Bernd Schmidt [Wed, 7 Jan 2009 15:14:39 +0000 (23:14 +0800)]
Blackfin arch: fix bug - crashes in tcp_v4_send_reset
Michael says that some bugs are crashes in tcp_v4_send_reset.
There's a missing clobber of "CC" in our checksum assembly
statement; fixing this makes the generated code look much saner.
Mike Frysinger [Wed, 7 Jan 2009 15:14:39 +0000 (23:14 +0800)]
Blackfin arch: rewrite dma_memcpy() and dma in/out functions
- unify all dma in/out functions (takes ~35 lines of code now)
- unify dma_memcpy with dma in/out functions (1 place that touches MDMA0
registers)
- add support for 32bit transfers
- cleanup dma_memcpy code to be much more readable
- irqs are disabled only while programming MDMA registers rather than
the entire transaction
Signed-off-by: Mike Frysinger <vapier.adi@gmail.com> Signed-off-by: Bryan Wu <cooloney@kernel.org>
Blackfin arch: Add BF537-STAMP platform support for ENC28J60 SPI Ethernet MAC
This device shouldn't be considered as an alternative to a Memory Mapped
or built-in Ethernet MAC. Throughput is slow (~460kByte/s) while generating
a very high system load (~60%).
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com> Signed-off-by: Bryan Wu <cooloney@kernel.org>
Bernd Schmidt [Wed, 7 Jan 2009 15:14:39 +0000 (23:14 +0800)]
Blackfin arch: fix bugs in linker script when using upstream binutils
Fix a few problems I discovered when building a kernel with upstream CVS
binutils.
We have to add the NOTES macro to our linker script, since a kernel
built with --build-id is otherwise unable to boot. Last time NOTES was
added, it broke things, but the definition of the macro has changed not
to rely on parts of the linker script that aren't present on Blackfin.
I also noticed that _l2_lma_start does not point into the kernel image,
but rather somewhere in L1/L2 space, which seems unintended. Also, when
the L2 section was added to the linker script, the part following it which
computes then length of the init section was not updated.
Blackfin arch: fix bugs and unify BFIN_KERNEL_CLOCK option
- remove duplicated code and headers
- add option allowing arbitrary SDRAM/DDR Timing parameters.
- mark automatically calculated timings as EXPERIMENTAL
- fix comment header block
Related to BUGs:
- kernel boot up fails with CONFIG_BFIN_KERNEL_CLOCK item on.
- kernel does not boot if re-program clocks
[ Mike Frysinger <vapier.adi@gmail.com>
- fix comment header
- mark do_sync static
- document the DMA shutdown
- simplify SIC_IWR handling
- fix ANOMALY_05000265 handling to work as intended ]
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com> Signed-off-by: Mike Frysinger <vapier.adi@gmail.com> Signed-off-by: Bryan Wu <cooloney@kernel.org>
Graf Yang [Wed, 7 Jan 2009 15:14:38 +0000 (23:14 +0800)]
Blackfin arch: Allow a gpio pin be requested both as gpio and irq.
[Mike Frysinger <vapier.adi@gmail.com>:
- use KERN_NOTICE when using gpios as both irq and non
rather than KERN_ERR
- embedded newlines in printk() does not fly]
Signed-off-by: Graf Yang <graf.yang@analog.com> Signed-off-by: Mike Frysinger <vapier.adi@gmail.com> Signed-off-by: Bryan Wu <cooloney@kernel.org>
Peter Zijlstra [Wed, 7 Jan 2009 14:28:57 +0000 (15:28 +0100)]
sched: fix possible recursive rq->lock
Vaidyanathan Srinivasan reported:
> =============================================
> [ INFO: possible recursive locking detected ]
> 2.6.28-autotest-tip-sv #1
> ---------------------------------------------
> klogd/5062 is trying to acquire lock:
> (&rq->lock){++..}, at: [<ffffffff8022aca2>] task_rq_lock+0x45/0x7e
>
> but task is already holding lock:
> (&rq->lock){++..}, at: [<ffffffff805f7354>] schedule+0x158/0xa31
With sched_mc at 2. (it is default-off)
Strictly speaking we'll not deadlock, because ttwu will not be able to
place the migration task on our rq, but since the code can deal with
both rqs getting unlocked, this seems the easiest way out.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Jean Delvare [Wed, 7 Jan 2009 13:29:17 +0000 (14:29 +0100)]
Input: apanel - convert to new i2c binding
Convert the apanel driver to the new i2c device driver binding model,
as the legacy model is going away soon. In the new model, the apanel
driver is no longer scanning all the i2c adapters, instead the
relevant bus driver (i2c-i801) is instantiating the device as needed.
