Al Viro [Fri, 1 Aug 2008 08:29:18 +0000 (04:29 -0400)]
[PATCH] fix races and leaks in vfs_quota_on() users
* new helper: vfs_quota_on_path(); equivalent of vfs_quota_on() sans the
pathname resolution.
* callers of vfs_quota_on() that do their own pathname resolution and
checks based on it are switched to vfs_quota_on_path(); that way we
avoid the races.
* reiserfs leaked dentry/vfsmount references on several failure exits.
Stephen Smalley [Mon, 28 Jul 2008 17:32:38 +0000 (13:32 -0400)]
Re: BUG at security/selinux/avc.c:883 (was: Re: linux-next: Tree
for July 17: early crash on x86-64)
SELinux needs MAY_APPEND to be passed down to the security hook.
Otherwise, we get permission denials when only append permission is
granted by policy even if the opening process specified O_APPEND.
Shows up as a regression in the ltp selinux testsuite, fixed by
this patch.
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Heiko Carstens [Fri, 1 Aug 2008 14:39:23 +0000 (16:39 +0200)]
[S390] dont use kthread for smp_rescan_cpus().
Since git commit 3da1c84c00c7e5fa8348336bd8c342f9128b0f14
"workqueues: make get_online_cpus() useable for work->func()"
it is safe to call get_online_cpus() from workqueue context.
So remove the kthread workaround again.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
WARNING: vmlinux.o(.text+0x2e1cc4): Section mismatch in reference from the function s390_virtio_console_init() to the function .init.text:early_put_chars()
The function s390_virtio_console_init() references
the function __init early_put_chars().
This is often because s390_virtio_console_init lacks a __init
annotation or the annotation of early_put_chars is wrong.
WARNING: vmlinux.o(.text+0x2e1cd0): Section mismatch in reference from the function s390_virtio_console_init() to the function .init.text:virtio_cons_early_init()
The function s390_virtio_console_init() references
the function __init virtio_cons_early_init().
This is often because s390_virtio_console_init lacks a __init
annotation or the annotation of virtio_cons_early_init is wrong.
Cc: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Heiko Carstens [Fri, 1 Aug 2008 14:39:20 +0000 (16:39 +0200)]
[S390] qdio: fix section mismatch bug.
Fix the two section mismatch warnings below.
This fixes two real bugs since the code which has __exit annotations
may already be gone when it is called.
WARNING: vmlinux.o(.init.text+0x1cc4a): Section mismatch in reference from the function init_QDIO() to the function .exit.text:qdio_setup_exit()
The function __init init_QDIO() references
a function __exit qdio_setup_exit().
This is often seen when error handling in the init function
uses functionality in the exit path.
The fix is often to remove the __exit annotation of
qdio_setup_exit() so it may be used outside an exit section.
WARNING: vmlinux.o(.init.text+0x1cc7a): Section mismatch in reference from the function init_QDIO() to the function .exit.text:qdio_remove_perf_stats()
The function __init init_QDIO() references
a function __exit qdio_remove_perf_stats().
This is often seen when error handling in the init function
uses functionality in the exit path.
The fix is often to remove the __exit annotation of
qdio_remove_perf_stats() so it may be used outside an exit section.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Heiko Carstens [Fri, 1 Aug 2008 14:39:19 +0000 (16:39 +0200)]
[S390] stp: fix section mismatch warning.
Fix these two (false positive) warnings by adding an __init annoation:
WARNING: vmlinux.o(.text+0x7e6a): Section mismatch in reference from the function stp_reset() to the function .init.text:__alloc_bootmem()
The function stp_reset() references
the function __init __alloc_bootmem().
This is often because stp_reset lacks a __init
annotation or the annotation of __alloc_bootmem is wrong.
WARNING: vmlinux.o(.text+0x7ece): Section mismatch in reference from the function stp_reset() to the function .init.text:free_bootmem()
The function stp_reset() references
the function __init free_bootmem().
This is often because stp_reset lacks a __init
annotation or the annotation of free_bootmem is wrong.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Frank Blaschka [Fri, 1 Aug 2008 14:39:13 +0000 (16:39 +0200)]
[S390] qeth: preallocated qeth header for hiper socket
For hiper socket devices this patch will economize the reallocation
of the tx skb data segment by allocating separate memory for the qdio
transport information (qeth header).
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
[S390] Optimize storage key operations for anon pages
For anonymous pages without a swap cache backing the check in
page_remove_rmap for the physical dirty bit in page_remove_rmap is
unnecessary. The instructions that are used to check and reset the dirty
bit are expensive. Removing the check noticably speeds up process exit.
