Merge branch 'x86/uv' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'x86/uv' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: UV BAU distribution and payload MMRs
x86: UV: BAU partition-relative distribution map
x86, uv: add Kconfig dependency on NUMA for UV systems
x86: prevent /sys/firmware/sgi_uv from being created on non-uv systems
x86, UV: Fix for nodes with memory and no cpus
x86, UV: system table in bios accessed after unmap
x86: UV BAU messaging timeouts
x86: UV BAU and nodes with no memory
Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
ata: Report 16/32bit PIO as best we can
libata: use ATA_ID_CFA_*
pata_legacy: fix no device fail path
pata_hpt37x: fix HPT370 DMA timeouts
libata: handle SEMB signature better
Tetsuo Handa reports seeing the WARN_ON(current->mm == NULL) in
security_vm_enough_memory(), when do_execve() is touching the
target mm's stack, to set up its args and environment.
Yes, a UMH_NO_WAIT or UMH_WAIT_PROC call_usermodehelper() spawns
an mm-less kernel thread to do the exec. And in any case, that
vm_enough_memory check when growing stack ought to be done on the
target mm, not on the execer's mm (though apart from the warning,
it only makes a slight tweak to OVERCOMMIT_NEVER behaviour).
Tetsuo Handa, running a kernel with CONFIG_DEBUG_PAGEALLOC=y and
CONFIG_UEVENT_HELPER_PATH=/sbin/hotplug, has been hitting RCU detected
CPU stalls: it's been spinning in the loop where do_execve() counts up
the args (but why wasn't fixup_exception working? dunno).
The recent change, switching kobject_uevent_env() from UMH_WAIT_EXEC
to UMH_NO_WAIT, is broken: the exec uses args on the local stack here,
and an env which is kfreed as soon as call_usermodehelper() returns.
It very much needs to wait for the exec to be done.
An alternative would be to keep the UMH_NO_WAIT, and complicate the code
to allocate and free these resources correctly? but no, as GregKH
pointed out when making the commit, CONFIG_UEVENT_HELPER_PATH="" is a
much better optimization - though some distros are still saying
/sbin/hotplug in their .config, yet with no such binary in their initrd
or their root.
Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Acked-by: Greg Kroah-Hartman <gregkh@suse.de> Acked-by: Arjan van de Ven <arjan@linux.intel.com> Acked-by: Will Newton <will.newton@gmail.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Alan Cox [Thu, 9 Apr 2009 16:31:17 +0000 (17:31 +0100)]
ata: Report 16/32bit PIO as best we can
The legacy old IDE ioctl API for this is a bit primitive so we try
and map stuff sensibly onto it.
- Set PIO over DMA devices to report 32bit
- Add ability to change the PIO32 settings if the controller permits it
- Add that functionality into the sff drivers
- Add that functionality into the VLB legacy driver
- Turn on the 32bit PIO on the ninja32 and add support there
Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
When pata_legacy can't detect any device, it unregisters the
platform_device and fails detection. However, it forgets to detach
ata host triggering weird failures as the host later gets freed by
devres while still attached. Fix it.
Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Sergei Shtylyov [Tue, 14 Apr 2009 14:39:14 +0000 (18:39 +0400)]
pata_hpt37x: fix HPT370 DMA timeouts
The libata driver has copied the code from the IDE driver which caused a post
2.4.18 regression on many HPT370[A] chips -- DMA stopped to work completely,
only causing timeouts. Now remove hpt370_bmdma_start() for good...
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
WDC WD1600JS-62MHB5 successfully hits the window between ATA/ATAPI-7
and Serial ATA II standards and reports 3c/c3 signature which now is
assigned to SEMB. Make ata_dev_classify() report ATA_DEV_SEMB on the
sig and let ata_dev_read_id() work around it by trying IDENTIFY once.
This fixes bko#11579.
Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: David Haun <drhaun88@gmail.com> Reported-by: Lars Wirzenius <liw@liw.fi> Reported-by: Juan Manuel <jmcarranza@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
This patch correctly sets BAU memory mapped registers to point
to the sending activation descriptor table and target payload table.
The "Broadcast Assist Unit" is used for TLB shootdown in UV.
The memory mapped registers that point to sending and receiving
memory structures contain node numbers.
In one case the __pa() function did not provide the node id of
memory on blade zero in configurations where that id is nonzero.
In another case, it was assumed that memory was allocated on
the local node. That assumption is not true in a configuration
in which the node has no memory.
