Dan Williams [Tue, 2 Jan 2007 20:52:30 +0000 (13:52 -0700)]
md: raid5_run_ops - run stripe operations outside sh->lock
When the raid acceleration work was proposed, Neil laid out the following
attack plan:
1/ move the xor and copy operations outside spin_lock(&sh->lock)
2/ find/implement an asynchronous offload api
The raid5_run_ops routine uses the asynchronous offload api (async_tx) and
the stripe_operations member of a stripe_head to carry out xor+copy
operations asynchronously, outside the lock.
To perform operations outside the lock a new set of state flags is needed
to track new requests, in-flight requests, and completed requests. In this
new model handle_stripe is tasked with scanning the stripe_head for work,
updating the stripe_operations structure, and finally dropping the lock and
calling raid5_run_ops for processing. The following flags outline the
requests that handle_stripe can make of raid5_run_ops:
STRIPE_OP_BIOFILL
- copy data into request buffers to satisfy a read request
STRIPE_OP_COMPUTE_BLK
- generate a missing block in the cache from the other blocks
STRIPE_OP_PREXOR
- subtract existing data as part of the read-modify-write process
STRIPE_OP_BIODRAIN
- copy data out of request buffers to satisfy a write request
STRIPE_OP_POSTXOR
- recalculate parity for new data that has entered the cache
STRIPE_OP_CHECK
- verify that the parity is correct
STRIPE_OP_IO
- submit i/o to the member disks (note this was already performed outside
the stripe lock, but it made sense to add it as an operation type
The flow is:
1/ handle_stripe sets STRIPE_OP_* in sh->ops.pending
2/ raid5_run_ops reads sh->ops.pending, sets sh->ops.ack, and submits the
operation to the async_tx api
3/ async_tx triggers the completion callback routine to set
sh->ops.complete and release the stripe
4/ handle_stripe runs again to finish the operation and optionally submit
new operations that were previously blocked
Note this patch just defines raid5_run_ops, subsequent commits (one per
major operation type) modify handle_stripe to take advantage of this
routine.
Changelog:
* removed ops_complete_biodrain in favor of ops_complete_postxor and
ops_complete_write.
* removed the raid5_run_ops workqueue
* call bi_end_io for reads in ops_complete_biofill, saves a call to
handle_stripe
* explicitly handle the 2-disk raid5 case (xor becomes memcpy), Neil Brown
* fix race between async engines and bi_end_io call for reads, Neil Brown
* remove unnecessary spin_lock from ops_complete_biofill
* remove test_and_set/test_and_clear BUG_ONs, Neil Brown
* remove explicit interrupt handling for channel switching, this feature
was absorbed (i.e. it is now implicit) by the async_tx api
* use return_io in ops_complete_biofill
Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-By: NeilBrown <neilb@suse.de>
Dan Williams [Mon, 9 Jul 2007 18:56:43 +0000 (11:56 -0700)]
raid5: replace custom debug PRINTKs with standard pr_debug
Replaces PRINTK with pr_debug, and kills the RAID5_DEBUG definition in
favor of the global DEBUG definition. To get local debug messages just add
'#define DEBUG' to the top of the file.
Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-By: NeilBrown <neilb@suse.de>
Dan Williams [Mon, 9 Jul 2007 18:56:43 +0000 (11:56 -0700)]
raid5: refactor handle_stripe5 and handle_stripe6 (v3)
handle_stripe5 and handle_stripe6 have very deep logic paths handling the
various states of a stripe_head. By introducing the 'stripe_head_state'
and 'r6_state' objects, large portions of the logic can be moved to
sub-routines.
'struct stripe_head_state' consumes all of the automatic variables that previously
stood alone in handle_stripe5,6. 'struct r6_state' contains the handle_stripe6
specific variables like p_failed and q_failed.
One of the nice side effects of the 'stripe_head_state' change is that it
allows for further reductions in code duplication between raid5 and raid6.
The following new routines are shared between raid5 and raid6:
Changes:
* v2: fixed 'conf->raid_disk-1' for the raid6 'handle_stripe_expansion' path
* v3: removed the unused 'dirty' field from struct stripe_head_state
* v3: coalesced open coded bi_end_io routines into return_io()
Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-By: NeilBrown <neilb@suse.de>
Dan Williams [Tue, 2 Jan 2007 18:10:44 +0000 (11:10 -0700)]
async_tx: add the async_tx api
The async_tx api provides methods for describing a chain of asynchronous
bulk memory transfers/transforms with support for inter-transactional
dependencies. It is implemented as a dmaengine client that smooths over
the details of different hardware offload engine implementations. Code
that is written to the api can optimize for asynchronous operation and the
api will fit the chain of operations to the available offload resources.
