);
- writel(0x000100ff, mmio + MV_PCI_XBAR_TMOUT);
+ writel(0x000100ff, mmio + MV_PCI_XBAR_TMOUT_OFS);
ZERO(PCI_HC_MAIN_IRQ_MASK_OFS);
ZERO(MV_PCI_SERR_MASK);
ZERO(hpriv->irq_cause_ofs);
mv5_reset_flash(hpriv, mmio);
- tmp = readl(mmio + MV_GPIO_PORT_CTL);
+ tmp = readl(mmio + MV_GPIO_PORT_CTL_OFS);
tmp &= 0x3;
tmp |= (1 << 5) | (1 << 6);
- writel(tmp, mmio + MV_GPIO_PORT_CTL);
+ writel(tmp, mmio + MV_GPIO_PORT_CTL_OFS);
}
/**
void __iomem *port_mmio;
u32 tmp;
- tmp = readl(mmio + MV_RESET_CFG);
+ tmp = readl(mmio + MV_RESET_CFG_OFS);
if ((tmp & (1 << 0)) == 0) {
hpriv->signal[idx].amps = 0x7 << 8;
hpriv->signal[idx].pre = 0x1 << 5;
static void mv6_enable_leds(struct mv_host_priv *hpriv, void __iomem *mmio)
{
- writel(0x00000060, mmio + MV_GPIO_PORT_CTL);
+ writel(0x00000060, mmio + MV_GPIO_PORT_CTL_OFS);
}
static void mv6_phy_errata(struct mv_host_priv *hpriv, void __iomem *mmio,
{
void __iomem *port_mmio = mv_port_base(mmio, port);
- /*
- * The datasheet warns against setting ATA_RST when EDMA is active
- * (but doesn't say what the problem might be). So we first try
- * to disable the EDMA engine before doing the ATA_RST operation.
- */
mv_reset_channel(hpriv, mmio, port);
ZERO(0x028); /* command */
ZERO(0x024); /* respq outp */
ZERO(0x020); /* respq inp */
ZERO(0x02c); /* test control */
- writel(0xbc, port_mmio + EDMA_IORDY_TMOUT);
+ writel(0xbc, port_mmio + EDMA_IORDY_TMOUT_OFS);
}
#undef ZERO
return;
}
-static void mv_setup_ifctl(void __iomem *port_mmio, int want_gen2i)
+static void mv_setup_ifcfg(void __iomem *port_mmio, int want_gen2i)
{
- u32 ifctl = readl(port_mmio + SATA_INTERFACE_CFG);
+ u32 ifcfg = readl(port_mmio + SATA_INTERFACE_CFG_OFS);
- ifctl = (ifctl & 0xf7f) | 0x9b1000; /* from chip spec */
+ ifcfg = (ifcfg & 0xf7f) | 0x9b1000; /* from chip spec */
if (want_gen2i)
- ifctl |= (1 << 7); /* enable gen2i speed */
- writelfl(ifctl, port_mmio + SATA_INTERFACE_CFG);
+ ifcfg |= (1 << 7); /* enable gen2i speed */
+ writelfl(ifcfg, port_mmio + SATA_INTERFACE_CFG_OFS);
}
-/*
- * Caller must ensure that EDMA is not active,
- * by first doing mv_stop_edma() where needed.
- */
static void mv_reset_channel(struct mv_host_priv *hpriv, void __iomem *mmio,
unsigned int port_no)
{
void __iomem *port_mmio = mv_port_base(mmio, port_no);
+ /*
+ * The datasheet warns against setting EDMA_RESET when EDMA is active
+ * (but doesn't say what the problem might be). So we first try
+ * to disable the EDMA engine before doing the EDMA_RESET operation.
+ */
mv_stop_edma_engine(port_mmio);
- writelfl(ATA_RST, port_mmio + EDMA_CMD_OFS);
+ writelfl(EDMA_RESET, port_mmio + EDMA_CMD_OFS);
if (!IS_GEN_I(hpriv)) {
- /* Enable 3.0gb/s link speed */
- mv_setup_ifctl(port_mmio, 1);
+ /* Enable 3.0gb/s link speed: this survives EDMA_RESET */
+ mv_setup_ifcfg(port_mmio, 1);
}
/*
- * Strobing ATA_RST here causes a hard reset of the SATA transport,
+ * Strobing EDMA_RESET here causes a hard reset of the SATA transport,
* link, and physical layers. It resets all SATA interface registers
* (except for SATA_INTERFACE_CFG), and issues a COMRESET to the dev.
*/
- writelfl(ATA_RST, port_mmio + EDMA_CMD_OFS);
+ writelfl(EDMA_RESET, port_mmio + EDMA_CMD_OFS);
udelay(25); /* allow reset propagation */
writelfl(0, port_mmio + EDMA_CMD_OFS);
sata_scr_read(link, SCR_STATUS, &sstatus);
if (!IS_GEN_I(hpriv) && ++attempts >= 5 && sstatus == 0x121) {
/* Force 1.5gb/s link speed and try again */
- mv_setup_ifctl(mv_ap_base(ap), 0);
+ mv_setup_ifcfg(mv_ap_base(ap), 0);
if (time_after(jiffies + HZ, deadline))
extra = HZ; /* only extend it once, max */
}
readl(port_mmio + EDMA_ERR_IRQ_MASK_OFS));
}
+static unsigned int mv_in_pcix_mode(struct ata_host *host)
+{
+ struct mv_host_priv *hpriv = host->private_data;
+ void __iomem *mmio = hpriv->base;
+ u32 reg;
+
+ if (!HAS_PCI(host) || !IS_PCIE(hpriv))
+ return 0; /* not PCI-X capable */
+ reg = readl(mmio + MV_PCI_MODE_OFS);
+ if ((reg & MV_PCI_MODE_MASK) == 0)
+ return 0; /* conventional PCI mode */
+ return 1; /* chip is in PCI-X mode */
+}
+
+static int mv_pci_cut_through_okay(struct ata_host *host)
+{
+ struct mv_host_priv *hpriv = host->private_data;
+ void __iomem *mmio = hpriv->base;
+ u32 reg;
+
+ if (!mv_in_pcix_mode(host)) {
+ reg = readl(mmio + PCI_COMMAND_OFS);
+ if (reg & PCI_COMMAND_MRDTRIG)
+ return 0; /* not okay */
+ }
+ return 1; /* okay */
+}
+
static int mv_chip_id(struct ata_host *host, unsigned int board_idx)
{
struct pci_dev *pdev = to_pci_dev(host->dev);
break;
case chip_7042:
- hp_flags |= MV_HP_PCIE;
+ hp_flags |= MV_HP_PCIE | MV_HP_CUT_THROUGH;
if (pdev->vendor == PCI_VENDOR_ID_TTI &&
(pdev->device == 0x2300 || pdev->device == 0x2310))
{
" and avoid the final two gigabytes on"
" all RocketRAID BIOS initialized drives.\n");
}
+ /* drop through */
case chip_6042:
hpriv->ops = &mv6xxx_ops;
hp_flags |= MV_HP_GEN_IIE;
+ if (board_idx == chip_6042 && mv_pci_cut_through_okay(host))
+ hp_flags |= MV_HP_CUT_THROUGH;
switch (pdev->revision) {
case 0x0:
}
/* Check whether this driver has already been added to a class. */
- if ((drv->entry.next != drv->entry.prev) ||
- (drv->entry.next != NULL)) {
+ if (drv->entry.next && !list_empty(&drv->entry)) {
printk(KERN_WARNING "sysdev: class %s: driver (%p) has already"
" been registered to a class, something is wrong, but "
"will forge on!\n", cls->name, drv);
return 0;
if (!info->xmit_buf)
- return;
+ return 0;
local_irq_save(flags);
if (info->xmit_cnt >= PAGE_SIZE - 1) {
#define PCI_VEND_DEV(vend, dev) PCI_VENDOR_ID_ ## vend, \
PCI_DEVICE_ID_ ## vend ## _ ## dev
-#define dev_name(dev) (dev)->dev_name
+#define edac_dev_name(dev) (dev)->dev_name
/* memory devices */
enum dev_type {
fail0:
edac_printk(KERN_WARNING, EDAC_MC,
"%s (%s) %s %s already assigned %d\n",
- rover->dev->bus_id, dev_name(rover),
+ rover->dev->bus_id, edac_dev_name(rover),
rover->mod_name, rover->ctl_name, rover->dev_idx);
return 1;
"'%s': DEV '%s' (%s)\n",
edac_dev->mod_name,
edac_dev->ctl_name,
- dev_name(edac_dev),
+ edac_dev_name(edac_dev),
edac_op_state_to_string(edac_dev->op_state));
mutex_unlock(&device_ctls_mutex);
edac_printk(KERN_INFO, EDAC_MC,
"Removed device %d for %s %s: DEV %s\n",
edac_dev->dev_idx,
- edac_dev->mod_name, edac_dev->ctl_name, dev_name(edac_dev));
+ edac_dev->mod_name, edac_dev->ctl_name, edac_dev_name(edac_dev));
return edac_dev;
}
fail0:
edac_printk(KERN_WARNING, EDAC_MC,
"%s (%s) %s %s already assigned %d\n", p->dev->bus_id,
- dev_name(mci), p->mod_name, p->ctl_name, p->mc_idx);
+ edac_dev_name(mci), p->mod_name, p->ctl_name, p->mc_idx);
return 1;
fail1:
/* Report action taken */
edac_mc_printk(mci, KERN_INFO, "Giving out device to '%s' '%s':"
- " DEV %s\n", mci->mod_name, mci->ctl_name, dev_name(mci));
+ " DEV %s\n", mci->mod_name, mci->ctl_name, edac_dev_name(mci));
mutex_unlock(&mem_ctls_mutex);
return 0;
edac_printk(KERN_INFO, EDAC_MC,
"Removed device %d for %s %s: DEV %s\n", mci->mc_idx,
- mci->mod_name, mci->ctl_name, dev_name(mci));
+ mci->mod_name, mci->ctl_name, edac_dev_name(mci));
return mci;
}
fail0:
edac_printk(KERN_WARNING, EDAC_PCI,
"%s (%s) %s %s already assigned %d\n",
- rover->dev->bus_id, dev_name(rover),
+ rover->dev->bus_id, edac_dev_name(rover),
rover->mod_name, rover->ctl_name, rover->pci_idx);
return 1;
" DEV '%s' (%s)\n",
pci->mod_name,
pci->ctl_name,
- dev_name(pci), edac_op_state_to_string(pci->op_state));
+ edac_dev_name(pci), edac_op_state_to_string(pci->op_state));
mutex_unlock(&edac_pci_ctls_mutex);
return 0;
edac_printk(KERN_INFO, EDAC_PCI,
"Removed device %d for %s %s: DEV %s\n",
- pci->pci_idx, pci->mod_name, pci->ctl_name, dev_name(pci));
+ pci->pci_idx, pci->mod_name, pci->ctl_name, edac_dev_name(pci));
return pci;
}
unsigned long data_addr = drive->hwif->io_ports.data_addr;
if (drive->media == ide_disk && rq && rq->cmd_type == REQ_TYPE_FS)
- return outsw(data_adr, buf, (len + 1) / 2);
+ return outsw(data_addr, buf, (len + 1) / 2);
outsw_swapw(data_addr, buf, (len + 1) / 2);
}
cq->sw_wptr++;
}
-void cxio_flush_rq(struct t3_wq *wq, struct t3_cq *cq, int count)
+int cxio_flush_rq(struct t3_wq *wq, struct t3_cq *cq, int count)
{
u32 ptr;
+ int flushed = 0;
PDBG("%s wq %p cq %p\n", __func__, wq, cq);
PDBG("%s rq_rptr %u rq_wptr %u skip count %u\n", __func__,
wq->rq_rptr, wq->rq_wptr, count);
ptr = wq->rq_rptr + count;
- while (ptr++ != wq->rq_wptr)
+ while (ptr++ != wq->rq_wptr) {