In addition the clearing of the dirty bit in __SetPageUptodate is
pointless as well. With these two changes there is no storage key
operation for an anonymous page anymore if it does not hit the swap
space.
The micro benchmark which repeatedly executes an empty shell script
gets about 5% faster.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Heiko Carstens [Fri, 1 Aug 2008 14:39:11 +0000 (16:39 +0200)]
[S390] nohz/sclp: disable timer on synchronous waits.
sclp_sync_wait wait synchronously for an sclp interrupt and disables
timer interrupts. However on the irq enter paths there is an extra
check if a timer interrupt would be due and calls the timer callback.
This would schedule softirqs in the wrong context.
So introduce local_tick_enable/disable which prevents this.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Michael Holzheu [Fri, 1 Aug 2008 14:39:10 +0000 (16:39 +0200)]
[S390] ipl: Reboot from alternate device does not work when booting from file
During startup we check if diag308 works using diag 308 subcode 6,
which stores the actual ipl information. This fails with rc = 0x102, if
the system has been ipled from the HMC using load from CD or load from file.
In the case of rc = 0x102 we have to assume that diag 308 is working,
since it still can be used to ipl from an alternative device.
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
When z/VM provides two virtual devices (minidisks) that reside on the
same real device, both will receive the configuration data from the
real device and thus get the same uid. To fix this problem, z/VM
provides an additional configuration data record that allows to
distinguish between minidisks.
z/VM APAR VM64273 needs be installed so this fix has an effect.
Signed-off-by: Stefan Weinhuber <wein@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
drivers/mtd/nand/nandsim.c: In function 'divide':
drivers/mtd/nand/nandsim.c:462: error: implicit declaration of function 'do_div'
Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Jason Wessel [Fri, 1 Aug 2008 13:39:35 +0000 (08:39 -0500)]
kgdb: fix gdb serial thread queries
The command "info threads" did not work correctly with kgdb. It would
result in a silent kernel hang if used.
This patach addresses several problems.
- Fix use of deprecated NR_CPUS
- Fix kgdb to not walk linearly through the pid space
- Correctly implement shadow pids
- Change the threads per query to a #define
- Fix kgdb_hex2long to work with negated values
The threads 0 and -1 are reserved to represent the current task. That
means that CPU 0 will start with a shadow thread id of -2, and CPU 1
will have a shadow thread id of -3, etc...
From the debugger you can switch to a shadow thread to see what one of
the other cpus was doing, however it is not possible to execute run
control operations on any other cpu execept the cpu executing the
kgdb_handle_exception().
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Jason Wessel [Fri, 1 Aug 2008 13:39:34 +0000 (08:39 -0500)]
kgdb: fix kgdb_validate_break_address to perform a mem write
A regression to the kgdb core was found in the case of using the
CONFIG_DEBUG_RODATA kernel option. When this option is on, a breakpoint
cannot be written into any readonly memory page. When an external
debugger requests a breakpoint to get set, the
kgdb_validate_break_address() was only checking to see if the address
to place the breakpoint was readable and lacked a write check.
This patch changes the validate routine to try reading (via the
breakpoint set request) and also to try immediately writing the break
point. If either fails, an error is correctly returned and the
debugger behaves correctly. Then an end user can make the
descision to use hardware breakpoints.
Also update the documentation to reflect that using
CONFIG_DEBUG_RODATA will inhibit the use of software breakpoints.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
David Woodhouse [Fri, 1 Aug 2008 09:07:51 +0000 (10:07 +0100)]
[JFFS2] Fix allocation of summary buffer
We can't use vmalloc for the buffer we use for writing summaries,
because some drivers may want to DMA from it. So limit the size to 64KiB
and use kmalloc for it instead.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Pieter du Preez [Fri, 1 Aug 2008 09:06:40 +0000 (10:06 +0100)]
Fix rename of at91_nand -> atmel_nand
Structs called at91_nand_data where renamed to atmel_nand_data
and configs called *MTD_NAND_AT91* where renamed to
*MTD_NAND_ATMEL*. This was unfortunately not done consistently,
causing NAND chips not being initialised on several ARM boards.
I am aware that the author of the original change did not rename
MTD_NAND_AT91_BUSWIDTH to MTD_NAND_ATMEL_BUSWIDTH, for example.
All *MTD_NAND_AT91* where renamed to *MTD_NAND_ATMEL* in order
to keep naming consistency.
This patch was only tested on a MACH_SAM9_L9260, as this is the
only ARM board I have to my disposal.