Tested on the UV hardware simulator.
[ Impact: fix possible runtime crash due to incorrect TLB logic ]
Chris Mason [Wed, 15 Apr 2009 17:22:38 +0000 (13:22 -0400)]
Add block_write_full_page_endio for passing endio handler
block_write_full_page doesn't allow the caller to control what happens
when the IO is over. This adds a new call named block_write_full_page_endio
so the buffer head end_io handler can be provided by the caller.
This will be used by the ext3 data=guarded mode to do i_size updates in
a workqueue based end_io handler. end_buffer_async_write is also
exported so it can be called to do the dirty work of managing page
writeback for the higher level end_io handler.
Signed-off-by: Chris Mason <chris.mason@oracle.com> Acked-by: Theodore Tso <tytso@mit.edu> Acked-by: Jan Kara <jack@suse.cz> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (64 commits)
phylib: Fix delay argument of schedule_delayed_work
NET/ixgbe: Fix powering off during shutdown
NET/e1000e: Fix powering off during shutdown
NET/e1000: Fix powering off during shutdown
packet: avoid warnings when high-order page allocation fails
gianfar: stop send queue before resetting gianfar
myr10ge: again fix lro_gen_skb() alignment
declance: convert to net_device_ops
bfin_mac: convert to net_device_ops
au1000: convert to net_device_ops
atarilance: convert to net_device_ops
a2065: convert to net_device_ops
ixgbe: update real_num_tx_queues on changing num_rx_queues
ixgbe: fix tx queue index
Revert "rose: zero length frame filtering in af_rose.c"
sfc: Use correct macro to set event bitfield
sfc: Match calls to netif_napi_add() and netif_napi_del()
bonding: Remove debug printk
e1000/e1000: fix compile warning
ehea: Fix incomplete conversion to net_device_ops
...
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
sparc: remove some pointless conditionals before kfree()
sbus: changed ioctls to unlocked
sparc: asm/atomic.h on 32bit should include asm/system.h for xchg
sparc64: Fix smp_callin() locking.
phylib: Fix delay argument of schedule_delayed_work
The commit a390d1f3 ("phylib: convert state_queue work to
delayed_work") missed converting 'expires' value to 'delay' value.
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp> Acked-by: Marcin Slusarz <marcin.slusarz@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Prevent ixgbe from putting the adapter into D3 during shutdown except when
we're going to power off the system, since doing that may generally cause
problems with kexec to happen (such problems were observed for igb and
forcedeth). For this purpose seperate ixgbe_shutdown() from ixgbe_suspend()
and use the appropriate PCI PM callbacks in both of them.
Signed-off-by: "Rafael J. Wysocki" <rjw@sisk.pl> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Prevent e1000e from putting the adapter into D3 during shutdown except when
we're going to power off the system, since doing that may generally cause
problems with kexec to happen (such problems were observed for igb and
forcedeth). For this purpose seperate e1000e_shutdown() from e1000e_suspend()
and use the appropriate PCI PM callbacks in both of them.
Signed-off-by: "Rafael J. Wysocki" <rjw@sisk.pl> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Prevent e1000 from putting the adapter into D3 during shutdown except when
we're going to power off the system, since doing that may generally cause
problems with kexec to happen (such problems were observed for igb and
forcedeth). For this purpose seperate e1000_shutdown() from e1000_suspend()
and use the appropriate PCI PM callbacks in both of them.
Signed-off-by: "Rafael J. Wysocki" <rjw@sisk.pl> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
as it upsets some versions of gcc under some circumstances:
kernel/rcupdate.c: In function `_rcu_barrier':
kernel/rcupdate.c:125: sorry, unimplemented: inlining failed in call to 'wait_migrated_callbacks': function body not available
kernel/rcupdate.c:152: sorry, unimplemented: called from here
This can be dealt with by simply putting the static variables (rcu_migrate_*)
at the top, and moving the implementation of the function up so that it
replaces its forward declaration.
Signed-off-by: David Howells <dhowells@redhat.com> Cc: Dipankar Sarma <dipankar@in.ibm.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
David Howells [Wed, 15 Apr 2009 18:34:51 +0000 (19:34 +0100)]
MN10300: Stop gcc from generating uninitialised variable warnings after BUG()
Stop gcc from generating uninitialised variable warnings after BUG(). The
problem is that MN10300's implementation of BUG() invokes system call 15 which
doesn't return - but there's no way to tell the compiler that and also emit the
bug table element with the correct file and line data.