insert_recv_cqe(wq, cq);
+ flushed++;
+ }
+ return flushed;
}
static void insert_sq_cqe(struct t3_wq *wq, struct t3_cq *cq,
cq->sw_wptr++;
}
-void cxio_flush_sq(struct t3_wq *wq, struct t3_cq *cq, int count)
+int cxio_flush_sq(struct t3_wq *wq, struct t3_cq *cq, int count)
{
__u32 ptr;
+ int flushed = 0;
struct t3_swsq *sqp = wq->sq + Q_PTR2IDX(wq->sq_rptr, wq->sq_size_log2);
ptr = wq->sq_rptr + count;
insert_sq_cqe(wq, cq, sqp);
sqp++;
ptr++;
+ flushed++;
}
+ return flushed;
}
/*
void cxio_hal_put_pdid(struct cxio_hal_resource *rscp, u32 pdid);
int __init cxio_hal_init(void);
void __exit cxio_hal_exit(void);
-void cxio_flush_rq(struct t3_wq *wq, struct t3_cq *cq, int count);
-void cxio_flush_sq(struct t3_wq *wq, struct t3_cq *cq, int count);
+int cxio_flush_rq(struct t3_wq *wq, struct t3_cq *cq, int count);
+int cxio_flush_sq(struct t3_wq *wq, struct t3_cq *cq, int count);
void cxio_count_rcqes(struct t3_cq *cq, struct t3_wq *wq, int *count);
void cxio_count_scqes(struct t3_cq *cq, struct t3_wq *wq, int *count);
void cxio_flush_hw_cq(struct t3_cq *cq);
module_param(peer2peer, int, 0644);
MODULE_PARM_DESC(peer2peer, "Support peer2peer ULPs (default=0)");
-static int ep_timeout_secs = 10;
+static int ep_timeout_secs = 60;
module_param(ep_timeout_secs, int, 0644);
MODULE_PARM_DESC(ep_timeout_secs, "CM Endpoint operation timeout "
- "in seconds (default=10)");
+ "in seconds (default=60)");
static int mpa_rev = 1;
module_param(mpa_rev, int, 0644);
release = 1;
break;
case ABORTING:
- break;
case DEAD:
+ break;
default:
BUG_ON(1);
break;
{
struct iwch_cq *rchp, *schp;
int count;
+ int flushed;
rchp = get_chp(qhp->rhp, qhp->attr.rcq);
schp = get_chp(qhp->rhp, qhp->attr.scq);
spin_lock(&qhp->lock);
cxio_flush_hw_cq(&rchp->cq);
cxio_count_rcqes(&rchp->cq, &qhp->wq, &count);
- cxio_flush_rq(&qhp->wq, &rchp->cq, count);
+ flushed = cxio_flush_rq(&qhp->wq, &rchp->cq, count);
spin_unlock(&qhp->lock);
spin_unlock_irqrestore(&rchp->lock, *flag);
- (*rchp->ibcq.comp_handler)(&rchp->ibcq, rchp->ibcq.cq_context);
+ if (flushed)
+ (*rchp->ibcq.comp_handler)(&rchp->ibcq, rchp->ibcq.cq_context);
/* locking heirarchy: cq lock first, then qp lock. */
spin_lock_irqsave(&schp->lock, *flag);
spin_lock(&qhp->lock);
cxio_flush_hw_cq(&schp->cq);
cxio_count_scqes(&schp->cq, &qhp->wq, &count);
- cxio_flush_sq(&qhp->wq, &schp->cq, count);
+ flushed = cxio_flush_sq(&qhp->wq, &schp->cq, count);
spin_unlock(&qhp->lock);
spin_unlock_irqrestore(&schp->lock, *flag);
- (*schp->ibcq.comp_handler)(&schp->ibcq, schp->ibcq.cq_context);
+ if (flushed)
+ (*schp->ibcq.comp_handler)(&schp->ibcq, schp->ibcq.cq_context);
/* deref */
if (atomic_dec_and_test(&qhp->refcnt))
ep = qhp->ep;
get_ep(&ep->com);
}
- flush_qp(qhp, &flag);
break;
case IWCH_QP_STATE_TERMINATE:
qhp->attr.state = IWCH_QP_STATE_TERMINATE;
}
switch (attrs->next_state) {
case IWCH_QP_STATE_IDLE:
+ flush_qp(qhp, &flag);
qhp->attr.state = IWCH_QP_STATE_IDLE;
qhp->attr.llp_stream_handle = NULL;
put_ep(&qhp->ep->com);
props->max_ee = limit_uint(rblock->max_rd_ee_context);
props->max_rdd = limit_uint(rblock->max_rd_domain);
props->max_fmr = limit_uint(rblock->max_mr);
- props->local_ca_ack_delay = limit_uint(rblock->local_ca_ack_delay);
props->max_qp_rd_atom = limit_uint(rblock->max_rr_qp);
props->max_ee_rd_atom = limit_uint(rblock->max_rr_ee_context);
props->max_res_rd_atom = limit_uint(rblock->max_rr_hca);
}
props->max_pkeys = 16;
- props->local_ca_ack_delay = limit_uint(rblock->local_ca_ack_delay);
+ props->local_ca_ack_delay = min_t(u8, rblock->local_ca_ack_delay, 255);
props->max_raw_ipv6_qp = limit_uint(rblock->max_raw_ipv6_qp);
props->max_raw_ethy_qp = limit_uint(rblock->max_raw_ethy_qp);
props->max_mcast_grp = limit_uint(rblock->max_mcast_grp);
return ret;
}
-static int map_mtu(struct ehca_shca *shca, u32 fw_mtu)
+static enum ib_mtu map_mtu(struct ehca_shca *shca, u32 fw_mtu)
{
switch (fw_mtu) {
case 0x1:
}
}
-static int map_number_of_vls(struct ehca_shca *shca, u32 vl_cap)
+static u8 map_number_of_vls(struct ehca_shca *shca, u32 vl_cap)
{
switch (vl_cap) {
case 0x1:
if (context)
ib_umem_release(cq->umem);
else
- mlx4_ib_free_cq_buf(dev, &cq->buf, entries);
+ mlx4_ib_free_cq_buf(dev, &cq->buf, cq->ibcq.cqe);
err_db:
if (!context)
mlx4_ib_db_unmap_user(to_mucontext(cq->uobject->context), &mcq->db);
ib_umem_release(mcq->umem);
} else {
- mlx4_ib_free_cq_buf(dev, &mcq->buf, cq->cqe + 1);
+ mlx4_ib_free_cq_buf(dev, &mcq->buf, cq->cqe);
mlx4_db_free(dev->dev, &mcq->db);
}
#endif
int hca_caps;
struct ipoib_ethtool_st ethtool;
+ struct timer_list poll_timer;
};
struct ipoib_ah {
int ipoib_poll(struct napi_struct *napi, int budget);
void ipoib_ib_completion(struct ib_cq *cq, void *dev_ptr);
+void ipoib_send_comp_handler(struct ib_cq *cq, void *dev_ptr);
struct ipoib_ah *ipoib_create_ah(struct net_device *dev,
struct ib_pd *pd, struct ib_ah_attr *attr);
netif_rx_schedule(dev, &priv->napi);
}
+static void drain_tx_cq(struct net_device *dev)
+{
+ struct ipoib_dev_priv *priv = netdev_priv(dev);
+ unsigned long flags;
+
+ spin_lock_irqsave(&priv->tx_lock, flags);
+ while (poll_tx(priv))
+ ; /* nothing */
+
+ if (netif_queue_stopped(dev))
+ mod_timer(&priv->poll_timer, jiffies + 1);
+
+ spin_unlock_irqrestore(&priv->tx_lock, flags);
+}
+
+void ipoib_send_comp_handler(struct ib_cq *cq, void *dev_ptr)
+{
+ drain_tx_cq((struct net_device *)dev_ptr);
+}
+
static inline int post_send(struct ipoib_dev_priv *priv,
unsigned int wr_id,
struct ib_ah *address, u32 qpn,
else
priv->tx_wr.send_flags &= ~IB_SEND_IP_CSUM;
+ if (++priv->tx_outstanding == ipoib_sendq_size) {
+ ipoib_dbg(priv, "TX ring full, stopping kernel net queue\n");
+ if (ib_req_notify_cq(priv->send_cq, IB_CQ_NEXT_COMP))
+ ipoib_warn(priv, "request notify on send CQ failed\n");
+ netif_stop_queue(dev);
+ }
+
if (unlikely(post_send(priv, priv->tx_head & (ipoib_sendq_size - 1),
address->ah, qpn, tx_req, phead, hlen))) {
ipoib_warn(priv, "post_send failed\n");
++dev->stats.tx_errors;
+ --priv->tx_outstanding;
ipoib_dma_unmap_tx(priv->ca, tx_req);
dev_kfree_skb_any(skb);
+ if (netif_queue_stopped(dev))
+ netif_wake_queue(dev);
} else {
dev->trans_start = jiffies;
++priv->tx_head;
skb_orphan(skb);
- if (++priv->tx_outstanding == ipoib_sendq_size) {
- ipoib_dbg(priv, "TX ring full, stopping kernel net queue\n");
- netif_stop_queue(dev);
- }
}
if (unlikely(priv->tx_outstanding > MAX_SEND_CQE))
- poll_tx(priv);
+ while (poll_tx(priv))
+ ; /* nothing */
}
static void __ipoib_reap_ah(struct net_device *dev)
round_jiffies_relative(HZ));
}
+static void ipoib_ib_tx_timer_func(unsigned long ctx)
+{
+ drain_tx_cq((struct net_device *)ctx);
+}
+
int ipoib_ib_dev_open(struct net_device *dev)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
queue_delayed_work(ipoib_workqueue, &priv->ah_reap_task,
round_jiffies_relative(HZ));
+ init_timer(&priv->poll_timer);
+ priv->poll_timer.function = ipoib_ib_tx_timer_func;
+ priv->poll_timer.data = (unsigned long)dev;
+
set_bit(IPOIB_FLAG_INITIALIZED, &priv->flags);
return 0;
ipoib_dbg(priv, "All sends and receives done.\n");
timeout:
+ del_timer_sync(&priv->poll_timer);
qp_attr.qp_state = IB_QPS_RESET;
if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE))
ipoib_warn(priv, "Failed to modify QP to RESET state\n");
goto out_free_mr;
}
- priv->send_cq = ib_create_cq(priv->ca, NULL, NULL, dev, ipoib_sendq_size, 0);
+ priv->send_cq = ib_create_cq(priv->ca, ipoib_send_comp_handler, NULL,
+ dev, ipoib_sendq_size, 0);
if (IS_ERR(priv->send_cq)) {
printk(KERN_WARNING "%s: failed to create send CQ\n", ca->name);
goto out_free_recv_cq;
#include <linux/time.h>
#include <linux/slab.h>
#include <linux/hil.h>
+#include <linux/semaphore.h>
#include <asm/io.h>
#include <asm/system.h>
#include <linux/device.h>
#include <linux/kthread.h>
#include <linux/platform_device.h>
-#include <linux/semaphore.h>
+#include <linux/mutex.h>
#include <asm/uaccess.h>
#ifdef CONFIG_PPC
} adb_handler[16];
/*
- * The adb_handler_sem mutex protects all accesses to the original_address
+ * The adb_handler_mutex mutex protects all accesses to the original_address
* and handler_id fields of adb_handler[i] for all i, and changes to the
* handler field.
* Accesses to the handler field are protected by the adb_handler_lock
* time adb_unregister returns, we know that the old handler isn't being
* called.