One side benefit is that the apanel driver will now load automatically
on all systems where it is needed.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Stephen Hemminger <shemminger@linux-foundation.org>
Jean Delvare [Wed, 7 Jan 2009 13:29:17 +0000 (14:29 +0100)]
i2c: Drop I2C_CLASS_CAM_DIGITAL
There are a number of drivers which set their i2c bus class to
I2C_CLASS_CAM_DIGITAL, however no chip driver actually checks for this
flag, so we might as well drop it now.
Jean Delvare [Wed, 7 Jan 2009 13:29:16 +0000 (14:29 +0100)]
i2c: Drop I2C_CLASS_ALL
I2C_CLASS_ALL is almost never what bus driver authors really want.
These i2c classes are really only about which devices must be probed,
not what devices can be present. As device drivers get converted to the
new i2c device driver model, only a few device types will keep relying
on probing.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Acked-by: Sonic Zhang <sonic.zhang@analog.com>
Kay Sievers [Wed, 7 Jan 2009 13:29:16 +0000 (14:29 +0100)]
i2c: Replace bus_id with dev_name(), dev_set_name()
This patch is part of a larger patch series which will remove
the "char bus_id[20]" name string from struct device. The device
name is managed in the kobject anyway, and without any size
limitation, and just needlessly copied into "struct device".
To set and read the device name dev_name(dev) and dev_set_name(dev)
must be used. If your code uses static kobjects, which it shouldn't
do, "const char *init_name" can be used to statically provide the
name the registered device should have. At registration time, the
init_name field is cleared, to enforce the use of dev_name(dev) to
access the device name at a later time.
We need to get rid of all occurrences of bus_id in the entire tree
to be able to enable the new interface. Please apply this patch,
and possibly convert any remaining remaining occurrences of bus_id.
We want to submit a patch to -next, which will remove bus_id from
"struct device", to find the remaining pieces to convert, and finally
switch over to the new api, which will remove the 20 bytes array
and does no longer have a size limitation.
Acked-by: Greg Kroah-Hartman <gregkh@suse.de> Signed-off-by: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by: Jean Delvare <khali@linux-fr.org>
FUJITA Tomonori [Tue, 6 Jan 2009 17:13:42 +0000 (02:13 +0900)]
convert to use generic dma_map_ops struct, cleanup
Ingo Molnar <mingo@elte.hu> wrote:
> looks good on x86 but on ia64 there's a problem with one of the
> prototypes:
>
> In file included from tip/arch/ia64/include/asm/io.h:72,
> from tip/arch/ia64/include/asm/smp.h:20,
> from tip/include/linux/smp.h:33,
> from tip/include/linux/sched.h:68,
> from tip/arch/ia64/kernel/asm-offsets.c:9:
> tip/arch/ia64/include/asm/machvec.h:101: warning: parameter has incomplete type
> tip/arch/ia64/include/asm/machvec.h:103: warning: parameter has incomplete type
>
> that's about "enum dma_data_direction".
>
> I dont think enums can be forward declared like that.
>
> machvec.h is a fairly lowlevel include file - so including
> linux/dma-mapping.h probably wont work. We could do a
> linux/dma-mapping-types.h file that is more lowlevel, or we could move the
> machvec_dma_sync_single() and machvec_dma_sync_sg() declarations to a more
> highlevel file - like arch/ia64/include/asm/dma-mapping.h.
>
> To me the latter looks cleaner but no strong feelings.
Yeah, agreed.
They are generic IA64 DMA operations so I think that it makes sense to
move them to dma-mapping.h.
x86: fix section mismatch warnings in mcheck/mce_amd_64.c
Mark the function local_allocate_threshold_blocks() with __cpuinit,
in order to remove the following section mismatch messages:
WARNING: arch/x86/kernel/cpu/mcheck/built-in.o(.text+0x1363): Section mismatch in reference from the function local_allocate_threshold_blocks() to the function .cpuinit.text:allocate_threshold_blocks()
The function local_allocate_threshold_blocks() references
the function __cpuinit allocate_threshold_blocks().