So instead, we make the do...while wrapper in _debug_bug_trap() an endless loop
from which there's no escape.
Also, while we're at it, (1) get rid of _debug_bug_trap() and just implement
directly as BUG(), and (2) make the implementation of BUG() contingent on
CONFIG_BUG=y.
Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge branch 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6
* 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6:
[S390] boot cputime accounting
[S390] add read_persistent_clock
[S390] cpu hotplug and accounting values
[S390] fix idle time accounting
[S390] smp: fix cpu_possible_map initialization
[S390] dasd: fix idaw boundary checking for track based ccw
[S390] dasd: Use the new async framework for autoonlining.
[S390] qdio: remove dead timeout handler
[S390] appldata: Use new mod_virt_timer_periodic() function.
[S390] extend virtual timer interface by mod_virt_timer_periodic
[S390] stp synchronization retry timer
[S390] call nmi_enter/nmi_exit on machine checks
[S390] wire up preadv/pwritev system calls
[S390] s390: move machine flags to lowcore
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
ALSA: hda - Fix the cmd cache keys for amp verbs
ALSA: add missing definitions(letters) to HD-Audio.txt
ALSA: hda - Add quirk mask for Fujitsu Amilo laptops with ALC883
[ALSA] intel8x0: add one retry to the ac97_clock measurement routine
[ALSA] intel8x0: fix wrong conditions in ac97_clock measure routine
ALSA: hda - Avoid call of snd_jack_report at release
ALSA: add private_data to struct snd_jack
ALSA: snd-usb-caiaq: rename files to remove redundant information in file pathes
ALSA: snd-usb-caiaq: clean up header includes
ALSA: sound/pci: use memdup_user()
ALSA: sound/usb: use memdup_user()
ALSA: sound/isa: use memdup_user()
ALSA: sound/core: use memdup_user()
[ALSA] intel8x0: do not use zero value from PICB register
[ALSA] intel8x0: an attempt to make ac97_clock measurement more reliable
[ALSA] pcm-midlevel: Add more strict buffer position checks based on jiffies
[ALSA] hda_intel: fix unexpected ring buffer positions
ASoC: Disable S3C64xx support in Kconfig
ASoC: magician: remove un-necessary #include of pxa-regs.h and hardware.h
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
* 'for-linus' of git://git.kernel.dk/linux-2.6-block: (28 commits)
cfq-iosched: add close cooperator code
cfq-iosched: log responsible 'cfqq' in idle timer arm
cfq-iosched: tweak kick logic a bit more
cfq-iosched: no need to save interrupts in cfq_kick_queue()
brd: fix cacheflushing
brd: support barriers
swap: Remove code handling bio_alloc failure with __GFP_WAIT
gfs2: Remove code handling bio_alloc failure with __GFP_WAIT
ext4: Remove code handling bio_alloc failure with __GFP_WAIT
dio: Remove code handling bio_alloc failure with __GFP_WAIT
block: Remove code handling bio_alloc failure with __GFP_WAIT
bio: add documentation to bio_alloc()
splice: add helpers for locking pipe inode
splice: remove generic_file_splice_write_nolock()
ocfs2: fix i_mutex locking in ocfs2_splice_to_file()
splice: fix i_mutex locking in generic_splice_write()
splice: remove i_mutex locking in splice_from_pipe()
splice: split up __splice_from_pipe()
block: fix SG_IO to return a proper error value
cfq-iosched: don't delay queue kick for a merged request
...
Fix the key value generation for get/set amp verbs. The upper bits of
the parameter have to be combined with the verb value to be unique for
each direction/index of amp access.
This fixes the resume problem on some hardwares like Macbook after
the channel mode is changed.
Tested-by: Johannes Berg <johannes@sipsolutions.net> Cc: <stable@kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
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: pseries/dtl.c should include asm/firmware.h
powerpc: Fix data-corrupting bug in __futex_atomic_op
powerpc/pseries: Set error_state to pci_channel_io_normal in eeh_report_reset()
powerpc: Allow 256kB pages with SHMEM
powerpc: Document new FSL I2C bindings and cleanup
powerpc/mm: Fix compile warning
powerpc/85xx: TQM8548: update defconfig
powerpc/85xx: TQM8548: use proper phy-handles for enet2 and enet3
powerpc/85xx: TQM85xx: correct address of LM75 I2C device nodes
powerpc: Add support for early tlbilx opcode
powerpc: Fix tlbilx opcode
It turns out that 'smp_call_function_many()' doesn't work at all like
'smp_call_function_single()', and my change to Andrew's patch to use it
rather than a loop over all CPU's acpi-cpufreq doesn't work.