*/
-static DECLARE_MUTEX(adb_handler_sem);
+static DEFINE_MUTEX(adb_handler_mutex);
static DEFINE_RWLOCK(adb_handler_lock);
#if 0
msleep(500);
}
- down(&adb_handler_sem);
+ mutex_lock(&adb_handler_mutex);
write_lock_irq(&adb_handler_lock);
memset(adb_handler, 0, sizeof(adb_handler));
write_unlock_irq(&adb_handler_lock);
if (adb_controller->autopoll)
adb_controller->autopoll(autopoll_devs);
}
- up(&adb_handler_sem);
+ mutex_unlock(&adb_handler_mutex);
blocking_notifier_call_chain(&adb_client_list,
ADB_MSG_POST_RESET, NULL);
{
int i;
- down(&adb_handler_sem);
+ mutex_lock(&adb_handler_mutex);
ids->nids = 0;
for (i = 1; i < 16; i++) {
if ((adb_handler[i].original_address == default_id) &&
ids->id[ids->nids++] = i;
}
}
- up(&adb_handler_sem);
+ mutex_unlock(&adb_handler_mutex);
return ids->nids;
}
{
int ret = -ENODEV;
- down(&adb_handler_sem);
+ mutex_lock(&adb_handler_mutex);
write_lock_irq(&adb_handler_lock);
if (adb_handler[index].handler) {
while(adb_handler[index].busy) {
adb_handler[index].handler = NULL;
}
write_unlock_irq(&adb_handler_lock);
- up(&adb_handler_sem);
+ mutex_unlock(&adb_handler_mutex);
return ret;
}
{
int ret;
- down(&adb_handler_sem);
+ mutex_lock(&adb_handler_mutex);
ret = try_handler_change(address, new_id);
- up(&adb_handler_sem);
+ mutex_unlock(&adb_handler_mutex);
return ret;
}
int
adb_get_infos(int address, int *original_address, int *handler_id)
{
- down(&adb_handler_sem);
+ mutex_lock(&adb_handler_mutex);
*original_address = adb_handler[address].original_address;
*handler_id = adb_handler[address].handler_id;
- up(&adb_handler_sem);
+ mutex_unlock(&adb_handler_mutex);
return (*original_address != 0);
}
case ADB_QUERY_GETDEVINFO:
if (req->nbytes < 3)
break;
- down(&adb_handler_sem);
+ mutex_lock(&adb_handler_mutex);
req->reply[0] = adb_handler[req->data[2]].original_address;
req->reply[1] = adb_handler[req->data[2]].handler_id;
- up(&adb_handler_sem);
+ mutex_unlock(&adb_handler_mutex);
req->complete = 1;
req->reply_len = 2;
adb_write_done(req);
#include <linux/kmod.h>
#include <linux/i2c.h>
#include <linux/kthread.h>
+#include <linux/mutex.h>
#include <asm/prom.h>
#include <asm/machdep.h>
#include <asm/io.h>
static s32 dimm_output_clamp;
static int fcu_rpm_shift;
static int fcu_tickle_ticks;
-static DECLARE_MUTEX(driver_lock);
+static DEFINE_MUTEX(driver_lock);
/*
* We have 3 types of CPU PID control. One is "split" old style control
static ssize_t show_##name(struct device *dev, struct device_attribute *attr, char *buf) \
{ \
ssize_t r; \
- down(&driver_lock); \
+ mutex_lock(&driver_lock); \
r = sprintf(buf, "%d.%03d", FIX32TOPRINT(data)); \
- up(&driver_lock); \
+ mutex_unlock(&driver_lock); \
return r; \
}
#define BUILD_SHOW_FUNC_INT(name, data) \
{
DBG("main_control_loop started\n");
- down(&driver_lock);
+ mutex_lock(&driver_lock);
if (start_fcu() < 0) {
printk(KERN_ERR "kfand: failed to start FCU\n");
- up(&driver_lock);
+ mutex_unlock(&driver_lock);
goto out;
}
fcu_tickle_ticks = FCU_TICKLE_TICKS;
- up(&driver_lock);
+ mutex_unlock(&driver_lock);
while (state == state_attached) {
unsigned long elapsed, start;
start = jiffies;
- down(&driver_lock);
+ mutex_lock(&driver_lock);
/* Tickle the FCU just in case */
if (--fcu_tickle_ticks < 0) {
do_monitor_slots(&slots_state);
else
do_monitor_drives(&drives_state);
- up(&driver_lock);
+ mutex_unlock(&driver_lock);
if (critical_state == 1) {
printk(KERN_WARNING "Temperature control detected a critical condition\n");
*/
static int therm_pm72_attach(struct i2c_adapter *adapter)
{
- down(&driver_lock);
+ mutex_lock(&driver_lock);
/* Check state */
if (state == state_detached)
state = state_attaching;
if (state != state_attaching) {
- up(&driver_lock);
+ mutex_unlock(&driver_lock);
return 0;
}
state = state_attached;
start_control_loops();
}
- up(&driver_lock);
+ mutex_unlock(&driver_lock);
return 0;
}
*/
static int therm_pm72_detach(struct i2c_adapter *adapter)
{
- down(&driver_lock);
+ mutex_lock(&driver_lock);
if (state != state_detached)
state = state_detaching;
/* Stop control loops if any */
DBG("stopping control loops\n");
- up(&driver_lock);
+ mutex_unlock(&driver_lock);
stop_control_loops();
- down(&driver_lock);
+ mutex_lock(&driver_lock);
if (u3_0 != NULL && !strcmp(adapter->name, "u3 0")) {
DBG("lost U3-0, disposing control loops\n");
if (u3_0 == NULL && u3_1 == NULL)
state = state_detached;
- up(&driver_lock);
+ mutex_unlock(&driver_lock);
return 0;
}
#include <linux/init.h>
#include <linux/wait.h>
#include <linux/i2c.h>
-#include <linux/semaphore.h>
+#include <linux/mutex.h>
#include <asm/prom.h>
#include <asm/smu.h>
#include <asm/pmac_low_i2c.h>
struct wf_sat {
int nr;
atomic_t refcnt;
- struct semaphore mutex;
+ struct mutex mutex;
unsigned long last_read; /* jiffies when cache last updated */
u8 cache[16];
struct i2c_client i2c;
if (sat->i2c.adapter == NULL)
return -ENODEV;
- down(&sat->mutex);
+ mutex_lock(&sat->mutex);
if (time_after(jiffies, (sat->last_read + MAX_AGE))) {
err = wf_sat_read_cache(sat);
if (err)
err = 0;
fail:
- up(&sat->mutex);
+ mutex_unlock(&sat->mutex);
return err;
}
sat->nr = -1;
sat->node = of_node_get(dev);
atomic_set(&sat->refcnt, 0);
- init_MUTEX(&sat->mutex);
+ mutex_init(&sat->mutex);
sat->i2c.addr = (addr >> 1) & 0x7f;
sat->i2c.adapter = adapter;
sat->i2c.driver = &wf_sat_driver;
static int nsc_ircc_init_108(nsc_chip_t *chip, chipio_t *info);
static int nsc_ircc_init_338(nsc_chip_t *chip, chipio_t *info);
static int nsc_ircc_init_39x(nsc_chip_t *chip, chipio_t *info);
+#ifdef CONFIG_PNP
static int nsc_ircc_pnp_probe(struct pnp_dev *dev, const struct pnp_device_id *id);
+#endif
/* These are the known NSC chips */
static nsc_chip_t chips[] = {
MODULE_DEVICE_TABLE(pnp, nsc_ircc_pnp_table);
static struct pnp_driver nsc_ircc_pnp_driver = {
+#ifdef CONFIG_PNP
.name = "nsc-ircc",
.id_table = nsc_ircc_pnp_table,
.probe = nsc_ircc_pnp_probe,
+#endif
};
/* Some prototypes */
return 0;
}
+#ifdef CONFIG_PNP
/* PNP probing */
static int nsc_ircc_pnp_probe(struct pnp_dev *dev, const struct pnp_device_id *id)
{
return 0;
}
+#endif
/*
* Function nsc_ircc_setup (info)
static int pnp_driver_registered;
+#ifdef CONFIG_PNP
static int __init smsc_ircc_pnp_probe(struct pnp_dev *dev,
const struct pnp_device_id *dev_id)
{
.id_table = smsc_ircc_pnp_table,
.probe = smsc_ircc_pnp_probe,
};
-
+#else /* CONFIG_PNP */
+static struct pnp_driver smsc_ircc_pnp_driver;
+#endif
/*******************************************************************************
*
u64 mtt_seg;
int err = -ENOMEM;
- if (page_shift < 12 || page_shift >= 32)
+ if (page_shift < (ffs(dev->caps.page_size_cap) - 1) || page_shift >= 32)
return -EINVAL;
/* All MTTs must fit in the same page */
/* niu.c: Neptune ethernet driver.
*
- * Copyright (C) 2007 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 2007, 2008 David S. Miller (davem@davemloft.net)
*/
#include <linux/module.h>
#define DRV_MODULE_NAME "niu"
#define PFX DRV_MODULE_NAME ": "
-#define DRV_MODULE_VERSION "0.8"
-#define DRV_MODULE_RELDATE "April 24, 2008"
+#define DRV_MODULE_VERSION "0.9"
+#define DRV_MODULE_RELDATE "May 4, 2008"
static char version[] __devinitdata =
DRV_MODULE_NAME ".c:v" DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")\n";
parent->num_ports = nr64(ESPC_NUM_PORTS_MACS) &
ESPC_NUM_PORTS_MACS_VAL;
+ /* All of the current probing methods fail on
+ * Maramba on-board parts.
+ */
if (!parent->num_ports)
- return -ENODEV;
+ parent->num_ports = 4;
}
}
}
list_for_each_safe(entry, tmp, &lapbeth_devices) {
lapbeth = list_entry(entry, struct lapbethdev, node);
+ dev_put(lapbeth->ethdev);
unregister_netdevice(lapbeth->axdev);
}
rtnl_unlock();
config IWLWIFI
- bool
- default n
+ tristate
config IWLCORE
tristate "Intel Wireless Wifi Core"
* reading the dword at 0x100 which must either be 0 or a valid extended
* capability header.
*/
-int pci_cfg_space_size_ext(struct pci_dev *dev, unsigned check_exp_pcix)
+int pci_cfg_space_size_ext(struct pci_dev *dev)
{
- int pos;
u32 status;
- if (!check_exp_pcix)
- goto skip;
+ if (pci_read_config_dword(dev, 256, &status) != PCIBIOS_SUCCESSFUL)
+ goto fail;
+ if (status == 0xffffffff)
+ goto fail;
+
+ return PCI_CFG_SPACE_EXP_SIZE;
+
+ fail:
+ return PCI_CFG_SPACE_SIZE;
+}
+
+int pci_cfg_space_size(struct pci_dev *dev)
+{
+ int pos;
+ u32 status;
pos = pci_find_capability(dev, PCI_CAP_ID_EXP);
if (!pos) {
goto fail;
}
- skip:
- if (pci_read_config_dword(dev, 256, &status) != PCIBIOS_SUCCESSFUL)
- goto fail;
- if (status == 0xffffffff)
- goto fail;
-
- return PCI_CFG_SPACE_EXP_SIZE;
+ return pci_cfg_space_size_ext(dev);
fail:
return PCI_CFG_SPACE_SIZE;
}
-int pci_cfg_space_size(struct pci_dev *dev)
-{
- return pci_cfg_space_size_ext(dev, 1);
-}
-
static void pci_release_bus_bridge_dev(struct device *dev)
{
kfree(dev);
#define BPP_DELAY 100
static const unsigned BPP_MAJOR = LP_MAJOR;
-static const char* dev_name = "bpp";
+static const char *bpp_dev_name = "bpp";
/* When switching from compatibility to a mode where I can read, try
the following mode first. */
config SCSI_MAC_ESP
tristate "Macintosh NCR53c9[46] SCSI"
depends on MAC && SCSI
+ select SCSI_SPI_ATTRS
help
This is the NCR 53c9x SCSI controller found on most of the 68040
based Macintoshes.
config USB_ISP1760_OF
bool "Support for the OF platform bus"
- depends on USB_ISP1760_HCD && OF
+ depends on USB_ISP1760_HCD && PPC_OF
---help---
Enables support for the device present on the PowerPC
OpenFirmware platform bus.
if (!info->screen_base)
goto out_unmap_regs;
- bw2_blank(0, info);
+ bw2_blank(FB_BLANK_UNBLANK, info);
bw2_init_fix(info, linebytes);
if (!info->screen_base)
goto out_unmap_regs;
- cg3_blank(0, info);
+ cg3_blank(FB_BLANK_UNBLANK, info);
if (!of_find_property(dp, "width", NULL)) {
err = cg3_do_default_mode(par);
cg6_bt_init(par);
cg6_chip_init(info);
- cg6_blank(0, info);
+ cg6_blank(FB_BLANK_UNBLANK, info);
if (fb_alloc_cmap(&info->cmap, 256, 0))
goto out_unmap_regs;
* chosen console, it will have video outputs off in
* the DAC.