I imagine that any piece of ADMA hardware would register with the
'async_*' subsystem, and a call to async_X would be routed as
appropriate, or be run in-line. - Neil Brown
async_tx exploits the capabilities of struct dma_async_tx_descriptor to
provide an api of the following general format:
async_tx_find_channel() returns a capable channel from its pool. The
channel pool is organized as a per-cpu array of channel pointers. The
async_tx_rebalance() routine is tasked with managing these arrays. In the
uniprocessor case async_tx_rebalance() tries to spread responsibility
evenly over channels of similar capabilities. For example if there are two
copy+xor channels, one will handle copy operations and the other will
handle xor. In the SMP case async_tx_rebalance() attempts to spread the
operations evenly over the cpus, e.g. cpu0 gets copy channel0 and xor
channel0 while cpu1 gets copy channel 1 and xor channel 1. When a
dependency is specified async_tx_find_channel defaults to keeping the
operation on the same channel. A xor->copy->xor chain will stay on one
channel if it supports both operation types, otherwise the transaction will
transition between a copy and a xor resource.
Currently the raid5 implementation in the MD raid456 driver has been
converted to the async_tx api. A driver for the offload engines on the
Intel Xscale series of I/O processors, iop-adma, is provided in a later
commit. With the iop-adma driver and async_tx, raid456 is able to offload
copy, xor, and xor-zero-sum operations to hardware engines.
On iop342 tiobench showed higher throughput for sequential writes (20 - 30%
improvement) and sequential reads to a degraded array (40 - 55%
improvement). For the other cases performance was roughly equal, +/- a few
percentage points. On a x86-smp platform the performance of the async_tx
implementation (in synchronous mode) was also +/- a few percentage points
of the original implementation. According to 'top' on iop342 CPU
utilization drops from ~50% to ~15% during a 'resync' while the speed
according to /proc/mdstat doubles from ~25 MB/s to ~50 MB/s.
The tiobench command line used for testing was: tiobench --size 2048
--block 4096 --block 131072 --dir /mnt/raid --numruns 5
* iop342 had 1GB of memory available
Details:
* if CONFIG_DMA_ENGINE=n the asynchronous path is compiled away by making
async_tx_find_channel a static inline routine that always returns NULL
* when a callback is specified for a given transaction an interrupt will
fire at operation completion time and the callback will occur in a
tasklet. if the the channel does not support interrupts then a live
polling wait will be performed
* the api is written as a dmaengine client that requests all available
channels
* In support of dependencies the api implicitly schedules channel-switch
interrupts. The interrupt triggers the cleanup tasklet which causes
pending operations to be scheduled on the next channel
* Xor engines treat an xor destination address differently than a software
xor routine. To the software routine the destination address is an implied
source, whereas engines treat it as a write-only destination. This patch
modifies the xor_blocks routine to take a an explicit destination address
to mirror the hardware.
Changelog:
* fixed a leftover debug print
* don't allow callbacks in async_interrupt_cond
* fixed xor_block changes
* fixed usage of ASYNC_TX_XOR_DROP_DEST
* drop dma mapping methods, suggested by Chris Leech
* printk warning fixups from Andrew Morton
* don't use inline in C files, Adrian Bunk
* select the API when MD is enabled
* BUG_ON xor source counts <= 1
* implicitly handle hardware concerns like channel switching and
interrupts, Neil Brown
* remove the per operation type list, and distribute operation capabilities
evenly amongst the available channels
* simplify async_tx_find_channel to optimize the fast path
* introduce the channel_table_initialized flag to prevent early calls to
the api
* reorganize the code to mimic crypto
* include mm.h as not all archs include it in dma-mapping.h
* make the Kconfig options non-user visible, Adrian Bunk
* move async_tx under crypto since it is meant as 'core' functionality, and
the two may share algorithms in the future
* move large inline functions into c files
* checkpatch.pl fixes
* gpl v2 only correction
Cc: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-By: NeilBrown <neilb@suse.de>
Dan Williams [Mon, 9 Jul 2007 18:56:42 +0000 (11:56 -0700)]
xor: make 'xor_blocks' a library routine for use with async_tx
The async_tx api tries to use a dma engine for an operation, but will fall
back to an optimized software routine otherwise. Xor support is
implemented using the raid5 xor routines. For organizational purposes this
routine is moved to a common area.
The following fixes are also made:
* rename xor_block => xor_blocks, suggested by Adrian Bunk
* ensure that xor.o initializes before md.o in the built-in case
* checkpatch.pl fixes
* mark calibrate_xor_blocks __init, Adrian Bunk
Cc: Adrian Bunk <bunk@stusta.de> Cc: NeilBrown <neilb@suse.de> Cc: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Dan Williams [Mon, 9 Jul 2007 18:56:42 +0000 (11:56 -0700)]
dmaengine: make clients responsible for managing channels
The current implementation assumes that a channel will only be used by one
client at a time. In order to enable channel sharing the dmaengine core is
changed to a model where clients subscribe to channel-available-events.