Adam Langley [Fri, 1 Aug 2008 04:36:07 +0000 (21:36 -0700)]
tcp: MD5: Fix IPv6 signatures
Reported by Stefanos Harhalakis; although 2.6.27-rc1 talks to itself using IPv6
TCP MD5 packets just fine, Stefanos noted that tcpdump claimed that the
signatures were invalid.
Ingo Molnar [Fri, 1 Aug 2008 03:51:22 +0000 (20:51 -0700)]
net/ipv4/route.c: fix build error
fix:
net/ipv4/route.c: In function 'ip_static_sysctl_init':
net/ipv4/route.c:3225: error: 'ipv4_route_path' undeclared (first use in this function)
net/ipv4/route.c:3225: error: (Each undeclared identifier is reported only once
net/ipv4/route.c:3225: error: for each function it appears in.)
net/ipv4/route.c:3225: error: 'ipv4_route_table' undeclared (first use in this function)
Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: David S. Miller <davem@davemloft.net>
Adam Langley [Fri, 1 Aug 2008 03:49:48 +0000 (20:49 -0700)]
tcp: MD5: Fix MD5 signatures on certain ACK packets
I noticed, looking at tcpdumps, that timewait ACKs were getting sent
with an incorrect MD5 signature when signatures were enabled.
I broke this in 49a72dfb8814c2d65bd9f8c9c6daf6395a1ec58d ("tcp: Fix
MD5 signatures for non-linear skbs"). I didn't take into account that
the skb passed to tcp_*_send_ack was the inbound packet, thus the
source and dest addresses need to be swapped when calculating the MD5
pseudoheader.
Signed-off-by: Adam Langley <agl@imperialviolet.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Yongjun [Fri, 1 Aug 2008 03:46:47 +0000 (20:46 -0700)]
ipv6: Fix ip6_xmit to send fragments if ipfragok is true
SCTP used ip6_xmit() to send fragments after received ICMP packet too
big message. But while send packet used ip6_xmit, the skb->local_df is
not initialized. So when skb if enter ip6_fragment(), the following
code will discard the skb.
SCTP do the following step:
1. send packet ip6_xmit(skb, ipfragok=0)
2. received ICMP packet too big message
3. if PMTUD_ENABLE: ip6_xmit(skb, ipfragok=1)
This patch fixed the problem by set local_df if ipfragok is true.
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Julius Volz [Fri, 1 Aug 2008 03:45:24 +0000 (20:45 -0700)]
ipvs: Move userspace definitions to include/linux/ip_vs.h
Current versions of ipvsadm include "/usr/src/linux/include/net/ip_vs.h"
directly. This file also contains kernel-only definitions. Normally, public
definitions should live in include/linux, so this patch moves the
definitions shared with userspace to a new file, "include/linux/ip_vs.h".
This also removes the unused NFC_IPVS_PROPERTY bitmask, which was once
used to point into skb->nfcache.
To make old ipvsadms still compile with this, the old header file includes
the new one.
Thanks to Dave Miller and Horms for noting/adding the missing Kbuild entry
for the new header file.
Signed-off-by: Julius Volz <juliusv@google.com> Acked-by: Simon Horman <horms@verge.net.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Jean Delvare [Wed, 25 Jun 2008 13:10:30 +0000 (09:10 -0400)]
hwmon: (lm85) Simplify device initialization function
Clean up and simplify the device initialization function:
* Degrade error messages to warnings - what they really are.
* Stop warning about VxI mode, we don't really care.
* Drop comment about lack of limit initialization - that's the standard
way, all hardware monitoring drivers do that.
* Only read the configuration register once.
* Only write back to the configuration register if needed.
* Don't attempt to clear the lock bit, it locks itself to 1.
* Move the function to before it's called, so that we no longer need to
forware declare it.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
Jean Delvare [Wed, 25 Jun 2008 12:47:35 +0000 (08:47 -0400)]
hwmon: (lm85) Misc cleanups
Misc cleanups to the lm85 hardware monitoring driver:
* Mark constant arrays as const.
* Remove useless masks.
* Have lm85_write_value return void - nobody is checking the returned
value anyway and in some cases it was plain wrong.
* Remove useless initializations.
* Rename new_client to client in lm85_detect.
* Replace cascaded if/else with a switch/case in lm85_detect.
* Group similar loops in lm85_update_device.