*/
- ffb_blank(0, info);
+ ffb_blank(FB_BLANK_UNBLANK, info);
if (fb_alloc_cmap(&info->cmap, 256, 0))
goto out_unmap_dac;
leo_init_wids(info);
leo_init_hw(info);
- leo_blank(0, info);
+ leo_blank(FB_BLANK_UNBLANK, info);
if (fb_alloc_cmap(&info->cmap, 256, 0))
goto out_unmap_regs;
if (!info->screen_base)
goto out_unmap_regs;
- p9100_blank(0, info);
+ p9100_blank(FB_BLANK_UNBLANK, info);
if (fb_alloc_cmap(&info->cmap, 256, 0))
goto out_unmap_screen;
struct tcx_thc {
u32 thc_rev;
- u32 thc_pad0[511];
+ u32 thc_pad0[511];
u32 thc_hs; /* hsync timing */
u32 thc_hsdvs;
u32 thc_hd;
};
/* Reset control plane so that WID is 8-bit plane. */
-static void __tcx_set_control_plane (struct tcx_par *par)
+static void __tcx_set_control_plane(struct tcx_par *par)
{
u32 __iomem *p, *pend;
-
+
if (par->lowdepth)
return;
sbus_writel(tmp, p);
}
}
-
-static void tcx_reset (struct fb_info *info)
+
+static void tcx_reset(struct fb_info *info)
{
struct tcx_par *par = (struct tcx_par *) info->par;
unsigned long flags;
info->screen_base, par->fbsize);
}
-static int __devinit tcx_init_one(struct of_device *op)
+static int __devinit tcx_probe(struct of_device *op,
+ const struct of_device_id *match)
{
struct device_node *dp = op->node;
struct fb_info *info;
return err;
}
-static int __devinit tcx_probe(struct of_device *dev, const struct of_device_id *match)
-{
- struct of_device *op = to_of_device(&dev->dev);
-
- return tcx_init_one(op);
-}
-
static int __devexit tcx_remove(struct of_device *op)
{
struct fb_info *info = dev_get_drvdata(&op->dev);
* affs fs inode data in memory
*/
struct affs_inode_info {
- u32 i_opencnt;
+ atomic_t i_opencnt;
struct semaphore i_link_lock; /* Protects internal inode access. */
struct semaphore i_ext_lock; /* Protects internal inode access. */
#define i_hash_lock i_ext_lock
extern unsigned long affs_parent_ino(struct inode *dir);
extern struct inode *affs_new_inode(struct inode *dir);
extern int affs_notify_change(struct dentry *dentry, struct iattr *attr);
-extern void affs_put_inode(struct inode *inode);
-extern void affs_drop_inode(struct inode *inode);
extern void affs_delete_inode(struct inode *inode);
extern void affs_clear_inode(struct inode *inode);
extern struct inode *affs_iget(struct super_block *sb,
{
if (atomic_read(&filp->f_count) != 1)
return 0;
- pr_debug("AFFS: open(%d)\n", AFFS_I(inode)->i_opencnt);
- AFFS_I(inode)->i_opencnt++;
+ pr_debug("AFFS: open(%lu,%d)\n",
+ inode->i_ino, atomic_read(&AFFS_I(inode)->i_opencnt));
+ atomic_inc(&AFFS_I(inode)->i_opencnt);
return 0;
}
{
if (atomic_read(&filp->f_count) != 0)
return 0;
- pr_debug("AFFS: release(%d)\n", AFFS_I(inode)->i_opencnt);
- AFFS_I(inode)->i_opencnt--;
- if (!AFFS_I(inode)->i_opencnt)
+ pr_debug("AFFS: release(%lu, %d)\n",
+ inode->i_ino, atomic_read(&AFFS_I(inode)->i_opencnt));
+
+ if (atomic_dec_and_test(&AFFS_I(inode)->i_opencnt)) {
+ mutex_lock(&inode->i_mutex);
+ if (inode->i_size != AFFS_I(inode)->mmu_private)
+ affs_truncate(inode);
affs_free_prealloc(inode);
+ mutex_unlock(&inode->i_mutex);
+ }
return 0;
}
/* inline the simplest case: same extended block as last time */
struct buffer_head *bh = AFFS_I(inode)->i_ext_bh;
if (ext == AFFS_I(inode)->i_ext_last)
- atomic_inc(&bh->b_count);
+ get_bh(bh);
else
/* we have to do more (not inlined) */
bh = affs_get_extblock_slow(inode, ext);
affs_brelse(AFFS_I(inode)->i_ext_bh);
AFFS_I(inode)->i_ext_last = ext;
AFFS_I(inode)->i_ext_bh = bh;
- atomic_inc(&bh->b_count);
+ get_bh(bh);
return bh;
pr_debug("AFFS: get_block(%u, %lu)\n", (u32)inode->i_ino, (unsigned long)block);
-
BUG_ON(block > (sector_t)0x7fffffffUL);
if (block >= AFFS_I(inode)->i_blkcnt) {
res = mapping->a_ops->write_begin(NULL, mapping, size, 0, 0, &page, &fsdata);
if (!res)
res = mapping->a_ops->write_end(NULL, mapping, size, 0, 0, page, fsdata);
+ else
+ inode->i_size = AFFS_I(inode)->mmu_private;
mark_inode_dirty(inode);
return;
} else if (inode->i_size == AFFS_I(inode)->mmu_private)
blk++;
} else
AFFS_HEAD(ext_bh)->first_data = 0;
+ AFFS_HEAD(ext_bh)->block_count = cpu_to_be32(i);
size = AFFS_SB(sb)->s_hashsize;
if (size > blkcnt - blk + i)
size = blkcnt - blk + i;
AFFS_I(inode)->i_extcnt = 1;
AFFS_I(inode)->i_ext_last = ~1;
AFFS_I(inode)->i_protect = prot;
- AFFS_I(inode)->i_opencnt = 0;
+ atomic_set(&AFFS_I(inode)->i_opencnt, 0);
AFFS_I(inode)->i_blkcnt = 0;
AFFS_I(inode)->i_lc = NULL;
AFFS_I(inode)->i_lc_size = 0;
inode->i_mode |= S_IFDIR;
} else
inode->i_mode = S_IRUGO | S_IXUGO | S_IWUSR | S_IFDIR;
- if (tail->link_chain)
- inode->i_nlink = 2;
/* Maybe it should be controlled by mount parameter? */
//inode->i_mode |= S_ISVTX;
inode->i_op = &affs_dir_inode_operations;
return error;
}
-void
-affs_put_inode(struct inode *inode)
-{
- pr_debug("AFFS: put_inode(ino=%lu, nlink=%u)\n", inode->i_ino, inode->i_nlink);
- affs_free_prealloc(inode);
-}
-
-void
-affs_drop_inode(struct inode *inode)
-{
- mutex_lock(&inode->i_mutex);
- if (inode->i_size != AFFS_I(inode)->mmu_private)
- affs_truncate(inode);
- mutex_unlock(&inode->i_mutex);
-
- generic_drop_inode(inode);
-}
-
void
affs_delete_inode(struct inode *inode)
{
pr_debug("AFFS: delete_inode(ino=%lu, nlink=%u)\n", inode->i_ino, inode->i_nlink);
truncate_inode_pages(&inode->i_data, 0);
inode->i_size = 0;
- if (S_ISREG(inode->i_mode))
- affs_truncate(inode);
+ affs_truncate(inode);
clear_inode(inode);
affs_free_block(inode->i_sb, inode->i_ino);
}
void
affs_clear_inode(struct inode *inode)
{
- unsigned long cache_page = (unsigned long) AFFS_I(inode)->i_lc;
+ unsigned long cache_page;
pr_debug("AFFS: clear_inode(ino=%lu, nlink=%u)\n", inode->i_ino, inode->i_nlink);
+
+ affs_free_prealloc(inode);
+ cache_page = (unsigned long)AFFS_I(inode)->i_lc;
if (cache_page) {
pr_debug("AFFS: freeing ext cache\n");
AFFS_I(inode)->i_lc = NULL;
inode->i_ino = block;
inode->i_nlink = 1;
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC;
- AFFS_I(inode)->i_opencnt = 0;
+ atomic_set(&AFFS_I(inode)->i_opencnt, 0);
AFFS_I(inode)->i_blkcnt = 0;
AFFS_I(inode)->i_lc = NULL;
AFFS_I(inode)->i_lc_size = 0;
switch (type) {
case ST_LINKFILE:
case ST_LINKDIR:
- inode_bh = bh;
retval = -ENOSPC;
block = affs_alloc_block(dir, dir->i_ino);
if (!block)
goto err;
retval = -EIO;
+ inode_bh = bh;
bh = affs_getzeroblk(sb, block);
if (!bh)
goto err;
int
affs_unlink(struct inode *dir, struct dentry *dentry)
{
- pr_debug("AFFS: unlink(dir=%d, \"%.*s\")\n", (u32)dir->i_ino,
+ pr_debug("AFFS: unlink(dir=%d, %lu \"%.*s\")\n", (u32)dir->i_ino,
+ dentry->d_inode->i_ino,
(int)dentry->d_name.len, dentry->d_name.name);
return affs_remove_header(dentry);
int
affs_rmdir(struct inode *dir, struct dentry *dentry)
{
- pr_debug("AFFS: rmdir(dir=%u, \"%.*s\")\n", (u32)dir->i_ino,
+ pr_debug("AFFS: rmdir(dir=%u, %lu \"%.*s\")\n", (u32)dir->i_ino,
+ dentry->d_inode->i_ino,
(int)dentry->d_name.len, dentry->d_name.name);
return affs_remove_header(dentry);
static struct inode *affs_alloc_inode(struct super_block *sb)
{
- struct affs_inode_info *ei;
- ei = (struct affs_inode_info *)kmem_cache_alloc(affs_inode_cachep, GFP_KERNEL);
- if (!ei)
+ struct affs_inode_info *i;
+
+ i = kmem_cache_alloc(affs_inode_cachep, GFP_KERNEL);
+ if (!i)
return NULL;
- ei->vfs_inode.i_version = 1;
- return &ei->vfs_inode;
+
+ i->vfs_inode.i_version = 1;
+ i->i_lc = NULL;
+ i->i_ext_bh = NULL;
+ i->i_pa_cnt = 0;
+
+ return &i->vfs_inode;
}
static void affs_destroy_inode(struct inode *inode)
.alloc_inode = affs_alloc_inode,
.destroy_inode = affs_destroy_inode,
.write_inode = affs_write_inode,
- .put_inode = affs_put_inode,
- .drop_inode = affs_drop_inode,
.delete_inode = affs_delete_inode,
.clear_inode = affs_clear_inode,
.put_super = affs_put_super,
void iput(struct inode *inode)
{
if (inode) {
- const struct super_operations *op = inode->i_sb->s_op;
-
BUG_ON(inode->i_state == I_CLEAR);
- if (op && op->put_inode)
- op->put_inode(inode);
-
if (atomic_dec_and_lock(&inode->i_count, &inode_lock))
iput_final(inode);
}
struct file_lock *file_lock = locks_alloc_lock();
struct flock flock;
struct inode *inode;
+ struct file *f;
int error;
if (file_lock == NULL)
* Attempt to detect a close/fcntl race and recover by
* releasing the lock that was just acquired.
*/
- if (!error && fcheck(fd) != filp && flock.l_type != F_UNLCK) {
+ /*
+ * we need that spin_lock here - it prevents reordering between
+ * update of inode->i_flock and check for it done in close().
+ * rcu_read_lock() wouldn't do.
+ */
+ spin_lock(¤t->files->file_lock);
+ f = fcheck(fd);
+ spin_unlock(¤t->files->file_lock);
+ if (!error && f != filp && flock.l_type != F_UNLCK) {
flock.l_type = F_UNLCK;
goto again;
}
struct file_lock *file_lock = locks_alloc_lock();
struct flock64 flock;
struct inode *inode;
+ struct file *f;
int error;
if (file_lock == NULL)
* Attempt to detect a close/fcntl race and recover by
* releasing the lock that was just acquired.
*/
- if (!error && fcheck(fd) != filp && flock.l_type != F_UNLCK) {
+ spin_lock(¤t->files->file_lock);
+ f = fcheck(fd);
+ spin_unlock(¤t->files->file_lock);
+ if (!error && f != filp && flock.l_type != F_UNLCK) {
flock.l_type = F_UNLCK;
goto again;
}
# define mac_scc ((*(volatile struct SCC*)MAC_SCC_BAS))
#endif
-/* hardware stuff */
-
-#define MACHW_DECLARE(name) unsigned name : 1
-#define MACHW_SET(name) (mac_hw_present.name = 1)
-#define MACHW_PRESENT(name) (mac_hw_present.name)
-
-struct mac_hw_present {
- /* video hardware */
- /* sound hardware */
- /* disk storage interfaces */
- MACHW_DECLARE(MAC_SCSI_80); /* Directly mapped NCR5380 */
- MACHW_DECLARE(MAC_SCSI_96); /* 53c9[46] */
- MACHW_DECLARE(MAC_SCSI_96_2); /* 2nd 53c9[46] Q900 and Q950 */
- MACHW_DECLARE(IDE); /* IDE Interface */
- /* other I/O hardware */
- MACHW_DECLARE(SCC); /* Serial Communications Contr. */
- /* DMA */
- MACHW_DECLARE(SCSI_DMA); /* DMA for the NCR5380 */
- /* real time clocks */
- MACHW_DECLARE(RTC_CLK); /* clock chip */
- /* supporting hardware */
- MACHW_DECLARE(VIA1); /* Versatile Interface Ad. 1 */
- MACHW_DECLARE(VIA2); /* Versatile Interface Ad. 2 */
- MACHW_DECLARE(RBV); /* Versatile Interface Ad. 2+ */
- /* NUBUS */
- MACHW_DECLARE(NUBUS); /* NUBUS */
-};
-
-extern struct mac_hw_present mac_hw_present;
-
#endif /* __ASSEMBLY__ */
#endif /* linux/machw.h */
#define _ASM_POWERPC_IO_H
#ifdef __KERNEL__
-/*
+/*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
#define _PNPWRP 0xa79
#define PNPBIOS_BASE 0xf000
+#include <linux/device.h>
+#include <linux/io.h>
+
#include <linux/compiler.h>
#include <asm/page.h>
#include <asm/byteorder.h>
#define clrsetbits_8(addr, clear, set) clrsetbits(8, addr, clear, set)
+void __iomem *devm_ioremap_prot(struct device *dev, resource_size_t offset,
+ size_t size, unsigned long flags);
+
#endif /* __KERNEL__ */
#endif /* _ASM_POWERPC_IO_H */
/* To avoid harmful races, pmd_none(x) should check only the lower when PAE */
#define pmd_none(x) (!(unsigned long)pmd_val((x)))
#define pmd_present(x) (pmd_val((x)) & _PAGE_PRESENT)
-
-extern int pmd_bad(pmd_t pmd);
-
-#define pmd_bad_v1(x) \
- (_KERNPG_TABLE != (pmd_val((x)) & ~(PAGE_MASK | _PAGE_USER)))
-#define pmd_bad_v2(x) \
- (_KERNPG_TABLE != (pmd_val((x)) & ~(PAGE_MASK | _PAGE_USER | \
- _PAGE_PSE | _PAGE_NX)))
+#define pmd_bad(x) ((pmd_val(x) & (~PAGE_MASK & ~_PAGE_USER)) != _KERNPG_TABLE)
#define pages_to_mb(x) ((x) >> (20-PAGE_SHIFT))
static inline unsigned long pud_bad(pud_t pud)
{
- return pud_val(pud) &
- ~(PTE_MASK | _KERNPG_TABLE | _PAGE_USER | _PAGE_PSE | _PAGE_NX);
+ return pud_val(pud) & ~(PTE_MASK | _KERNPG_TABLE | _PAGE_USER);
}
static inline unsigned long pmd_bad(pmd_t pmd)
{
- return pmd_val(pmd) &
- ~(PTE_MASK | _KERNPG_TABLE | _PAGE_USER | _PAGE_PSE | _PAGE_NX);
+ return pmd_val(pmd) & ~(PTE_MASK | _KERNPG_TABLE | _PAGE_USER);
}
#define pte_none(x) (!pte_val((x)))
extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
unsigned long, loff_t *);
-/*
- * NOTE: write_inode, delete_inode, clear_inode, put_inode can be called
- * without the big kernel lock held in all filesystems.