Instead of tracking how many channels a client wants and how many it has
received the core just broadcasts the available channels and lets the
clients optionally take a reference. The core learns about the clients'
needs at dma_event_callback time.
In support of multiple operation types, clients can specify a capability
mask to only be notified of channels that satisfy a certain set of
capabilities.
Changelog:
* removed DMA_TX_ARRAY_INIT, no longer needed
* dma_client_chan_free -> dma_chan_release: switch to global reference
counting only at device unregistration time, before it was also happening
at client unregistration time
* clients now return dma_state_client to dmaengine (ack, dup, nak)
* checkpatch.pl fixes
* fixup merge with git-ioat
Cc: Chris Leech <christopher.leech@intel.com> Signed-off-by: Shannon Nelson <shannon.nelson@intel.com> Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-by: David S. Miller <davem@davemloft.net>
Dan Williams [Tue, 2 Jan 2007 18:10:43 +0000 (11:10 -0700)]
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org> Cc: Chris Leech <christopher.leech@intel.com> Signed-off-by: Shannon Nelson <shannon.nelson@intel.com> Signed-off-by: Dan Williams <dan.j.williams@intel.com> Acked-by: David S. Miller <davem@davemloft.net>
Dan Aloni [Thu, 8 Mar 2007 17:57:36 +0000 (09:57 -0800)]
I/OAT: fix I/OAT for kexec
Under kexec, I/OAT initialization breaks over busy resources because the
previous kernel did not release them.
I'm not sure this fix can be considered a complete one but it works for me.
I guess something similar to the *_remove method should occur there..
Signed-off-by: Dan Aloni <da-x@monatomic.org> Signed-off-by: Chris Leech <christopher.leech@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Chris Leech [Thu, 8 Mar 2007 17:57:36 +0000 (09:57 -0800)]
I/OAT: Only offload copies for TCP when there will be a context switch
The performance wins come with having the DMA copy engine doing the copies
in parallel with the context switch. If there is enough data ready on the
socket at recv time just use a regular copy.
Signed-off-by: Chris Leech <christopher.leech@intel.com>
Chris Leech [Thu, 8 Mar 2007 17:57:35 +0000 (09:57 -0800)]
ioatdma: Remove the use of writeq from the ioatdma driver
There's only one now anyway, and it's not in a performance path,
so make it behave the same on 32-bit and 64-bit CPUs.
Signed-off-by: Chris Leech <christopher.leech@intel.com>
Chris Leech [Thu, 8 Mar 2007 17:57:33 +0000 (09:57 -0800)]
ioatdma: Push pending transactions to hardware more frequently
Every 20 descriptors turns out to be to few append commands with
newer/faster CPUs. Pushing every 4 still cuts down on MMIO writes to an
acceptable level without letting the DMA engine run out of work.
Signed-off-by: Chris Leech <christopher.leech@intel.com>
Fix permission checking for the new utimensat() system call
Commit 1c710c896eb461895d3c399e15bb5f20b39c9073 added the utimensat()
system call, but didn't handle the case of checking for the writability
of the target right, when the target was a file descriptor, not a
filename.
We cannot use vfs_permission(MAY_WRITE) for that case, and need to
simply check whether the file descriptor is writable. The oops from
using the wrong function was noticed and narrowed down by Markus
Trippelsdorf.
Cc: Ulrich Drepper <drepper@redhat.com> Cc: Markus Trippelsdorf <markus@trippelsdorf.de> Cc: Andrew Morton <akpm@linux-foundation.org> Acked-by: Al Viro <viro@ftp.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
PIO4 is a maximum PIO mode supported by a driver. Using "255" as a max_mode
argument to ide_get_best_pio_mode() could result in wrong timings being used
by a driver (for "pio" equal to 5) or OOPS (for "pio" values > 5 && < 255).
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com> Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com> Reviewed-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
If the chipset is using this PCI-ID, the primary channel is connected to the
first PATA-port. The secondary channel is connected to SATA-ports in IDE
emulation mode. The legacy IO-ports are used.
The including of the PCI-ID into pata_sis is not sufficient, because the legacy
driver in drivers/ide is initialized before pata_sis.
Adrian Bunk [Sat, 7 Jul 2007 04:14:02 +0000 (06:14 +0200)]
DLM must depend on SYSFS
The dependency of DLM on SYSFS got lost in
commit 6ed7257b46709e87d79ac2b6b819b7e0c9184998 resulting in the
following compile error with CONFIG_DLM=y, CONFIG_SYSFS=n:
<-- snip -->
...