* Remove legacy comments.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Acked-by: Juerg Haefliger <juergh at gmail.com> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
Jean Delvare [Sat, 12 Apr 2008 17:56:35 +0000 (19:56 +0200)]
hwmon: (lm85) Don't write back cached values
In set_pwm_auto_pwm_minctl, we write cached register bits back to the
chip. This is a bad idea as we have no guarantee that the cache is
up-to-date. Better read a fresh register value from the chip, it's
safer and in fact it is also more simple.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Acked-by: Juerg Haefliger <juergh at gmail.com> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
Jean Delvare [Thu, 1 May 2008 06:47:33 +0000 (08:47 +0200)]
hwmon: (lm85) Drop dead code
Drop a lot of useless register defines, conversion macros, data structure
members and update code. All these register values were read from the
device but nothing is done out of them, so this is all dead code in
practice.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Acked-by: Juerg Haefliger <juergh at gmail.com> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
Jean Delvare [Tue, 29 Apr 2008 12:03:37 +0000 (14:03 +0200)]
hwmon: (lm85) Coding-style cleanups
Fix most style issues reported by checkpatch, including:
* Trailing, missing and extra whitespace
* Extra parentheses, curly braces and semi-colons
* Broken indentation
* Lines too long
I verified that the generated code is the same before and after
these changes.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Acked-by: Juerg Haefliger <juergh at gmail.com> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
David Brownell [Sun, 4 May 2008 02:33:15 +0000 (19:33 -0700)]
hwmon: (lm75) add new-style driver binding
More LM75 updates:
- Teach the LM75 driver to use new-style driver binding:
* Create a second driver struct, using new-style driver binding
methods cribbed from the legacy code.
* Add a MODULE_DEVICE_TABLE (for "newER-style binding")
* The legacy probe logic delegates its work to this new code.
* The legacy driver now uses the name "lm75_legacy".
- More careful initialization. Chips are put into 9-bit mode so
the current interconversion routines will never fail.
- Save the original chip configuration, and restore it on exit.
(Among other things, this normally turns off the mode where
the chip is constantly sampling ... and thus saves power.)
So the new-style code should catch all chips that boards declare,
while the legacy code catches others. This particular coexistence
strategy may need some work yet ... legacy modes might best be set
up explicitly by some tool not unlike "sensors-detect". (Or else
completely eradicated...)
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Acked-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
David Brownell [Mon, 21 Apr 2008 19:10:53 +0000 (12:10 -0700)]
hwmon: (lm75) cleanup/reorg
Minor cleanup and reorg of the lm75 code.
- Kconfig provides a larger list of lm75-compatible chips
- A top comment now says what the driver does (!) ... as in, just
what sort of sensor is this??
- Section comments now delineate the various sections of the driver:
hwmon attributes, driver binding, register access, module glue.
One driver binding function moved out of the attribute section,
as did the driver struct itself.
- Minor tweaks to legacy probe logic: correct a comment, and
remove a pointless variable.
- Whitespace, linelength, and comment fixes.
This patch should include no functional changes. It's preparation
for adding new-style (driver model) I2C driver binding.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Acked-by: Jean Delvare <khali@linux-fr.org> Acked-by: Laurent Pinchart <laurentp@cse-semaphore.com> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
Juerg Haefliger [Wed, 26 Mar 2008 04:49:15 +0000 (21:49 -0700)]
hwmon: (dme1737) fix voltage scaling
This patch fixes a voltage scaling issue for the sch311x device.
Signed-Off-By: Juerg Haefliger <juergh at gmail.com> Acked-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
This patch adds a module load parameter to enable probing of
non-standard LPC addresses 0x162e and 0x164e when scanning for supported
ISA chips.
Signed-Off-By: Juerg Haefliger <juergh at gmail.com> Acked-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
Juerg Haefliger [Wed, 26 Mar 2008 04:49:02 +0000 (21:49 -0700)]
hwmon: (dme1737) demacrofy for readability
This patch gets rid of a couple of macros previously used for sysfs attribute
generation and manipulation. This makes the source a little bigger but a lot
more readable and maintainable. It also fixes an issue with pwm5 & pwm6
attributes not being created read-only initially.
Signed-Off-By: Juerg Haefliger <juergh at gmail.com> Acked-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
David S. Miller [Fri, 1 Aug 2008 03:33:43 +0000 (20:33 -0700)]
sparc64: Kill __show_regs().
The story is that what we used to do when we actually used
smp_report_regs() is that if you specifically only wanted to have the
current cpu's registers dumped you would call "__show_regs()"
otherwise you would call show_regs() which also invoked
smp_report_regs().
Now that we killed off smp_report_regs() there is no longer any
reason to have these two routines, just show_regs() is sufficient.