- */
struct super_operations {
struct inode *(*alloc_inode)(struct super_block *sb);
void (*destroy_inode)(struct inode *);
void (*dirty_inode) (struct inode *);
int (*write_inode) (struct inode *, int);
- void (*put_inode) (struct inode *);
void (*drop_inode) (struct inode *);
void (*delete_inode) (struct inode *);
void (*put_super) (struct super_block *);
void devm_iounmap(struct device *dev, void __iomem *addr);
int check_signature(const volatile void __iomem *io_addr,
const unsigned char *signature, int length);
+void devm_ioremap_release(struct device *dev, void *res);
#endif /* _LINUX_IO_H */
extern void ata_eh_qc_complete(struct ata_queued_cmd *qc);
extern void ata_eh_qc_retry(struct ata_queued_cmd *qc);
+extern void ata_eh_analyze_ncq_error(struct ata_link *link);
extern void ata_do_eh(struct ata_port *ap, ata_prereset_fn_t prereset,
ata_reset_fn_t softreset, ata_reset_fn_t hardreset,
return *(struct ata_port **)&host->hostdata[0];
}
+static inline int ata_check_ready(u8 status)
+{
+ /* Some controllers report 0x77 or 0x7f during intermediate
+ * not-ready stages.
+ */
+ if (status == 0x77 || status == 0x7f)
+ return 0;
+
+ /* 0xff indicates either no device or device not ready */
+ if (status == 0xff)
+ return -ENODEV;
+
+ return !(status & ATA_BUSY);
+}
+
/**************************************************************************
* PMP - drivers/ata/libata-pmp.c
#include <linux/mod_devicetable.h>
#include <linux/types.h>
+#include <linux/init.h>
#include <linux/ioport.h>
#include <linux/list.h>
#include <linux/compiler.h>
void pci_bus_add_devices(struct pci_bus *bus);
struct pci_bus *pci_scan_bus_parented(struct device *parent, int bus,
struct pci_ops *ops, void *sysdata);
-static inline struct pci_bus *pci_scan_bus(int bus, struct pci_ops *ops,
+static inline struct pci_bus * __devinit pci_scan_bus(int bus, struct pci_ops *ops,
void *sysdata)
{
struct pci_bus *root_bus;
void pci_walk_bus(struct pci_bus *top, void (*cb)(struct pci_dev *, void *),
void *userdata);
-int pci_cfg_space_size_ext(struct pci_dev *dev, unsigned check_exp_pcix);
+int pci_cfg_space_size_ext(struct pci_dev *dev);
int pci_cfg_space_size(struct pci_dev *dev);
unsigned char pci_bus_max_busnr(struct pci_bus *bus);
}
#endif
+extern unsigned long long time_sync_thresh;
+
/*
* Task state bitmask. NOTE! These bits are also
* encoded in fs/proc/array.c: get_task_state().
extern unsigned long long sched_clock(void);
+#ifndef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
+static inline void sched_clock_init(void)
+{
+}
+
+static inline u64 sched_clock_cpu(int cpu)
+{
+ return sched_clock();
+}
+
+static inline void sched_clock_tick(void)
+{
+}
+
+static inline void sched_clock_idle_sleep_event(void)
+{
+}
+
+static inline void sched_clock_idle_wakeup_event(u64 delta_ns)
+{
+}
+#else
+extern void sched_clock_init(void);
+extern u64 sched_clock_cpu(int cpu);
+extern void sched_clock_tick(void);
+extern void sched_clock_idle_sleep_event(void);
+extern void sched_clock_idle_wakeup_event(u64 delta_ns);
+#endif
+
/*
* For kernel-internal use: high-speed (but slightly incorrect) per-cpu
* clock constructed from sched_clock():
clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
}
+static inline int test_tsk_need_resched(struct task_struct *tsk)
+{
+ return unlikely(test_tsk_thread_flag(tsk,TIF_NEED_RESCHED));
+}
+
static inline int signal_pending(struct task_struct *p)
{
return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
static inline int need_resched(void)
{
- return unlikely(test_thread_flag(TIF_NEED_RESCHED));
+ return unlikely(test_tsk_need_resched(current));
}
/*
{
return (inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
(inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
- !(dst_metric(dst, RTAX_LOCK)&(1<<RTAX_MTU))));
+ !(dst_metric_locked(dst, RTAX_MTU))));
}
extern void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
extern void xfrm_audit_state_icvfail(struct xfrm_state *x,
struct sk_buff *skb, u8 proto);
#else
-#define xfrm_audit_policy_add(x, r, a, se, s) do { ; } while (0)
-#define xfrm_audit_policy_delete(x, r, a, se, s) do { ; } while (0)
-#define xfrm_audit_state_add(x, r, a, se, s) do { ; } while (0)
-#define xfrm_audit_state_delete(x, r, a, se, s) do { ; } while (0)
-#define xfrm_audit_state_replay_overflow(x, s) do { ; } while (0)
-#define xfrm_audit_state_notfound_simple(s, f) do { ; } while (0)
-#define xfrm_audit_state_notfound(s, f, sp, sq) do { ; } while (0)
-#define xfrm_audit_state_icvfail(x, s, p) do { ; } while (0)
+
+static inline void xfrm_audit_policy_add(struct xfrm_policy *xp, int result,
+ u32 auid, u32 ses, u32 secid)
+{
+}
+
+static inline void xfrm_audit_policy_delete(struct xfrm_policy *xp, int result,
+ u32 auid, u32 ses, u32 secid)
+{
+}
+
+static inline void xfrm_audit_state_add(struct xfrm_state *x, int result,
+ u32 auid, u32 ses, u32 secid)
+{
+}
+
+static inline void xfrm_audit_state_delete(struct xfrm_state *x, int result,
+ u32 auid, u32 ses, u32 secid)
+{
+}
+
+static inline void xfrm_audit_state_replay_overflow(struct xfrm_state *x,
+ struct sk_buff *skb)
+{
+}
+
+static inline void xfrm_audit_state_notfound_simple(struct sk_buff *skb,
+ u16 family)
+{
+}
+
+static inline void xfrm_audit_state_notfound(struct sk_buff *skb, u16 family,
+ __be32 net_spi, __be32 net_seq)
+{
+}
+
+static inline void xfrm_audit_state_icvfail(struct xfrm_state *x,
+ struct sk_buff *skb, u8 proto)
+{
+}
#endif /* CONFIG_AUDITSYSCALL */
static inline void xfrm_pol_hold(struct xfrm_policy *policy)
Say N if unsure.
+#
+# Architectures with an unreliable sched_clock() should select this:
+#
+config HAVE_UNSTABLE_SCHED_CLOCK
+ bool
+
config GROUP_SCHED
bool "Group CPU scheduler"
- default y
+ depends on EXPERIMENTAL
+ default n
help
This feature lets CPU scheduler recognize task groups and control CPU
bandwidth allocation to such task groups.
config FAIR_GROUP_SCHED
bool "Group scheduling for SCHED_OTHER"
depends on GROUP_SCHED
- default y
+ default GROUP_SCHED
config RT_GROUP_SCHED
bool "Group scheduling for SCHED_RR/FIFO"
softirq_init();
timekeeping_init();
time_init();
+ sched_clock_init();
profile_init();
if (!irqs_disabled())
printk("start_kernel(): bug: interrupts were enabled early\n");
rcupdate.o extable.o params.o posix-timers.o \
kthread.o wait.o kfifo.o sys_ni.o posix-cpu-timers.o mutex.o \
hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \
- notifier.o ksysfs.o pm_qos_params.o
+ notifier.o ksysfs.o pm_qos_params.o sched_clock.o
obj-$(CONFIG_SYSCTL_SYSCALL_CHECK) += sysctl_check.o
obj-$(CONFIG_STACKTRACE) += stacktrace.o
#include <asm/tlb.h>
#include <asm/irq_regs.h>
-/*
- * Scheduler clock - returns current time in nanosec units.
- * This is default implementation.
- * Architectures and sub-architectures can override this.
- */
-unsigned long long __attribute__((weak)) sched_clock(void)
-{
- return (unsigned long long)jiffies * (NSEC_PER_SEC / HZ);
-}
-
/*
* Convert user-nice values [ -20 ... 0 ... 19 ]
* to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ],
}
#endif
+/*
+ * sched_domains_mutex serializes calls to arch_init_sched_domains,
+ * detach_destroy_domains and partition_sched_domains.
+ */
+static DEFINE_MUTEX(sched_domains_mutex);
+
#ifdef CONFIG_GROUP_SCHED
#include <linux/cgroup.h>
*/
static DEFINE_SPINLOCK(task_group_lock);
-/* doms_cur_mutex serializes access to doms_cur[] array */
-static DEFINE_MUTEX(doms_cur_mutex);
-
#ifdef CONFIG_FAIR_GROUP_SCHED
#ifdef CONFIG_USER_SCHED
# define INIT_TASK_GROUP_LOAD (2*NICE_0_LOAD)
# define INIT_TASK_GROUP_LOAD NICE_0_LOAD
#endif
+/*
+ * A weight of 0, 1 or ULONG_MAX can cause arithmetics problems.
+ * (The default weight is 1024 - so there's no practical
+ * limitation from this.)
+ */
#define MIN_SHARES 2
+#define MAX_SHARES (ULONG_MAX - 1)
static int init_task_group_load = INIT_TASK_GROUP_LOAD;
#endif
#endif
}
-static inline void lock_doms_cur(void)
-{
- mutex_lock(&doms_cur_mutex);
-}
-
-static inline void unlock_doms_cur(void)
-{
- mutex_unlock(&doms_cur_mutex);
-}
-
#else
static inline void set_task_rq(struct task_struct *p, unsigned int cpu) { }
-static inline void lock_doms_cur(void) { }
-static inline void unlock_doms_cur(void) { }
#endif /* CONFIG_GROUP_SCHED */
unsigned long next_balance;
struct mm_struct *prev_mm;
- u64 clock, prev_clock_raw;
- s64 clock_max_delta;
-
- unsigned int clock_warps, clock_overflows, clock_underflows;
- u64 idle_clock;
- unsigned int clock_deep_idle_events;
- u64 tick_timestamp;
+ u64 clock;
atomic_t nr_iowait;
#endif
}
-#ifdef CONFIG_NO_HZ
-static inline bool nohz_on(int cpu)
-{
- return tick_get_tick_sched(cpu)->nohz_mode != NOHZ_MODE_INACTIVE;
-}
-
-static inline u64 max_skipped_ticks(struct rq *rq)
-{
- return nohz_on(cpu_of(rq)) ? jiffies - rq->last_tick_seen + 2 : 1;
-}
-
-static inline void update_last_tick_seen(struct rq *rq)
-{
- rq->last_tick_seen = jiffies;
-}
-#else
-static inline u64 max_skipped_ticks(struct rq *rq)
-{
- return 1;
-}
-
-static inline void update_last_tick_seen(struct rq *rq)
-{
-}
-#endif
-
-/*
- * Update the per-runqueue clock, as finegrained as the platform can give
- * us, but without assuming monotonicity, etc.:
- */
-static void __update_rq_clock(struct rq *rq)
-{
- u64 prev_raw = rq->prev_clock_raw;
- u64 now = sched_clock();
- s64 delta = now - prev_raw;
- u64 clock = rq->clock;
-
-#ifdef CONFIG_SCHED_DEBUG
- WARN_ON_ONCE(cpu_of(rq) != smp_processor_id());
-#endif
- /*
- * Protect against sched_clock() occasionally going backwards:
- */
- if (unlikely(delta < 0)) {
- clock++;
- rq->clock_warps++;
- } else {
- /*
- * Catch too large forward jumps too:
- */
- u64 max_jump = max_skipped_ticks(rq) * TICK_NSEC;
- u64 max_time = rq->tick_timestamp + max_jump;
-
- if (unlikely(clock + delta > max_time)) {
- if (clock < max_time)
- clock = max_time;
- else
- clock++;
- rq->clock_overflows++;
- } else {
- if (unlikely(delta > rq->clock_max_delta))
- rq->clock_max_delta = delta;
- clock += delta;
- }
- }
-
- rq->prev_clock_raw = now;
- rq->clock = clock;
-}
-
-static void update_rq_clock(struct rq *rq)
-{
- if (likely(smp_processor_id() == cpu_of(rq)))
- __update_rq_clock(rq);
-}
-
/*
* The domain tree (rq->sd) is protected by RCU's quiescent state transition.
* See detach_destroy_domains: synchronize_sched for details.