LD .tmp_vmlinux1
fs/built-in.o: In function `dlm_lockspace_init':
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/fs/dlm/lockspace.c:231: undefined reference to `kernel_subsys'
fs/built-in.o: In function `configfs_init':
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/fs/configfs/mount.c:143: undefined reference to `kernel_subsys'
make[1]: *** [.tmp_vmlinux1] Error 1
<-- snip -->
Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Adrian Bunk [Fri, 6 Jul 2007 22:54:09 +0000 (00:54 +0200)]
include/linux/kallsyms.h must #include <linux/errno.h>
This patch fixes the following 2.6.22 regression with CONFIG_KALLSYMS=n:
<-- snip -->
...
CC arch/m32r/kernel/traps.o
In file included from /home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/arch/m32r/kernel/traps.c:14:
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/include/linux/kallsyms.h: In function 'lookup_symbol_name':
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/include/linux/kallsyms.h:66: error: 'ERANGE' undeclared (first use in this function)
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/include/linux/kallsyms.h:66: error: (Each undeclared identifier is reported only once
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/include/linux/kallsyms.h:66: error: for each function it appears in.)
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/include/linux/kallsyms.h: In function 'lookup_symbol_attrs':
/home/bunk/linux/kernel-2.6/linux-2.6.22-rc6-mm1/include/linux/kallsyms.h:71: error: 'ERANGE' undeclared (first use in this function)
make[2]: *** [arch/m32r/kernel/traps.o] Error 1
<-- snip -->
Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
David Woodhouse [Sat, 7 Jul 2007 18:58:39 +0000 (14:58 -0400)]
Fix use-after-free oops in Bluetooth HID.
When cleaning up HIDP sessions, we currently close the ACL connection
before deregistering the input device. Closing the ACL connection
schedules a workqueue to remove the associated objects from sysfs, but
the input device still refers to them -- and if the workqueue happens to
run before the input device removal, the kernel will oops when trying to
look up PHYSDEVPATH for the removed input device.
Fix this by deregistering the input device before closing the
connections.
Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
[NETPOLL]: Fixups for 'fix soft lockup when removing module'
[NET]: net/core/netevent.c should #include <net/netevent.h>
[NETFILTER]: nf_conntrack_h323: add checking of out-of-range on choices' index values
[NET] skbuff: remove export of static symbol
SCTP: Add scope_id validation for link-local binds
SCTP: Check to make sure file is valid before setting timeout
SCTP: Fix thinko in sctp_copy_laddrs()
He correctly updated many of the function definitions that were using this
extra regs pointer parameter but forgot to update some caller sites of
those functions. The reason the modifications was not properly done on all
drivers is that some drivers were rarely compiled because they are for
AMIGA, or that some code sites were inside #ifdefs where the option is not
set or inside #if 0.
Here is the semantic patch that found the occurences
and fixed the problem.
arch/i386/kernel/built-in.o(.init.text+0x4fa9): In function `acpi_parse_madt':
: undefined reference to `acpi_madt_oem_check'
arch/i386/kernel/built-in.o(.init.text+0x7406): In function `smp_read_mpc':
: undefined reference to `mps_oem_check'
arch/i386/kernel/built-in.o(.init.text+0x8990): In function
`connect_bsp_APIC':
: undefined reference to `enable_apic_mode'
make: *** [.tmp_vmlinux1] Error 1
o Fix the build issue. Provided the definitions of missing functions.
o Don't have ES7000 machine. Only compile tested.
Cc: Len Brown <lenb@kernel.org> Cc: Natalie Protasevich <protasnb@gmail.com> Cc: Roland Dreier <rolandd@cisco.com> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
PNP SMCf010 quirk: work around Toshiba Portege 4000 ACPI issues
When we enable the SMCf010 IR device, the Toshiba Portege 4000 BIOS claims
the device is working, but it really isn't configured correctly. The BIOS
*will* configure it, but only if we call _SRS after (1) reversing the order
of the SIR and FIR I/O port regions and (2) changing the IRQ from
active-high to active-low.
This patch addresses the 2.6.22 regression:
"no irda0 interface (2.6.21 was OK), smsc does not find chip"
I tested this on a Portege 4000. The smsc-ircc2 driver correctly detects
the device, and "irattach irda0 -s && irdadump" shows transmitted and
received packets.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com> Cc: Andrey Borzenkov <arvidjaar@mail.ru> Cc: Samuel Ortiz <samuel@sortiz.org> Cc: "Linus Walleij (LD/EAB)" <linus.walleij@ericsson.com> Cc: Michal Piotrowski <michal.k.k.piotrowski@gmail.com> Cc: Adam Belay <ambx1@neo.rr.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
David Woodhouse [Fri, 6 Jul 2007 09:39:52 +0000 (02:39 -0700)]
x86_64: fix headers_install
A bug in headers_install for ARCH=x86_64 yields an asm/ directory full of
files all of which are using the same #ifdef guard, "__ASM_STUB_" with no
postfix. So the second and later asm files #included in the same C file
(often through standard headers like ioctl.h) yields no symbols.