Also kill off a stray declaration of show_regs() in sparc64_ksym.c
Signed-off-by: David S. Miller <davem@davemloft.net>
Arthur Jones [Fri, 25 Jul 2008 19:03:38 +0000 (12:03 -0700)]
md: raid10: wake up frozen array
When rescheduling a bio in raid10, we wake up
the md thread, but if the array is frozen, this
will have no effect. This causes the array to
remain frozen for eternity. We add a wake_up
to allow the array to de-freeze. This code is
nearly identical to the raid1 code, which has
this fix already.
Signed-off-by: Arthur Jones <ajones@riverbed.com> Signed-off-by: NeilBrown <neilb@suse.de>
David S. Miller [Thu, 31 Jul 2008 23:58:50 +0000 (16:58 -0700)]
netdev: Fix lockdep warnings in multiqueue configurations.
When support for multiple TX queues were added, the
netif_tx_lock() routines we converted to iterate over
all TX queues and grab each queue's spinlock.
This causes heartburn for lockdep and it's not a healthy
thing to do with lots of TX queues anyways.
So modify this to use a top-level lock and a "frozen"
state for the individual TX queues.
Signed-off-by: David S. Miller <davem@davemloft.net>
Julia Lawall [Mon, 21 Jul 2008 07:59:15 +0000 (09:59 +0200)]
[PATCH] ocfs2: Release mutex in error handling code
The mutex is released on a successful return, so it would seem that it
should be released on an error return as well.
The semantic patch finds this problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)
// <smpl>
@@
expression l;
@@
mutex_lock(l);
... when != mutex_unlock(l)
when any
when strict
(
if (...) { ... when != mutex_unlock(l)
+ mutex_unlock(l);
return ...;
}
|
mutex_unlock(l);
)
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
[PATCH 2/2] ocfs2: Fix race between mount and recovery
As the fs recovery is asynchronous, there is a small chance that another
node can mount (and thus recover) the slot before the recovery thread
gets to it.
If this happens, the recovery thread will block indefinitely on the
journal/slot lock as that lock will be held for the duration of the mount
(by design) by the node assigned to that slot.
The solution implemented is to keep track of the journal replays using
a recovery generation in the journal inode, which will be incremented by the
thread replaying that journal. The recovery thread, before attempting the
blocking lock on the journal/slot lock, will compare the generation on disk
with what it has cached and skip recovery if it does not match.
This bug appears to have been inadvertently introduced during the mount/umount
vote removal by mainline commit 34d024f84345807bf44163fac84e921513dde323. In the
mount voting scheme, the messaging would indirectly indicate that the slot
was being recovered.
Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
[PATCH 1/2] ocfs2: Add counter in struct ocfs2_dinode to track journal replays
This patch renames the ij_pad to ij_recovery_generation in struct ocfs2_dinode.
This will be used to keep count of journal replays after an unclean shutdown.
Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Joel Becker [Thu, 19 Jun 2008 02:29:05 +0000 (19:29 -0700)]
[PATCH] configfs: Convenience macros for attribute definition.
Sysfs has the _ATTR() and _ATTR_RO() macros to make defining extended
form attributes easier. configfs should have something similiar.
- _CONFIGFS_ATTR() and _CONFIGFS_ATTR_RO() are the counterparts to the
sysfs macros.
- CONFIGFS_ATTR_STRUCT() creates the extended form attribute structure.
- CONFIGFS_ATTR_OPS() defines the show_attribute()/store_attribute()
operations that call the show()/store() operations of the extended
form configfs_attributes.
Signed-off-by: Joel Becker <joel.becker@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Joel Becker [Tue, 17 Jun 2008 22:34:32 +0000 (15:34 -0700)]
[PATCH] configfs: Pin configfs subsystems separately from new config_items.
configfs_mkdir() creates a new item by calling its parent's
->make_item/group() functions. Once that object is created,
configfs_mkdir() calls try_module_get() on the new item's module. If it
succeeds, the module owning the new item cannot be unloaded, and
configfs is safe to reference the item.
If the item and the subsystem it belongs to are part of the same module,
the subsystem is also pinned. This is the common case.
However, if the subsystem is made up of multiple modules, this may not
pin the subsystem. Thus, it would be possible to unload the toplevel
subsystem module while there is still a child item. Thus, we now
try_module_get() the subsystem's module. This only really affects
children of the toplevel subsystem group. Deeper children already have
their parents pinned.
Signed-off-by: Joel Becker <joel.becker@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Louis Rilling [Fri, 27 Jun 2008 11:10:25 +0000 (13:10 +0200)]
[PATCH] configfs: Fix open directory making rmdir() fail
When checking for user-created elements under an item to be removed by rmdir(),
configfs_detach_prep() counts fake configfs_dirents created by dir_open() as
user-created and fails when finding one. It is however perfectly valid to remove
a directory that is open.