This is often because local_allocate_threshold_blocks lacks a __cpuinit
annotation or the annotation of allocate_threshold_blocks is wrong.
WARNING: arch/x86/kernel/cpu/built-in.o(.text+0x1def): Section mismatch in reference from the function local_allocate_threshold_blocks() to the function .cpuinit.text:allocate_threshold_blocks()
The function local_allocate_threshold_blocks() references
the function __cpuinit allocate_threshold_blocks().
This is often because local_allocate_threshold_blocks lacks a __cpuinit
annotation or the annotation of allocate_threshold_blocks is wrong.
WARNING: arch/x86/kernel/built-in.o(.text+0xef2b): Section mismatch in reference from the function local_allocate_threshold_blocks() to the function .cpuinit.text:allocate_threshold_blocks()
The function local_allocate_threshold_blocks() references
the function __cpuinit allocate_threshold_blocks().
This is often because local_allocate_threshold_blocks lacks a __cpuinit
annotation or the annotation of allocate_threshold_blocks is wrong.
All the callsites of this function are __cpuinit already, and all the
functions it calls are __cpuinit as well.
Ingo Molnar [Wed, 7 Jan 2009 10:05:10 +0000 (11:05 +0100)]
x86: offer frame pointers in all build modes
CONFIG_FRAME_POINTERS=y results in much better debug info for the
kernel (clear and precise backtraces), with the only drawback being
a ~1% increase in kernel size.
So offer it unconditionally and enable it by default.
Steven Rostedt [Tue, 6 Jan 2009 17:43:01 +0000 (12:43 -0500)]
ftrace: convert unsigned index to signed
Impact: fix to unsigned compared to less than zero
Roel Kluin pointed out that there is a compare of an unsigned number
to less than zero. A previous clean up had the unsigned index set
to -1 for certain cases, but never converted it to signed.
Frederic Weisbecker noticed that another index is used to compare
the above index to and it also needs to be converted to signed.
[
Converted ftrace_page->index to int from unsigned long as
Andrew Morton pointed out that there's no need for it to
stay a long.
]
gfs2_write_alloc_required() expects pos to be the offset in bytes. However,
instead of the page index being shifted by by PAGE_CACHE_SHIFT, it was shifted
by (PAGE_CACHE_SIZE - inode->i_blkbits). This patch simply shifts the page
index by the proper amount.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com> Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
As a result of this recent patch:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=b3a6ffe16b5cc48abe7db8d04882dc45280eb693
We only need to depend on LBD.
Reported-by: Fabio M. Di Nitto <fdinitto@redhat.com> Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Alain Knaff [Wed, 7 Jan 2009 08:10:27 +0000 (00:10 -0800)]
bzip2/lzma: fix built-in initramfs vs CONFIG_RD_GZIP
Impact: Resolves build failures in some configurations
Makes it possible to disable CONFIG_RD_GZIP . In that case, the
built-in initramfs will be compressed by whatever compressor is
available (bzip2 or lzma) or left uncompressed if none is available.
It also removes a couple of warnings which occur when no ramdisk
compression at all is chosen.
It also restores the select ZLIB_INFLATE in drivers/block/Kconfig
which somehow came missing. This is needed to activate compilation of
the stuff in zlib_deflate.
Signed-off-by: Alain Knaff <alain@knaff.lu> Signed-off-by: H. Peter Anvin <hpa@zytor.com>
H. Peter Anvin [Wed, 7 Jan 2009 08:03:49 +0000 (00:03 -0800)]
bzip2/lzma: move initrd/ramfs options out of BLK_DEV
Impact: Partial resolution of build failure
Move the initrd/initramfs configuration options from
drivers/block/Kconfig to usr/Kconfig, since they do not and should not
depend on CONFIG_BLK_DEV. This fixes builds when CONFIG_BLK_DEV=n.
Linus Torvalds [Wed, 7 Jan 2009 05:17:57 +0000 (21:17 -0800)]
Fix up 64-bit byte swaps for most 32-bit architectures
The __SWAB_64_THRU_32__ case of a 64-bit byte swap was depending on the
no-longer-existant ___swab32() method (three underscores). We got rid
of some of the worst indirection and complexity, and now it should just
use the 32-bit swab function that was defined right above it.
Reported-and-tested-by: Nicolas Pitre <nico@cam.org> Reported-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
projects / linux-2.6-omap-h63xx.git / log