Strangeness with the Ubuntu 'tell me if I support something that's not
explcitly mentioned in POSIX, and I'll strip it out' shell, I believe.
We don't need the 'export' but we do need a semicolon at the end of the
FNAME line:
Signed-off-by: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Rob Landley <rob@landley.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
MTRR: Fix race causing set_mtrr to go into infinite loop
Processors synchronization in set_mtrr requires the .gate field to be set
after .count field is properly initialized. Without an explicit barrier,
the compiler was reordering those memory stores. That was sometimes
causing a processor (in ipi_handler) to see the .gate change and decrement
.count before the latter is set by set_mtrr() (which then hangs in a
infinite loop with irqs disabled).
Jason Wessel [Fri, 6 Jul 2007 09:39:50 +0000 (02:39 -0700)]
i386: fix regression, endless loop in ptrace singlestep over an int80
The commit 635cf99a80f4ebee59d70eb64bb85ce829e4591f introduced a
regression. Executing a ptrace single step after certain int80
accesses will infinitely loop and never advance the PC.
The TIF_SINGLESTEP check should be done on the return from the syscall
and not before it.
I loops on each single step on the pop right after the int80 which writes out
to the console. At that point you can issue as many single steps as you want
and it will not advance any further.
The test case is below:
/* Test whether singlestep through an int80 syscall works.
*/
#define _GNU_SOURCE
#include <stdio.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/ptrace.h>
#include <sys/wait.h>
#include <sys/mman.h>
#include <asm/user.h>
#include <string.h>
static int child, status;
static struct user_regs_struct regs;
Fix elf_core_dump() when writing arch specific notes (spu coredumps)
elf_core_dump() supports dumping arch specific ELF notes, via the #define
ELF_CORE_WRITE_EXTRA_NOTES. Currently the only user of this is the powerpc
spu coredump code.
There is a bug in the handling of foffset WRT the arch notes, which causes
us to erroneously increment foffset by the size of the arch notes, leaving
a block of zeroes in the file, and causing all subsequent data in the file
to be at <supposed position> + <arch note size>. eg:
LOAD 0x050000 0x00100000 0x00000000 0x20000 0x20000 R E 0x10000
Tells us we should have a chunk of data at 0x50000. The truth is the data
is at 0x90dbc = 0x50000 + 0x40dbc (the size of the arch notes).
This bug prevents gdb from reading the core file correctly.
The simplest fix is to simply remember the size of the arch notes, and add
it to foffset after we've written the arch notes. The only drawback is
that if the arch code doesn't write as many bytes as it said it would, we
end up with a broken core dump again. For now I think that's a reasonable
requirement.
Tested on a Cell blade, gdb no longer complains about the core file being
bogus.
While I'm here I should point out that the spu coredump code does not work
if we're dumping to a pipe - we'll have to wait for 23 to fix that.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Acked-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ralf Baechle [Wed, 20 Jun 2007 23:22:34 +0000 (00:22 +0100)]
[MIPS] Fix scheduling latency issue on 24K, 34K and 74K cores
The idle loop goes to sleep using the WAIT instruction if !need_resched().
This has is suffering from from a race condition that if if just after
need_resched has returned 0 an interrupt might set TIF_NEED_RESCHED but
we've just completed the test so go to sleep anyway. This would be
trivial to fix by just disabling interrupts during that sequence as in:
local_irq_disable();
if (!need_resched())
__asm__("wait");
local_irq_enable();
but the processor architecture leaves it undefined if a processor calling
WAIT with interrupts disabled will ever restart its pipeline and indeed
some processors have made use of the freedom provided by the architecture
definition. This has been resolved and the Config7.WII bit indicates that
the use of WAIT is safe on 24K, 24KE and 34K cores. It also is safe on
74K starting revision 2.1.0 so enable the use of WAIT with interrupts
disabled for 74K based on a c0_prid of at least that.
Older processors used to encode processor version and revision in two
4-bit bitfields, the 4K seems to simply count up and even newer MTI cores
have switched to use the 8-bits as 3:3:2 bitfield with the last field as
the patch number.
The RM7000 processors and the E9000 cores have a bug (though PMC-Sierra
opposes it being called that) where invalid instructions in the same
I-cache line worth of instructions being fetched may case spurious
exceptions.
The workaround for this was only enabled for E9000 cores; enable it also
for all RM7000-based platforms.
[NETPOLL]: Fixups for 'fix soft lockup when removing module'
>From my recent patch:
> > #1
> > Until kernel ver. 2.6.21 (including) cancel_rearming_delayed_work()
> > required a work function should always (unconditionally) rearm with
> > delay > 0 - otherwise it would endlessly loop. This patch replaces
> > this function with cancel_delayed_work(). Later kernel versions don't
> > require this, so here it's only for uniformity.
But Oleg Nesterov <oleg@tv-sign.ru> found:
> But 2.6.22 doesn't need this change, why it was merged?