Simply make configfs_detach_prep() skip fake configfs_dirent, like it already
does for attributes, and like detach_groups() does.
Signed-off-by: Louis Rilling <louis.rilling@kerlabs.com> Signed-off-by: Joel Becker <joel.becker@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Louis Rilling [Fri, 4 Jul 2008 14:56:06 +0000 (16:56 +0200)]
[PATCH] configfs: Lock new directory inodes before removing on cleanup after failure
Once a new configfs directory is created by configfs_attach_item() or
configfs_attach_group(), a failure in the remaining initialization steps leads
to removing a directory which inode the VFS may have already accessed.
This commit adds the necessary inode locking to safely remove configfs
directories while cleaning up after a failure. As an advantage, the locking
rules of populate_groups() and detach_groups() become the same: the caller must
have the group's inode mutex locked.
Signed-off-by: Louis Rilling <louis.rilling@kerlabs.com> Signed-off-by: Joel Becker <joel.becker@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Louis Rilling [Fri, 4 Jul 2008 14:56:05 +0000 (16:56 +0200)]
[PATCH] configfs: Prevent userspace from creating new entries under attaching directories
process 1: process 2:
configfs_mkdir("A")
attach_group("A")
attach_item("A")
d_instantiate("A")
populate_groups("A")
mutex_lock("A")
attach_group("A/B")
attach_item("A")
d_instantiate("A/B")
mkdir("A/B/C")
do_path_lookup("A/B/C", LOOKUP_PARENT)
ok
lookup_create("A/B/C")
mutex_lock("A/B")
ok
configfs_mkdir("A/B/C")
ok
attach_group("A/C")
attach_item("A/C")
d_instantiate("A/C")
populate_groups("A/C")
mutex_lock("A/C")
attach_group("A/C/D")
attach_item("A/C/D")
failure
mutex_unlock("A/C")
detach_groups("A/C")
nothing to do
mkdir("A/C/E")
do_path_lookup("A/C/E", LOOKUP_PARENT)
ok
lookup_create("A/C/E")
mutex_lock("A/C")
ok
configfs_mkdir("A/C/E")
ok
detach_item("A/C")
d_delete("A/C")
mutex_unlock("A")
detach_groups("A")
mutex_lock("A/B")
detach_group("A/B")
detach_groups("A/B")
nothing since no _default_ group
detach_item("A/B")
mutex_unlock("A/B")
d_delete("A/B")
detach_item("A")
d_delete("A")
Two bugs:
1/ "A/B/C" and "A/C/E" are created, but never removed while their parent are
removed in the end. The same could happen with symlink() instead of mkdir().
2/ "A" and "A/C" inodes are not locked while detach_item() is called on them,
which may probably confuse VFS.
This commit fixes 1/, tagging new directories with CONFIGFS_USET_CREATING before
building the inode and instantiating the dentry, and validating the whole
group+default groups hierarchy in a second pass by clearing
CONFIGFS_USET_CREATING.
mkdir(), symlink(), lookup(), and dir_open() simply return -ENOENT if
called in (or linking to) a directory tagged with CONFIGFS_USET_CREATING. This
does not prevent userspace from calling stat() successfuly on such directories,
but this prevents userspace from adding (children to | symlinking from/to |
read/write attributes of | listing the contents of) not validated items. In
other words, userspace will not interact with the subsystem on a new item until
the new item creation completes correctly.
It was first proposed to re-use CONFIGFS_USET_IN_MKDIR instead of a new
flag CONFIGFS_USET_CREATING, but this generated conflicts when checking the
target of a new symlink: a valid target directory in the middle of attaching
a new user-created child item could be wrongly detected as being attached.
2/ is fixed by next commit.
Signed-off-by: Louis Rilling <louis.rilling@kerlabs.com> Signed-off-by: Joel Becker <joel.becker@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Louis Rilling [Fri, 20 Jun 2008 12:09:22 +0000 (14:09 +0200)]
[PATCH] configfs: Fix failing symlink() making rmdir() fail
On a similar pattern as mkdir() vs rmdir(), a failing symlink() may make rmdir()
fail for the symlink's parent and the symlink's target as well.
failing symlink() making target's rmdir() fail:
process 1: process 2:
symlink("A/S" -> "B")
allow_link()
create_link()
attach to "B" links list
rmdir("B")
detach_prep("B")
error because of new link
configfs_create_link("A", "S")
error (eg -ENOMEM)
failing symlink() making parent's rmdir() fail:
process 1: process 2:
symlink("A/D/S" -> "B")
allow_link()
create_link()
attach to "B" links list
configfs_create_link("A/D", "S")
make_dirent("A/D", "S")
rmdir("A")
detach_prep("A")
detach_prep("A/D")
error because of "S"
create("S")
error (eg -ENOMEM)
We cannot use the same solution as for mkdir() vs rmdir(), since rmdir() on the
target cannot wait on the i_mutex of the new symlink's parent without risking a
deadlock (with other symlink() or sys_rename()). Instead we define a global
mutex protecting all configfs symlinks attachment, so that rmdir() can avoid the
races above.