#define task_rq(p) cpu_rq(task_cpu(p))
#define cpu_curr(cpu) (cpu_rq(cpu)->curr)
+static inline void update_rq_clock(struct rq *rq)
+{
+ rq->clock = sched_clock_cpu(cpu_of(rq));
+}
+
/*
* Tunables that become constants when CONFIG_SCHED_DEBUG is off:
*/
#define SCHED_FEAT(name, enabled) \
#name ,
-__read_mostly char *sched_feat_names[] = {
+static __read_mostly char *sched_feat_names[] = {
#include "sched_features.h"
NULL
};
#undef SCHED_FEAT
-int sched_feat_open(struct inode *inode, struct file *filp)
+static int sched_feat_open(struct inode *inode, struct file *filp)
{
filp->private_data = inode->i_private;
return 0;
return (u64)sysctl_sched_rt_runtime * NSEC_PER_USEC;
}
-static const unsigned long long time_sync_thresh = 100000;
+unsigned long long time_sync_thresh = 100000;
static DEFINE_PER_CPU(unsigned long long, time_offset);
static DEFINE_PER_CPU(unsigned long long, prev_cpu_time);
static DEFINE_SPINLOCK(time_sync_lock);
static unsigned long long prev_global_time;
-static unsigned long long __sync_cpu_clock(cycles_t time, int cpu)
+static unsigned long long __sync_cpu_clock(unsigned long long time, int cpu)
{
- unsigned long flags;
-
- spin_lock_irqsave(&time_sync_lock, flags);
+ /*
+ * We want this inlined, to not get tracer function calls
+ * in this critical section:
+ */
+ spin_acquire(&time_sync_lock.dep_map, 0, 0, _THIS_IP_);
+ __raw_spin_lock(&time_sync_lock.raw_lock);
if (time < prev_global_time) {
per_cpu(time_offset, cpu) += prev_global_time - time;
prev_global_time = time;
}
- spin_unlock_irqrestore(&time_sync_lock, flags);
+ __raw_spin_unlock(&time_sync_lock.raw_lock);
+ spin_release(&time_sync_lock.dep_map, 1, _THIS_IP_);
return time;
}
static unsigned long long __cpu_clock(int cpu)
{
unsigned long long now;
- unsigned long flags;
- struct rq *rq;
/*
* Only call sched_clock() if the scheduler has already been
if (unlikely(!scheduler_running))
return 0;
- local_irq_save(flags);
- rq = cpu_rq(cpu);
- update_rq_clock(rq);
- now = rq->clock;
- local_irq_restore(flags);
+ now = sched_clock_cpu(cpu);
return now;
}
unsigned long long cpu_clock(int cpu)
{
unsigned long long prev_cpu_time, time, delta_time;
+ unsigned long flags;
+ local_irq_save(flags);
prev_cpu_time = per_cpu(prev_cpu_time, cpu);
time = __cpu_clock(cpu) + per_cpu(time_offset, cpu);
delta_time = time-prev_cpu_time;
- if (unlikely(delta_time > time_sync_thresh))
+ if (unlikely(delta_time > time_sync_thresh)) {
time = __sync_cpu_clock(time, cpu);
+ per_cpu(prev_cpu_time, cpu) = time;
+ }
+ local_irq_restore(flags);
return time;
}
return rq;
}
-/*
- * We are going deep-idle (irqs are disabled):
- */
-void sched_clock_idle_sleep_event(void)
-{
- struct rq *rq = cpu_rq(smp_processor_id());
-
- spin_lock(&rq->lock);
- __update_rq_clock(rq);
- spin_unlock(&rq->lock);
- rq->clock_deep_idle_events++;
-}
-EXPORT_SYMBOL_GPL(sched_clock_idle_sleep_event);
-
-/*
- * We just idled delta nanoseconds (called with irqs disabled):
- */
-void sched_clock_idle_wakeup_event(u64 delta_ns)
-{
- struct rq *rq = cpu_rq(smp_processor_id());
- u64 now = sched_clock();
-
- rq->idle_clock += delta_ns;
- /*
- * Override the previous timestamp and ignore all
- * sched_clock() deltas that occured while we idled,
- * and use the PM-provided delta_ns to advance the
- * rq clock:
- */
- spin_lock(&rq->lock);
- rq->prev_clock_raw = now;
- rq->clock += delta_ns;
- spin_unlock(&rq->lock);
- touch_softlockup_watchdog();
-}
-EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event);
-
static void __resched_task(struct task_struct *p, int tif_bit);
static inline void resched_task(struct task_struct *p)
enum {
HRTICK_SET, /* re-programm hrtick_timer */
HRTICK_RESET, /* not a new slice */
+ HRTICK_BLOCK, /* stop hrtick operations */
};
/*
{
if (!sched_feat(HRTICK))
return 0;
+ if (unlikely(test_bit(HRTICK_BLOCK, &rq->hrtick_flags)))
+ return 0;
return hrtimer_is_hres_active(&rq->hrtick_timer);
}
WARN_ON_ONCE(cpu_of(rq) != smp_processor_id());
spin_lock(&rq->lock);
- __update_rq_clock(rq);
+ update_rq_clock(rq);
rq->curr->sched_class->task_tick(rq, rq->curr, 1);
spin_unlock(&rq->lock);
return HRTIMER_NORESTART;
}
-static inline void init_rq_hrtick(struct rq *rq)
+static void hotplug_hrtick_disable(int cpu)
+{
+ struct rq *rq = cpu_rq(cpu);
+ unsigned long flags;
+
+ spin_lock_irqsave(&rq->lock, flags);
+ rq->hrtick_flags = 0;
+ __set_bit(HRTICK_BLOCK, &rq->hrtick_flags);
+ spin_unlock_irqrestore(&rq->lock, flags);
+
+ hrtick_clear(rq);
+}
+
+static void hotplug_hrtick_enable(int cpu)
+{
+ struct rq *rq = cpu_rq(cpu);
+ unsigned long flags;
+
+ spin_lock_irqsave(&rq->lock, flags);
+ __clear_bit(HRTICK_BLOCK, &rq->hrtick_flags);
+ spin_unlock_irqrestore(&rq->lock, flags);
+}
+
+static int
+hotplug_hrtick(struct notifier_block *nfb, unsigned long action, void *hcpu)
+{
+ int cpu = (int)(long)hcpu;
+
+ switch (action) {
+ case CPU_UP_CANCELED:
+ case CPU_UP_CANCELED_FROZEN:
+ case CPU_DOWN_PREPARE:
+ case CPU_DOWN_PREPARE_FROZEN:
+ case CPU_DEAD:
+ case CPU_DEAD_FROZEN:
+ hotplug_hrtick_disable(cpu);
+ return NOTIFY_OK;
+
+ case CPU_UP_PREPARE:
+ case CPU_UP_PREPARE_FROZEN:
+ case CPU_DOWN_FAILED:
+ case CPU_DOWN_FAILED_FROZEN:
+ case CPU_ONLINE:
+ case CPU_ONLINE_FROZEN:
+ hotplug_hrtick_enable(cpu);
+ return NOTIFY_OK;
+ }
+
+ return NOTIFY_DONE;
+}
+
+static void init_hrtick(void)
+{
+ hotcpu_notifier(hotplug_hrtick, 0);
+}
+
+static void init_rq_hrtick(struct rq *rq)
{
rq->hrtick_flags = 0;
hrtimer_init(&rq->hrtick_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
void hrtick_resched(void)
{
}
+
+static inline void init_hrtick(void)
+{
+}
#endif
/*
{
u64 tmp;
- if (unlikely(!lw->inv_weight))
- lw->inv_weight = (WMULT_CONST-lw->weight/2) / (lw->weight+1);
+ if (!lw->inv_weight)
+ lw->inv_weight = 1 + (WMULT_CONST-lw->weight/2)/(lw->weight+1);
tmp = (u64)delta_exec * weight;
/*
if (shares < MIN_SHARES)
shares = MIN_SHARES;
+ else if (shares > MAX_SHARES)
+ shares = MAX_SHARES;
__set_se_shares(tg->se[tcpu], shares);
}
struct rq *rq = this_rq();
cputime64_t tmp;
- if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0))
- return account_guest_time(p, cputime);
+ if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0)) {
+ account_guest_time(p, cputime);
+ return;
+ }
p->stime = cputime_add(p->stime, cputime);
int cpu = smp_processor_id();
struct rq *rq = cpu_rq(cpu);
struct task_struct *curr = rq->curr;
- u64 next_tick = rq->tick_timestamp + TICK_NSEC;
+
+ sched_clock_tick();
spin_lock(&rq->lock);
- __update_rq_clock(rq);
- /*
- * Let rq->clock advance by at least TICK_NSEC:
- */
- if (unlikely(rq->clock < next_tick)) {
- rq->clock = next_tick;
- rq->clock_underflows++;
- }
- rq->tick_timestamp = rq->clock;
- update_last_tick_seen(rq);
+ update_rq_clock(rq);
update_cpu_load(rq);
curr->sched_class->task_tick(rq, curr, 0);
spin_unlock(&rq->lock);
* Do the rq-clock update outside the rq lock:
*/
local_irq_disable();
- __update_rq_clock(rq);
+ update_rq_clock(rq);
spin_lock(&rq->lock);
clear_tsk_need_resched(prev);
prev->sched_class->put_prev_task(rq, prev);
next = pick_next_task(rq, prev);
- sched_info_switch(prev, next);
-
if (likely(prev != next)) {
+ sched_info_switch(prev, next);
+
rq->nr_switches++;
rq->curr = next;
++*switch_count;
{
int i, j;
- lock_doms_cur();
+ mutex_lock(&sched_domains_mutex);
/* always unregister in case we don't destroy any domains */
unregister_sched_domain_sysctl();
register_sched_domain_sysctl();
- unlock_doms_cur();
+ mutex_unlock(&sched_domains_mutex);
}
#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT)
int err;
get_online_cpus();
+ mutex_lock(&sched_domains_mutex);
detach_destroy_domains(&cpu_online_map);
err = arch_init_sched_domains(&cpu_online_map);
+ mutex_unlock(&sched_domains_mutex);
put_online_cpus();
return err;
BUG_ON(sched_group_nodes_bycpu == NULL);
#endif
get_online_cpus();
+ mutex_lock(&sched_domains_mutex);
arch_init_sched_domains(&cpu_online_map);
cpus_andnot(non_isolated_cpus, cpu_possible_map, cpu_isolated_map);
if (cpus_empty(non_isolated_cpus))
cpu_set(smp_processor_id(), non_isolated_cpus);
+ mutex_unlock(&sched_domains_mutex);
put_online_cpus();
/* XXX: Theoretical race here - CPU may be hotplugged now */
hotcpu_notifier(update_sched_domains, 0);
+ init_hrtick();
/* Move init over to a non-isolated CPU */
if (set_cpus_allowed_ptr(current, &non_isolated_cpus) < 0)
se->my_q = cfs_rq;
se->load.weight = tg->shares;
- se->load.inv_weight = div64_u64(1ULL<<32, se->load.weight);
+ se->load.inv_weight = 0;
se->parent = parent;
}
#endif
spin_lock_init(&rq->lock);
lockdep_set_class(&rq->lock, &rq->rq_lock_key);
rq->nr_running = 0;
- rq->clock = 1;
- update_last_tick_seen(rq);
init_cfs_rq(&rq->cfs, rq);
init_rt_rq(&rq->rt, rq);
#ifdef CONFIG_FAIR_GROUP_SCHED
static void normalize_task(struct rq *rq, struct task_struct *p)
{
int on_rq;
+
update_rq_clock(rq);
on_rq = p->se.on_rq;
if (on_rq)
p->se.sleep_start = 0;
p->se.block_start = 0;
#endif
- task_rq(p)->clock = 0;
if (!rt_task(p)) {
/*
dequeue_entity(cfs_rq, se, 0);
se->load.weight = shares;
- se->load.inv_weight = div64_u64((1ULL<<32), shares);
+ se->load.inv_weight = 0;
if (on_rq)
enqueue_entity(cfs_rq, se, 0);
if (!tg->se[0])
return -EINVAL;
- /*
- * A weight of 0 or 1 can cause arithmetics problems.
- * (The default weight is 1024 - so there's no practical
- * limitation from this.)
- */
if (shares < MIN_SHARES)
shares = MIN_SHARES;
+ else if (shares > MAX_SHARES)
+ shares = MAX_SHARES;
mutex_lock(&shares_mutex);
if (tg->shares == shares)
* force a rebalance
*/
cfs_rq_set_shares(tg->cfs_rq[i], 0);
- set_se_shares(tg->se[i], shares/nr_cpu_ids);
+ set_se_shares(tg->se[i], shares);
}
/*
--- /dev/null
+/*
+ * sched_clock for unstable cpu clocks
+ *
+ * Copyright (C) 2008 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
+ *
+ * Based on code by:
+ * Ingo Molnar <mingo@redhat.com>
+ * Guillaume Chazarain <guichaz@gmail.com>
+ *
+ * Create a semi stable clock from a mixture of other events, including:
+ * - gtod
+ * - jiffies
+ * - sched_clock()
+ * - explicit idle events
+ *
+ * We use gtod as base and the unstable clock deltas. The deltas are filtered,
+ * making it monotonic and keeping it within an expected window. This window
+ * is set up using jiffies.
+ *
+ * Furthermore, explicit sleep and wakeup hooks allow us to account for time
+ * that is otherwise invisible (TSC gets stopped).
+ *
+ * The clock: sched_clock_cpu() is monotonic per cpu, and should be somewhat
+ * consistent between cpus (never more than 1 jiffies difference).