>
> In fact, I suspect this change adds a race,
...
His description was right (thanks), so this patch reverts #1.
Signed-off-by: Jarek Poplawski <jarkao2@o2.pl> Signed-off-by: David S. Miller <davem@davemloft.net>
[NETFILTER]: nf_conntrack_h323: add checking of out-of-range on choices' index values
Choices' index values may be out of range while still encoded in the fixed
length bit-field. This bug may cause access to undefined types (NULL
pointers) and thus crashes (Reported by Zhongling Wen).
This patch also adds checking of decode flag when decoding SEQUENCEs.
Signed-off-by: Jing Min Zhao <zhaojingmin@vivecode.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
SCTP: Add scope_id validation for link-local binds
SCTP currently permits users to bind to link-local addresses,
but doesn't verify that the scope id specified at bind matches
the interface that the address is configured on. It was report
that this can hang a system.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
SCTP: Check to make sure file is valid before setting timeout
In-kernel sockets created with sock_create_kern don't usually
have a file and file descriptor allocated to them. As a result,
when SCTP tries to check the non-blocking flag, we Oops when
dereferencing a NULL file pointer.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* master.kernel.org:/home/rmk/linux-2.6-arm:
[ARM] always allow dump_stack() to produce a backtrace
[ARM] Fix non-page aligned boot time mappings
[ARM] 4458/1: pxa: Fix CKEN usage and hence fix pxa suspend/resume
[ARM] 4454/1: Use word accesses in Versatile PCI config reads
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
Input: document some of keycodes
Input: add a new EV_SW SW_RADIO event, for radio switches on laptops
Input: serio - take drv_mutex in serio_cleanup()
Input: atkbd - use printk_ratelimit for spurious ACK messages
Input: atkbd - throttle LED switching
Input: i8042 - add HP Pavilion ZT1000 to the MUX blacklist
Merge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
[POWERPC] Update defconfigs
[POWERPC] Uninline and export virq_to_hw() for the pasemi_mac driver
[POWERPC] Fix PMI breakage in cbe_cbufreq driver
[POWERPC] Disable old EMAC driver in arch/powerpc
David Woodhouse [Thu, 5 Jul 2007 01:26:44 +0000 (21:26 -0400)]
Fix slab redzone alignment
Commit b46b8f19c9cd435ecac4d9d12b39d78c137ecd66 fixed a couple of bugs
by switching the redzone to 64 bits. Unfortunately, it neglected to
ensure that the _second_ redzone, after the slab object, is aligned
correctly. This caused illegal instruction faults on sparc32, which for
some reason not entirely clear to me are not trapped and fixed up.
Two things need to be done to fix this:
- increase the object size, rounding up to alignof(long long) so
that the second redzone can be aligned correctly.
- If SLAB_STORE_USER is set but alignof(long long)==8, allow a
full 64 bits of space for the user word at the end of the buffer,
even though we may not _use_ the whole 64 bits.
This patch should be a no-op on any 64-bit architecture or any 32-bit
architecture where alignof(long long) == 4. Of the others, it's tested
on ppc32 by myself and a very similar patch was tested on sparc32 by
Mark Fortescue, who reported the new problem.
Also, fix the conditions for FORCED_DEBUG, which hadn't been adjusted to
the new sizes. Again noticed by Mark.
Signed-off-by: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Yeah, we could have just disabled it, but there's work on a new one that
isn't as fundamentally broken, so there really doesn't seem to be any
point in keeping it around.
The recent timer cleanup broke the only valid use, and when I say
"valid", I obviously mean "totally broken". So it's not like it works,
or really even can work in the current format that uses the unsafe
"panic" LED blinking routines..
Russell King [Wed, 4 Jul 2007 20:16:33 +0000 (21:16 +0100)]
[ARM] Fix non-page aligned boot time mappings
AT91SAM9260 stopped booting with the recent changes to MM
initialisation - it was asking for a non-aligned virtual address
which caused loops to be non-terminal. Fix this by rounding
virtual addresses down, but remember to include the offset in
the length, and round the length up to the following page.
This means that asking for a mapping of 4K starting at 2K into
a page maps two pages as one would expect.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Merge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
[MIPS] VSMP: Fix initialization ordering bug.
[MIPS] Add whitelists for checksyscalls.sh
[MIPS] die(): Properly declare as non-returning
[MIPS] Fix include wrapper symbol definitions in IP32 code.
This marks the declaration of die() correctly, removing "control reaches
end of non-void function" warnings from non-void functions that die() at
the end.
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org> Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
David Woodhouse [Tue, 3 Jul 2007 20:51:19 +0000 (16:51 -0400)]
[JFFS2] Fix readinode failure when read_dnode() detects CRC failure.