Signed-off-by: Louis Rilling <louis.rilling@kerlabs.com> Signed-off-by: Joel Becker <joel.becker@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Louis Rilling [Mon, 23 Jun 2008 12:16:17 +0000 (14:16 +0200)]
[PATCH] configfs: Fix symlink() to a removing item
The rule for configfs symlinks is that symlinks always point to valid
config_items, and prevent the target from being removed. However,
configfs_symlink() only checks that it can grab a reference on the target item,
without ensuring that it remains alive until the symlink is correctly attached.
This patch makes configfs_symlink() fail whenever the target is being removed,
using the CONFIGFS_USET_DROPPING flag set by configfs_detach_prep() and
protected by configfs_dirent_lock.
This patch introduces a similar (weird?) behavior as with mkdir failures making
rmdir fail: if symlink() races with rmdir() of the parent directory (or its
youngest user-created ancestor if parent is a default group) or rmdir() of the
target directory, and then fails in configfs_create(), this can make the racing
rmdir() fail despite the concerned directory having no user-created entry (resp.
no symlink pointing to it or one of its default groups) in the end.
This behavior is fixed in later patches.
Signed-off-by: Louis Rilling <louis.rilling@kerlabs.com> Signed-off-by: Joel Becker <joel.becker@oracle.com> Signed-off-by: Mark Fasheh <mfasheh@suse.com>
kbuild: genksyms parser: fix the __attribute__ rule
We are having two kinds of problems with genksyms today: fake checksum
changes without actual ABI changes, and changes which we would rather like
to ignore (such as an additional field at the end of a structure that
modules are not supposed to touch, for example).
I have thought about ways to improve genksyms and compute checksums
differently to avoid those problems, but in the end I don't see a
fundamentally better way. So here are some genksyms patches for at least
making the checksums more easily manageable, if we cannot fully fix them.
In addition to the bugfixes (the first two patches), this allows genksyms
to track checksum changes and report why a checksum changed (third patch),
and to selectively ignore changes (fourth patch).
This patch:
Gcc __attribute__ definitions may occur repeatedly, e.g.,
static int foo __attribute__((__used__))
__attribute__((aligned (16)));
The genksyms parser does not understand this, and generates a syntax error.
Fix this case.
Signed-off-by: Andreas Gruenbacher <agruen@suse.de> Cc: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
The z10 system supports large pages, kvm-s390 doesnt.
Make sure that we dont advertise large pages to avoid the guest crashing as
soon as the guest kernel activates DAT.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Avi Kivity <avi@qumranet.com>
David S. Miller [Thu, 31 Jul 2008 08:06:02 +0000 (01:06 -0700)]
sparc64: Kill smp_report_regs().
All the call sites are #if 0'd out and we have a much more
useful global cpu dumping facility these days. smp_report_regs()
is way too verbose to be usable.
Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Thu, 31 Jul 2008 07:38:52 +0000 (00:38 -0700)]
netfilter: xt_hashlimit: fix race between htable_destroy and htable_gc
Deleting a timer with del_timer doesn't guarantee, that the
timer function is not running at the moment of deletion. Thus
in the xt_hashlimit case we can get into a ticklish situation
when the htable_gc rearms the timer back and we'll actually
delete an entry with a pending timer.
Fix it with using del_timer_sync().
AFAIK del_timer_sync checks for the timer to be pending by
itself, so I remove the check.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Thu, 31 Jul 2008 07:38:31 +0000 (00:38 -0700)]
netfilter: ipt_recent: fix race between recent_mt_destroy and proc manipulations
The thing is that recent_mt_destroy first flushes the entries
from table with the recent_table_flush and only *after* this
removes the proc file, corresponding to that table.
Thus, if we manage to write to this file the '+XXX' command we
will leak some entries. If we manage to write there a 'clean'
command we'll race in two recent_table_flush flows, since the
recent_mt_destroy calls this outside the recent_lock.