+ */
+#include <linux/sched.h>
+#include <linux/percpu.h>
+#include <linux/spinlock.h>
+#include <linux/ktime.h>
+#include <linux/module.h>
+
+
+#ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK
+
+struct sched_clock_data {
+ /*
+ * Raw spinlock - this is a special case: this might be called
+ * from within instrumentation code so we dont want to do any
+ * instrumentation ourselves.
+ */
+ raw_spinlock_t lock;
+
+ unsigned long prev_jiffies;
+ u64 prev_raw;
+ u64 tick_raw;
+ u64 tick_gtod;
+ u64 clock;
+};
+
+static DEFINE_PER_CPU_SHARED_ALIGNED(struct sched_clock_data, sched_clock_data);
+
+static inline struct sched_clock_data *this_scd(void)
+{
+ return &__get_cpu_var(sched_clock_data);
+}
+
+static inline struct sched_clock_data *cpu_sdc(int cpu)
+{
+ return &per_cpu(sched_clock_data, cpu);
+}
+
+void sched_clock_init(void)
+{
+ u64 ktime_now = ktime_to_ns(ktime_get());
+ u64 now = 0;
+ int cpu;
+
+ for_each_possible_cpu(cpu) {
+ struct sched_clock_data *scd = cpu_sdc(cpu);
+
+ scd->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
+ scd->prev_jiffies = jiffies;
+ scd->prev_raw = now;
+ scd->tick_raw = now;
+ scd->tick_gtod = ktime_now;
+ scd->clock = ktime_now;
+ }
+}
+
+/*
+ * update the percpu scd from the raw @now value
+ *
+ * - filter out backward motion
+ * - use jiffies to generate a min,max window to clip the raw values
+ */
+static void __update_sched_clock(struct sched_clock_data *scd, u64 now)
+{
+ unsigned long now_jiffies = jiffies;
+ long delta_jiffies = now_jiffies - scd->prev_jiffies;
+ u64 clock = scd->clock;
+ u64 min_clock, max_clock;
+ s64 delta = now - scd->prev_raw;
+
+ WARN_ON_ONCE(!irqs_disabled());
+ min_clock = scd->tick_gtod + delta_jiffies * TICK_NSEC;
+
+ if (unlikely(delta < 0)) {
+ clock++;
+ goto out;
+ }
+
+ max_clock = min_clock + TICK_NSEC;
+
+ if (unlikely(clock + delta > max_clock)) {
+ if (clock < max_clock)
+ clock = max_clock;
+ else
+ clock++;
+ } else {
+ clock += delta;
+ }
+
+ out:
+ if (unlikely(clock < min_clock))
+ clock = min_clock;
+
+ scd->prev_raw = now;
+ scd->prev_jiffies = now_jiffies;
+ scd->clock = clock;
+}
+
+static void lock_double_clock(struct sched_clock_data *data1,
+ struct sched_clock_data *data2)
+{
+ if (data1 < data2) {
+ __raw_spin_lock(&data1->lock);
+ __raw_spin_lock(&data2->lock);
+ } else {
+ __raw_spin_lock(&data2->lock);
+ __raw_spin_lock(&data1->lock);
+ }
+}
+
+u64 sched_clock_cpu(int cpu)
+{
+ struct sched_clock_data *scd = cpu_sdc(cpu);
+ u64 now, clock;
+
+ WARN_ON_ONCE(!irqs_disabled());
+ now = sched_clock();
+
+ if (cpu != raw_smp_processor_id()) {
+ /*
+ * in order to update a remote cpu's clock based on our
+ * unstable raw time rebase it against:
+ * tick_raw (offset between raw counters)
+ * tick_gotd (tick offset between cpus)
+ */
+ struct sched_clock_data *my_scd = this_scd();
+
+ lock_double_clock(scd, my_scd);
+
+ now -= my_scd->tick_raw;
+ now += scd->tick_raw;
+
+ now -= my_scd->tick_gtod;
+ now += scd->tick_gtod;
+
+ __raw_spin_unlock(&my_scd->lock);
+ } else {
+ __raw_spin_lock(&scd->lock);
+ }
+
+ __update_sched_clock(scd, now);
+ clock = scd->clock;
+
+ __raw_spin_unlock(&scd->lock);
+
+ return clock;
+}
+
+void sched_clock_tick(void)
+{
+ struct sched_clock_data *scd = this_scd();
+ u64 now, now_gtod;
+
+ WARN_ON_ONCE(!irqs_disabled());
+
+ now = sched_clock();
+ now_gtod = ktime_to_ns(ktime_get());
+
+ __raw_spin_lock(&scd->lock);
+ __update_sched_clock(scd, now);
+ /*
+ * update tick_gtod after __update_sched_clock() because that will
+ * already observe 1 new jiffy; adding a new tick_gtod to that would
+ * increase the clock 2 jiffies.
+ */
+ scd->tick_raw = now;
+ scd->tick_gtod = now_gtod;
+ __raw_spin_unlock(&scd->lock);
+}
+
+/*
+ * We are going deep-idle (irqs are disabled):
+ */
+void sched_clock_idle_sleep_event(void)
+{
+ sched_clock_cpu(smp_processor_id());
+}
+EXPORT_SYMBOL_GPL(sched_clock_idle_sleep_event);
+
+/*
+ * We just idled delta nanoseconds (called with irqs disabled):
+ */
+void sched_clock_idle_wakeup_event(u64 delta_ns)
+{
+ struct sched_clock_data *scd = this_scd();
+ u64 now = sched_clock();
+
+ /*
+ * Override the previous timestamp and ignore all
+ * sched_clock() deltas that occured while we idled,
+ * and use the PM-provided delta_ns to advance the
+ * rq clock:
+ */
+ __raw_spin_lock(&scd->lock);
+ scd->prev_raw = now;
+ scd->clock += delta_ns;
+ __raw_spin_unlock(&scd->lock);
+
+ touch_softlockup_watchdog();
+}
+EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event);
+
+#endif
+
+/*
+ * Scheduler clock - returns current time in nanosec units.
+ * This is default implementation.
+ * Architectures and sub-architectures can override this.
+ */
+unsigned long long __attribute__((weak)) sched_clock(void)
+{
+ return (unsigned long long)jiffies * (NSEC_PER_SEC / HZ);
+}
PN(next_balance);
P(curr->pid);
PN(clock);
- PN(idle_clock);
- PN(prev_clock_raw);
- P(clock_warps);
- P(clock_overflows);
- P(clock_underflows);
- P(clock_deep_idle_events);
- PN(clock_max_delta);
P(cpu_load[0]);
P(cpu_load[1]);
P(cpu_load[2]);
* Update run-time statistics of the 'current'.
*/
update_curr(cfs_rq);
+ account_entity_enqueue(cfs_rq, se);
if (wakeup) {
place_entity(cfs_rq, se, 0);
check_spread(cfs_rq, se);
if (se != cfs_rq->curr)
__enqueue_entity(cfs_rq, se);
- account_entity_enqueue(cfs_rq, se);
}
static void update_avg(u64 *avg, u64 sample)
* queued ticks are scheduled to match the slice, so don't bother
* validating it and just reschedule.
*/
- if (queued)
- return resched_task(rq_of(cfs_rq)->curr);
+ if (queued) {
+ resched_task(rq_of(cfs_rq)->curr);
+ return;
+ }
/*
* don't let the period tick interfere with the hrtick preemption
*/
return;
if (likely(!sysctl_sched_compat_yield) && curr->policy != SCHED_BATCH) {
- __update_rq_clock(rq);
+ update_rq_clock(rq);
/*
* Update run-time statistics of the 'current'.
*/
* sibling runqueue info. This will avoid the checks and cache miss
* penalities associated with that.
*/
- if (idle_cpu(cpu) || cpu_rq(cpu)->nr_running > 1)
+ if (idle_cpu(cpu) || cpu_rq(cpu)->cfs.nr_running > 1)
return cpu;
for_each_domain(cpu, sd) {
};
#ifdef CONFIG_SCHED_DEBUG
-static void
-print_cfs_rq_tasks(struct seq_file *m, struct cfs_rq *cfs_rq, int depth)
-{
- struct sched_entity *se;
-
- if (!cfs_rq)
- return;
-
- list_for_each_entry_rcu(se, &cfs_rq->tasks, group_node) {
- int i;
-
- for (i = depth; i; i--)
- seq_puts(m, " ");
-
- seq_printf(m, "%lu %s %lu\n",
- se->load.weight,
- entity_is_task(se) ? "T" : "G",
- calc_delta_weight(SCHED_LOAD_SCALE, se)
- );
- if (!entity_is_task(se))
- print_cfs_rq_tasks(m, group_cfs_rq(se), depth + 1);
- }
-}
-
static void print_cfs_stats(struct seq_file *m, int cpu)
{
struct cfs_rq *cfs_rq;
rcu_read_lock();
for_each_leaf_cfs_rq(cpu_rq(cpu), cfs_rq)
print_cfs_rq(m, cpu, cfs_rq);
-
- seq_printf(m, "\nWeight tree:\n");
- print_cfs_rq_tasks(m, &cpu_rq(cpu)->cfs, 1);
rcu_read_unlock();
}
#endif
/*
* Simple, special scheduling class for the per-CPU idle tasks:
*/
-const struct sched_class idle_sched_class = {
+static const struct sched_class idle_sched_class = {
/* .next is NULL */
/* no enqueue/yield_task for idle tasks */
}
}
-
+/*
+ * If we are not running and we are not going to reschedule soon, we should
+ * try to push tasks away now
+ */
static void task_wake_up_rt(struct rq *rq, struct task_struct *p)
{
if (!task_running(rq, p) &&
- (p->prio >= rq->rt.highest_prio) &&
+ !test_tsk_need_resched(rq->curr) &&
rq->rt.overloaded)
push_rt_tasks(rq);
}
p->se.exec_start = rq->clock;
}
-const struct sched_class rt_sched_class = {
+static const struct sched_class rt_sched_class = {
.next = &fair_sched_class,
.enqueue_task = enqueue_task_rt,
.dequeue_task = dequeue_task_rt,
#include <linux/io.h>
#include <linux/module.h>
-static void devm_ioremap_release(struct device *dev, void *res)
+void devm_ioremap_release(struct device *dev, void *res)
{
iounmap(*(void __iomem **)res);
}
goto no_page_table;
pmd = pmd_offset(pud, address);
- if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
+ if (pmd_none(*pmd))
goto no_page_table;
if (pmd_huge(*pmd)) {
goto out;
}
+ if (unlikely(pmd_bad(*pmd)))
+ goto no_page_table;
+
ptep = pte_offset_map_lock(mm, pmd, address, &ptl);
if (!ptep)
goto out;
/* skb==NULL means VCC is being destroyed */
br2684_close_vcc(brvcc);
if (list_empty(&brdev->brvccs)) {
- read_lock(&devs_lock);
+ write_lock_irq(&devs_lock);
list_del(&brdev->br2684_devs);
- read_unlock(&devs_lock);
+ write_unlock_irq(&devs_lock);
unregister_netdev(net_dev);
free_netdev(net_dev);
}
rtnl_lock();
if (strchr(dev->name, '%')) {
ret = dev_alloc_name(dev, dev->name);
- if (ret < 0) {
- free_netdev(dev);
- goto out;
- }
+ if (ret < 0)
+ goto out_free;
}
ret = register_netdevice(dev);
if (ret)
- goto out;
+ goto out_free;
ret = br_sysfs_addbr(dev);
if (ret)
out:
rtnl_unlock();
return ret;
+
+out_free:
+ free_netdev(dev);
+ goto out;
}
int br_del_bridge(const char *name)
goto nodata;
/*
- * See comment in sk_buff definition, just before the 'tail' member
+ * Only clear those fields we need to clear, not those that we will
+ * actually initialise below. Hence, don't put any more fields after
+ * the tail pointer in struct sk_buff!