We should have stopped returning 1 from read_dnode() to indicate
failure. We can just mark the damn thing obsolete immediately. But I
missed a case where we don't.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
When Andi reverted the HPET resource reservation (in commit 0f8dc2f06560e2ca126d1670a24126ba08357d38), he didn't remove the now
unused variables, which just causes gcc to be noisy.
Zach Brown [Tue, 3 Jul 2007 22:28:55 +0000 (15:28 -0700)]
dio: remove bogus refcounting BUG_ON
Badari Pulavarty reported a case of this BUG_ON is triggering during
testing. It's completely bogus and should be removed.
It's trying to notice if we left references to the dio hanging around in
the sync case. They should have been dropped as IO completed while this
path was in dio_await_completion(). This condition will also be
checked, via some twisty logic, by the BUG_ON(ret != -EIOCBQUEUED) a few
lines lower. So to start this BUG_ON() is redundant.
More fatally, it's dereferencing dio-> after having dropped its
reference. It's only safe to dereference the dio after releasing the
lock if the final reference was just dropped. Another CPU might free
the dio in bio completion and reuse the memory after this path drops the
dio lock but before the BUG_ON() is evaluated.
This patch passed aio+dio regression unit tests and aio-stress on ext3.
Signed-off-by: Zach Brown <zach.brown@oracle.com> Cc: Badari Pulavarty <pbadari@us.ibm.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Andi Kleen [Fri, 29 Jun 2007 14:16:36 +0000 (16:16 +0200)]
Revert HPET resource reservation
Matthias Lenk reports that the PCI subsystem would move the HPET on
SB400/SB600-based systems, where the HPET is in BAR1 of the SMbus
controller.
The reason? The ACPI layer registered the PCI MMIO range as being busy
too early, before PCI enumeration had happened, causing the PCI layer to
decide that it should relocate the resources somewhere else.
Firmware resources should be marked busy _after_ the PCI enumeration and
probing has happened, not before.
Remove the too-early reservation, we'll fix it up to do it properly
later. In the meantime, this solves the regression.
Merge branch 'master' of ssh://master.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb
* 'master' of ssh://master.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb:
V4L/DVB (5822): Fix the return value in ttpci_budget_init()
V4L/DVB (5818): CinergyT2: fix flush_workqueue() vs work->func() deadlock
V4L/DVB (5816): Cx88-blackbird: fix vidioc_g_tuner never ending list of tuners
V4L/DVB (5808): Bttv: fix v4l1 breaking the driver
SLUB: Make lockdep happy by not calling add_partial with interrupts enabled during bootstrap
If we move the local_irq_enable() to the end of the function then
add_partial() in early_kmem_cache_node_alloc() will be called
with interrupts disabled like during regular operations.
This makes lockdep happy.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Tested-by: Andre Noll <maan@systemlinux.org> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Thomas Gleixner [Tue, 3 Jul 2007 18:05:20 +0000 (20:05 +0200)]
NTP: remove clock_was_set() call to prevent deadlock
The clock_was_set() call in seconds_overflow() which happens only when
leap seconds are inserted / deleted is wrong in two aspects:
1. it results in a call to on_each_cpu() with interrupts disabled
2. it is potential deadlock source vs. call_lock in smp_call_function()
The only possible side effect of the removal might be, that an absolute
CLOCK_REALTIME timer fires 1 second too late, in the rare case of leap
second deletion and an absolute CLOCK_REALTIME timer which expires in
the affected time frame. It will never fire too early.
This was probably observed by the reporter of a June 30th -> July 1st
hang: http://lkml.org/lkml/2007/7/3/103
A similar problem was observed by Dave Jones, who provided a screen shot
with a lockdep back trace, which allowed to analyse the problem.
Andrew Morton [Tue, 3 Jul 2007 20:28:36 +0000 (22:28 +0200)]
ide: ide_scan_pcibus(): check __pci_register_driver return value
drivers/ide/setup-pci.c: In function 'ide_scan_pcibus':
drivers/ide/setup-pci.c:879: warning: ignoring return value of '__pci_register_driver', declared with attribute warn_unused_result
Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Albert Lee [Tue, 3 Jul 2007 20:28:36 +0000 (22:28 +0200)]
ide: pdc202xx_new PLL input clock fix
Recently the PLL input clock of Promise 2027x is sometimes detected
higher than expected (e.g. 20.027 MHz compared to 16.714 MHz).
It seems sometimes the mdelay() function is not as precise as it
used to be. Per Alan's advice, HT or power management might affect
the precision of mdelay().
This patch calls gettimeofday() to measure the time elapsed and
calculate the PLL input clock accordingly.
Signed-off-by: Albert Lee <albertcc@tw.ibm.com> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Bahadir Balban <bahadir.balban@gmail.com> Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com> Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Sergei Shtylyov [Tue, 3 Jul 2007 20:28:35 +0000 (22:28 +0200)]
hpt366: use correct enablebits for HPT36x
The HPT36x chips finally turned out to have the channel enable bits -- however,
badly implemented. Make use of them despite it's probably only going to burden
the driver's code -- assuming both channels are always enabled by the HighPoint
BIOS anyway...