The proper solution as I see it is to remove the proc file first
and then go on with flushing the table. This flushing becomes
safe w/o the lock, since the table is already inaccessible from
the outside.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Alan Cox [Thu, 24 Jul 2008 16:16:06 +0000 (17:16 +0100)]
pata_it821x: Driver updates and reworking
- Add support for the RDC 1010 variant
- Rework the core library to have a read_id method. This allows the hacky
bits of it821x to go and prepares us for pata_hd
- Switch from WARN to BUG in ata_id_string as it will reboot if you get
it wrong so WARN won't be seen
- Allow the issue of command 0xFC on the 821x. This is needed to query
rebuild status.
- Tidy up printk formatting
- Do more ident rewriting on RAID volumes to handle firmware provided
ident data which is rather wonky
- Report the firmware revision and device layout in RAID mode
- Don't try and disable raid on the 8211 or RDC - they don't have the
relevant bits
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Elias Oltmanns [Wed, 30 Jul 2008 19:32:39 +0000 (12:32 -0700)]
libata-core: make sure that ata_force_tbl is freed in case of an error
Fix a potential memory leak when ata_init() encounters an error.
Signed-off-by: Elias Oltmanns <eo@nebensachen.de> Cc: Tejun Heo <tj@kernel.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Global and per-LLD ATAPI disable checks were done in the command issue
path probably because it was left out during EH conversion. On
affected machines, this can cause lots of warning messages. Move them
to where they belong - the probing path.
Reported by Chunbo Luo.
Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Chunbo Luo <chunbo.luo@windriver.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
pata_via: add VX800 flag; add function for fixing h/w bugs
Add flag VIA_SATA_PATA for vx800, VX800 uses the same
chipset(0x0581/0x5324) as CX700, which has 1 PATA channel(Master/Slave)
and 1 SATA channel(Master/Slave) Add function <via_ata_tf_load>. This is
to fix the internal bug of VIA chipsets, which will reset the device
register after changing the IEN bit in CTL register
Signed-off-by: Joseph Chan <josephchan@via.com.tw> Cc: Tejun Heo <tj@kernel.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
David S. Miller [Thu, 31 Jul 2008 04:57:59 +0000 (21:57 -0700)]
sparc64: Make global reg dumping even more useful.
Record one more level of stack frame program counter.
Particularly when lockdep and all sorts of spinlock debugging is
enabled, figuring out the caller of spin_lock() is difficult when the
cpu is stuck on the lock.
Signed-off-by: David S. Miller <davem@davemloft.net>
Harvey Harrison [Thu, 31 Jul 2008 00:20:18 +0000 (17:20 -0700)]
irda: replace __FUNCTION__ with __func__
__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Matthew Garrett [Thu, 31 Jul 2008 00:00:38 +0000 (17:00 -0700)]
nsc-ircc: default to dongle type 9 on IBM hardware
This is necessary to set the dongle type on the nsc driver in order to get
it to work correctly. Thinkpads all appear to use dongle type 9. This
patch defaults nsc devices with an IBM PnP descriptor to use type 9.
Signed-off-by: Matthew Garrett <mjg59@srcf.ucam.org> Signed-off-by: Ben Collins <ben.collins@canonical.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Michael Frey [Wed, 30 Jul 2008 23:59:15 +0000 (16:59 -0700)]
bluetooth: add quirks for a few hci_usb devices
Preface: The "Broadcom" device is on unreleased hardware, so I can't
disclose the actual model.
When the Dell 370 and 410 BT adapters are put into BT radio mode, they
need to be prepared like many other Broadcom adapters.
Also, add quirk Broadcom 2046 devices with HCI_RESET. Reference for this
bug: https://launchpad.net/bugs/249448
Signed-off-by: Michael Frey <michael.frey@canonical.com> Signed-off-by: Mario Limonciello <Mario_Limonciello@Dell.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Ben Collins <ben.collins@canonical.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
David Howells [Wed, 30 Jul 2008 23:48:05 +0000 (16:48 -0700)]
hysdn: remove the packed attribute from PofTimStamp_tag
Remove the packed attribute from PofTimStamp_tag in the hysdn driver as the
thing being packed is just an array of chars and so is unpackable.
This deals with a compiler warning:
In file included from drivers/isdn/hysdn/hysdn_boot.c:19:
drivers/isdn/hysdn/hysdn_pof.h:63: warning: 'packed' attribute ignored for field of type 'unsigned char[40]'
Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Harvey Harrison [Wed, 30 Jul 2008 23:40:22 +0000 (16:40 -0700)]
isdn: use the common ascii hex helpers
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Acked-by: Karsten Keil <kkeil@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-2.6-omap-h63xx.git / log