*/
memset(skb, 0, offsetof(struct sk_buff, tail));
skb->truesize = size + sizeof(struct sk_buff);
if (len > 3) {
DCCP_WARN("invalid length %d\n", len);
- return 1;
+ return -EINVAL;
}
/* XXX add further sanity checks */
else
min_mtu -= 21;
- if (dst->metrics[RTAX_MTU-1] > mtu && mtu >= min_mtu) {
+ if (dst_metric(dst, RTAX_MTU) > mtu && mtu >= min_mtu) {
if (!(dst_metric_locked(dst, RTAX_MTU))) {
dst->metrics[RTAX_MTU-1] = mtu;
dst_set_expires(dst, dn_rt_mtu_expires);
}
if (!(dst_metric_locked(dst, RTAX_ADVMSS))) {
u32 mss = mtu - DN_MAX_NSP_DATA_HEADER;
- if (dst->metrics[RTAX_ADVMSS-1] > mss)
+ if (dst_metric(dst, RTAX_ADVMSS) > mss)
dst->metrics[RTAX_ADVMSS-1] = mss;
}
}
rt->u.dst.neighbour = n;
}
- if (rt->u.dst.metrics[RTAX_MTU-1] == 0 ||
- rt->u.dst.metrics[RTAX_MTU-1] > rt->u.dst.dev->mtu)
+ if (dst_metric(&rt->u.dst, RTAX_MTU) == 0 ||
+ dst_metric(&rt->u.dst, RTAX_MTU) > rt->u.dst.dev->mtu)
rt->u.dst.metrics[RTAX_MTU-1] = rt->u.dst.dev->mtu;
mss = dn_mss_from_pmtu(dev, dst_mtu(&rt->u.dst));
- if (rt->u.dst.metrics[RTAX_ADVMSS-1] == 0 ||
- rt->u.dst.metrics[RTAX_ADVMSS-1] > mss)
+ if (dst_metric(&rt->u.dst, RTAX_ADVMSS) == 0 ||
+ dst_metric(&rt->u.dst, RTAX_ADVMSS) > mss)
rt->u.dst.metrics[RTAX_ADVMSS-1] = mss;
return 0;
}
/* BSD 4.2 compatibility hack :-( */
if (mtu == 0 &&
- old_mtu >= rth->u.dst.metrics[RTAX_MTU-1] &&
+ old_mtu >= dst_metric(&rth->u.dst, RTAX_MTU) &&
old_mtu >= 68 + (iph->ihl << 2))
old_mtu -= iph->ihl << 2;
mtu = guess_mtu(old_mtu);
}
- if (mtu <= rth->u.dst.metrics[RTAX_MTU-1]) {
- if (mtu < rth->u.dst.metrics[RTAX_MTU-1]) {
+ if (mtu <= dst_metric(&rth->u.dst, RTAX_MTU)) {
+ if (mtu < dst_metric(&rth->u.dst, RTAX_MTU)) {
dst_confirm(&rth->u.dst);
if (mtu < ip_rt_min_pmtu) {
mtu = ip_rt_min_pmtu;
static void ip_rt_update_pmtu(struct dst_entry *dst, u32 mtu)
{
- if (dst->metrics[RTAX_MTU-1] > mtu && mtu >= 68 &&
+ if (dst_metric(dst, RTAX_MTU) > mtu && mtu >= 68 &&
!(dst_metric_locked(dst, RTAX_MTU))) {
if (mtu < ip_rt_min_pmtu) {
mtu = ip_rt_min_pmtu;
sizeof(rt->u.dst.metrics));
if (fi->fib_mtu == 0) {
rt->u.dst.metrics[RTAX_MTU-1] = rt->u.dst.dev->mtu;
- if (rt->u.dst.metrics[RTAX_LOCK-1] & (1 << RTAX_MTU) &&
+ if (dst_metric_locked(&rt->u.dst, RTAX_MTU) &&
rt->rt_gateway != rt->rt_dst &&
rt->u.dst.dev->mtu > 576)
rt->u.dst.metrics[RTAX_MTU-1] = 576;
} else
rt->u.dst.metrics[RTAX_MTU-1]= rt->u.dst.dev->mtu;
- if (rt->u.dst.metrics[RTAX_HOPLIMIT-1] == 0)
+ if (dst_metric(&rt->u.dst, RTAX_HOPLIMIT) == 0)
rt->u.dst.metrics[RTAX_HOPLIMIT-1] = sysctl_ip_default_ttl;
- if (rt->u.dst.metrics[RTAX_MTU-1] > IP_MAX_MTU)
+ if (dst_metric(&rt->u.dst, RTAX_MTU) > IP_MAX_MTU)
rt->u.dst.metrics[RTAX_MTU-1] = IP_MAX_MTU;
- if (rt->u.dst.metrics[RTAX_ADVMSS-1] == 0)
+ if (dst_metric(&rt->u.dst, RTAX_ADVMSS) == 0)
rt->u.dst.metrics[RTAX_ADVMSS-1] = max_t(unsigned int, rt->u.dst.dev->mtu - 40,
ip_rt_min_advmss);
- if (rt->u.dst.metrics[RTAX_ADVMSS-1] > 65535 - 40)
+ if (dst_metric(&rt->u.dst, RTAX_ADVMSS) > 65535 - 40)
rt->u.dst.metrics[RTAX_ADVMSS-1] = 65535 - 40;
#ifdef CONFIG_NET_CLS_ROUTE
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/sysctl.h>
+#include <net/dst.h>
#include <net/tcp.h>
#include <net/inet_common.h>
#include <linux/ipsec.h>
u32 rto_min = TCP_RTO_MIN;
if (dst && dst_metric_locked(dst, RTAX_RTO_MIN))
- rto_min = dst->metrics[RTAX_RTO_MIN - 1];
+ rto_min = dst_metric(dst, RTAX_RTO_MIN);
return rto_min;
}
dst->metrics[RTAX_RTTVAR - 1] = m;
else
dst->metrics[RTAX_RTTVAR-1] -=
- (dst->metrics[RTAX_RTTVAR-1] - m)>>2;
+ (dst_metric(dst, RTAX_RTTVAR) - m)>>2;
}
if (tp->snd_ssthresh >= 0xFFFF) {
dst->metrics[RTAX_SSTHRESH-1] =
max(tp->snd_cwnd >> 1, tp->snd_ssthresh);
if (!dst_metric_locked(dst, RTAX_CWND))
- dst->metrics[RTAX_CWND-1] = (dst->metrics[RTAX_CWND-1] + tp->snd_cwnd) >> 1;
+ dst->metrics[RTAX_CWND-1] = (dst_metric(dst, RTAX_CWND) + tp->snd_cwnd) >> 1;
} else {
/* Else slow start did not finish, cwnd is non-sense,
ssthresh may be also invalid.
*/
if (!dst_metric_locked(dst, RTAX_CWND))
- dst->metrics[RTAX_CWND-1] = (dst->metrics[RTAX_CWND-1] + tp->snd_ssthresh) >> 1;
- if (dst->metrics[RTAX_SSTHRESH-1] &&
+ dst->metrics[RTAX_CWND-1] = (dst_metric(dst, RTAX_CWND) + tp->snd_ssthresh) >> 1;
+ if (dst_metric(dst, RTAX_SSTHRESH) &&
!dst_metric_locked(dst, RTAX_SSTHRESH) &&
- tp->snd_ssthresh > dst->metrics[RTAX_SSTHRESH-1])
+ tp->snd_ssthresh > dst_metric(dst, RTAX_SSTHRESH))
dst->metrics[RTAX_SSTHRESH-1] = tp->snd_ssthresh;
}
if (!dst_metric_locked(dst, RTAX_REORDERING)) {
- if (dst->metrics[RTAX_REORDERING-1] < tp->reordering &&
+ if (dst_metric(dst, RTAX_REORDERING) < tp->reordering &&
tp->reordering != sysctl_tcp_reordering)
dst->metrics[RTAX_REORDERING-1] = tp->reordering;
}
}
}
- if (rt->u.dst.metrics[RTAX_HOPLIMIT-1] == 0)
+ if (dst_metric(&rt->u.dst, RTAX_HOPLIMIT) == 0)
rt->u.dst.metrics[RTAX_HOPLIMIT-1] = -1;
- if (!rt->u.dst.metrics[RTAX_MTU-1])
+ if (!dst_metric(&rt->u.dst, RTAX_MTU))
rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(dev);
- if (!rt->u.dst.metrics[RTAX_ADVMSS-1])
+ if (!dst_metric(&rt->u.dst, RTAX_ADVMSS))
rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, dst_mtu(&rt->u.dst));
rt->u.dst.dev = dev;
rt->rt6i_idev = idev;
fail_rate:
ieee80211_debugfs_remove_netdev(IEEE80211_DEV_TO_SUB_IF(local->mdev));
unregister_netdevice(local->mdev);
+ local->mdev = NULL;
fail_dev:
rtnl_unlock();
sta_info_stop(local);
debugfs_hw_del(local);
destroy_workqueue(local->hw.workqueue);
fail_workqueue:
- ieee80211_if_free(local->mdev);
- local->mdev = NULL;
+ if (local->mdev != NULL) {
+ ieee80211_if_free(local->mdev);
+ local->mdev = NULL;
+ }
fail_mdev_alloc:
wiphy_unregister(local->hw.wiphy);
return result;
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
- * Authors: Jamal Hadi Salim (2005)
+ * Authors: Jamal Hadi Salim (2005-8)
*
*/
.lock = &simp_lock,
};
+#define SIMP_MAX_DATA 32
static int tcf_simp(struct sk_buff *skb, struct tc_action *a, struct tcf_result *res)
{
struct tcf_defact *d = a->priv;
return ret;
}
-static int alloc_defdata(struct tcf_defact *d, u32 datalen, void *defdata)
+static int alloc_defdata(struct tcf_defact *d, char *defdata)
{
- d->tcfd_defdata = kmemdup(defdata, datalen, GFP_KERNEL);
+ d->tcfd_defdata = kstrndup(defdata, SIMP_MAX_DATA, GFP_KERNEL);
if (unlikely(!d->tcfd_defdata))
return -ENOMEM;
- d->tcfd_datalen = datalen;
+
return 0;
}
-static int realloc_defdata(struct tcf_defact *d, u32 datalen, void *defdata)
+static void reset_policy(struct tcf_defact *d, char *defdata,
+ struct tc_defact *p)
{
- kfree(d->tcfd_defdata);
- return alloc_defdata(d, datalen, defdata);
+ spin_lock_bh(&d->tcf_lock);
+ d->tcf_action = p->action;
+ memset(d->tcfd_defdata, 0, SIMP_MAX_DATA);
+ strlcpy(d->tcfd_defdata, defdata, SIMP_MAX_DATA);
+ spin_unlock_bh(&d->tcf_lock);
}
static const struct nla_policy simple_policy[TCA_DEF_MAX + 1] = {
[TCA_DEF_PARMS] = { .len = sizeof(struct tc_defact) },
+ [TCA_DEF_DATA] = { .type = NLA_STRING, .len = SIMP_MAX_DATA },
};
static int tcf_simp_init(struct nlattr *nla, struct nlattr *est,
struct tc_defact *parm;
struct tcf_defact *d;
struct tcf_common *pc;
- void *defdata;
- u32 datalen = 0;
+ char *defdata;
int ret = 0, err;
if (nla == NULL)
return -EINVAL;
- err = nla_parse_nested(tb, TCA_DEF_MAX, nla, NULL);
+ err = nla_parse_nested(tb, TCA_DEF_MAX, nla, simple_policy);
if (err < 0)
return err;
if (tb[TCA_DEF_PARMS] == NULL)
return -EINVAL;
- parm = nla_data(tb[TCA_DEF_PARMS]);
- defdata = nla_data(tb[TCA_DEF_DATA]);
- if (defdata == NULL)
+ if (tb[TCA_DEF_DATA] == NULL)
return -EINVAL;
- datalen = nla_len(tb[TCA_DEF_DATA]);
- if (datalen == 0)
- return -EINVAL;
+ parm = nla_data(tb[TCA_DEF_PARMS]);
+ defdata = nla_data(tb[TCA_DEF_DATA]);
pc = tcf_hash_check(parm->index, a, bind, &simp_hash_info);
if (!pc) {
return -ENOMEM;
d = to_defact(pc);
- ret = alloc_defdata(d, datalen, defdata);
+ ret = alloc_defdata(d, defdata);
if (ret < 0) {
kfree(pc);
return ret;
}
+ d->tcf_action = parm->action;
ret = ACT_P_CREATED;
} else {
d = to_defact(pc);
tcf_simp_release(d, bind);
return -EEXIST;
}
- realloc_defdata(d, datalen, defdata);
+ reset_policy(d, defdata, parm);
}
- spin_lock_bh(&d->tcf_lock);
- d->tcf_action = parm->action;
- spin_unlock_bh(&d->tcf_lock);
-
if (ret == ACT_P_CREATED)
tcf_hash_insert(pc, &simp_hash_info);
return ret;
opt.bindcnt = d->tcf_bindcnt - bind;
opt.action = d->tcf_action;
NLA_PUT(skb, TCA_DEF_PARMS, sizeof(opt), &opt);
- NLA_PUT(skb, TCA_DEF_DATA, d->tcfd_datalen, d->tcfd_defdata);
+ NLA_PUT_STRING(skb, TCA_DEF_DATA, d->tcfd_defdata);
t.install = jiffies_to_clock_t(jiffies - d->tcf_tm.install);
t.lastuse = jiffies_to_clock_t(jiffies - d->tcf_tm.lastuse);
t.expires = jiffies_to_clock_t(d->tcf_tm.expires);
return 1;
}
-static void htb_parent_to_leaf(struct htb_class *cl, struct Qdisc *new_q)
+static void htb_parent_to_leaf(struct htb_sched *q, struct htb_class *cl,
+ struct Qdisc *new_q)
{
struct htb_class *parent = cl->parent;
BUG_TRAP(!cl->level && cl->un.leaf.q && !cl->prio_activity);
+ if (parent->cmode != HTB_CAN_SEND)
+ htb_safe_rb_erase(&parent->pq_node, q->wait_pq + parent->level);
+
parent->level = 0;
memset(&parent->un.inner, 0, sizeof(parent->un.inner));
INIT_LIST_HEAD(&parent->un.leaf.drop_list);
htb_deactivate(q, cl);
if (last_child)
- htb_parent_to_leaf(cl, new_q);
+ htb_parent_to_leaf(q, cl, new_q);
if (--cl->refcnt == 0)
htb_destroy_class(sch, cl);