/*
* Last sector was transfered, wait until drive is ready.
* This can take up to 10 usec, but we will wait max 1 ms
* (drive_cmd_intr() waits that long).
*/
while (((stat = hwif->INB(IDE_STATUS_REG)) & BUSY_STAT) && retries--)
udelay(10);
if (!retries)
printk(KERN_ERR "%s: drive still BUSY!\n", drive->name);
return stat;
}
`printk' is never called because `retries' never holds zero at the
outside of `while' loop: when `retries' holds zero at the while's loop
condition, `retries' will hold -1 at the if condition.
V4L/DVB (5818): CinergyT2: fix flush_workqueue() vs work->func() deadlock
Spotted and tested by Thomas Sattler <tsattler@gmx.de>.
cinergyT2.c does cancel_delayed_work() + flush_scheduled_work() while
holding cinergyt2->sem. This leads to deadlock because work->func()
needs the same mutex to complete. Another bug is that this code in fact
can't reliably stop the re-arming delayed_work.
Convert this code to use cancel_rearming_delayed_work() and move it
out of ->sem. Another mutex, ->wq_sem, was added to protect against the
concurrent open/resume.
This patch is a horrible hack to fix the lockup which happens in practice.
As Dmitry Torokhov pointed out this driver has other problems and needs
further changes.
Jelle Foks [Mon, 21 May 2007 17:56:17 +0000 (14:56 -0300)]
V4L/DVB (5816): Cx88-blackbird: fix vidioc_g_tuner never ending list of tuners
v4l-info and other programs would loop indefinitely while querying the
tuners for cx88-blackbird cards.
The cause was that vidioc_g_tuner didn't return an error value for
qctrl->id != 0, making the application think there is a never ending
list of tuners...
This patch adds the same index check as done in vidioc_g_tuner() in
cx88-video.
V4L/DVB (5808): Bttv: fix v4l1 breaking the driver
If one uses a V4L *one* application, such as vlc or mplayer's v4l driver, as
the first user after the driver is loaded, the driver wedges itself and will
never capture properly. Even if one uses a V4L2 application later, it still
won't work.
If one uses a V4L *two* application first, such as tvtime or mplayer's v4l2
driver, then the driver will be ok. One can then run a V4L1 application, and
it will work.
It turns out the problem is with norm changing and the crop support that was
added in 2.6.21. The driver defaults to PAL, and keeps the last norm it was
set too across opens. If one changes the norm via V4L1, the cropping
parameters are not reset like they should be, and they'll remain broken across
device opens.
This patch removes the direct setting of btv->tvnorm in the V4L1 ioctl
VIDIOCSCHAN handler. The norm is set via the existing call to set_input(),
which calls set_tvnorm(), which will reset the cropping values now that it is
able to detect the norm change.
host->irq and host->irq2 should be set before ata_host_register() for
IRQ reporting to work. Move up host->irq assignment in
ata_host_activate() and add it to ata_pci_init_one() native path and
pata_cs5520.
The port info printing in ata_host_register() doesn't fit all the
different controllers. It should probably be moved out to LLDs with
some helpers in the future.
Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/cooloney/blackfin-2.6
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/cooloney/blackfin-2.6:
Blackfin arch: remove zero-sized include/asm-blackfin/macros.h
Blackfin arch: update board defconfig files
Blackfin arch: Fix up remaining printks with proper log levels
Blackfin arch: Add proper -mcpu option according to the cpu and silicon revision configuration
Merge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
3c589_cs: fix local_bh_enable warning
RESEND [PATCH 3/3] NetXen: Graceful teardown of interface and hardware upon module unload
drivers/net/ns83820.c: fix a check-after-use
net/usb/cdc_ether minor sparse cleanup
RESEND [PATCH 2/3] NetXen: Support per PCI-function interrupt mask registers
RESEND [PATCH 1/3] NetXen: Fix issue of MSI not working correctly
dm9601: Return 0 from bind() on success
Update MAINTAINERS for USB network devices
usbnet: Zero padding byte if there is tail room in skb
dm9601: HW header size shouldn't be included in packet length
starfire list alpha as 64 bit arch
myri10ge: SET_NETDEV_DEV()
gianfar: Fix typo bug introduced by move to udp_hdr()
[PATCH] libertas: remove private ioctls
[PATCH] libertas: fix WPA associations by handling ENABLE_RSN correctly
[PATCH] libertas: kill wlan_scan_process_results
[PATCH] libertas: style fixes
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
firewire: add Kconfig help on building both stacks
firewire: fix async reception on big endian machines
projects / mirror_ubuntu-disco-kernel.git / log