My bad.
'smp_call_function_many()' has two "features" (aka "documented bugs"):
(a) it needs to be called with preemption disabled, because it uses
smp_processor_id() without guarding the CPU lookup with 'get_cpu()'
and 'put_cpu()' like the 'single' variant does.
(b) even if the current CPU is part of the CPU mask, it won't do the
call on that CPU.
Still, we're better off trying to use 'smp_call_function_many()' than
looping over CPU's, since it at least in theory allows us to use a
broadcast IPI and do it all in parallel. So let's just work around the
silly semantic bugs in that function.
Reported-and-tested-by: Ali Gholami Rudi <ali@rudi.ir> Cc: Ingo Molnar <mingo@elte.hu> Cc: Andrew Morton <akpm@linux-foundation.org>, Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Dave Jones <davej@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If we have processes that are working in close proximity to each
other on disk, we don't want to idle wait. Instead allow the close
process to issue a request, getting better aggregate bandwidth.
The anticipatory scheduler has similar checks, noop and deadline do
not need it since they don't care about process <-> io mappings.
The code for CFQ is a little more involved though, since we split
request queues into per-process contexts.
This fixes a performance problem with eg dump(8), since it uses
several processes in some silly attempt to speed IO up. Even if
dump(8) isn't really a valid case (it should be fixed by using
CLONE_IO), there are other cases where we see close processes
and where idling ends up hurting performance.
Credit goes to Jeff Moyer <jmoyer@redhat.com> for writing the
initial implementation.
We only kick the dispatch for an idling queue, if we think it's a
(somewhat) fully merged request. Also allow a kick if we have other
busy queues in the system, since we don't want to risk waiting for
a potential merge in that case. It's better to get some work done and
proceed.
Nick Piggin [Wed, 15 Apr 2009 08:32:07 +0000 (10:32 +0200)]
brd: fix cacheflushing
brd is missing a flush_dcache_page. On 2nd thoughts, perhaps it is the
pagecache's responsibility to flush user virtual aliases (the driver of
course should flush kernel virtual mappings)... but anyway, there
already exists cache flushing for one direction of transfer, so we
should add the other.
Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Nick Piggin [Wed, 15 Apr 2009 08:27:07 +0000 (10:27 +0200)]
brd: support barriers
brd is always ordered (not that it matters, as it is defined not to
survive when the system goes down). So tell the block layer it is
ordered, which might be of help with testing filesystems.
Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
There are lots of sequences like this, especially in splice code:
if (pipe->inode)
mutex_lock(&pipe->inode->i_mutex);
/* do something */
if (pipe->inode)
mutex_unlock(&pipe->inode->i_mutex);
so introduce helpers which do the conditional locking and unlocking.
Also replace the inode_double_lock() call with a pipe_double_lock()
helper to avoid spreading the use of this functionality beyond the
pipe code.
This patch is just a cleanup, and should cause no behavioral changes.
Remove the now unused generic_file_splice_write_nolock() function.
It's conceptually broken anyway, because splice may need to wait for
pipe events so holding locks across the whole operation is wrong.
ocfs2: fix i_mutex locking in ocfs2_splice_to_file()
Rearrange locking of i_mutex on destination and call to
ocfs2_rw_lock() so locks are only held while buffers are copied with
the pipe_to_file() actor, and not while waiting for more data on the
pipe.
splice: fix i_mutex locking in generic_splice_write()
Rearrange locking of i_mutex on destination so it's only held while
buffers are copied with the pipe_to_file() actor, and not while
waiting for more data on the pipe.
splice_from_pipe_next() will wait (if necessary) for more buffers to
be added to the pipe. splice_from_pipe_feed() will feed the buffers
to the supplied actor and return when there's no more data available
(or if all of the requested data has been copied).
This is necessary so that implementations can do locking around the
non-waiting splice_from_pipe_feed().
This patch should not cause any change in behavior.
blk_rq_unmap_user() returns -EFAULT if a program passes an invalid
address to kernel. SG_IO path needs to pass the returned value to user
space instead of ignoring it.
Markus Brunner [Wed, 15 Apr 2009 09:35:40 +0000 (02:35 -0700)]
gianfar: stop send queue before resetting gianfar
After a transmit timed out, the reset task will be called, which will free the
allocated resources(stop_gfar). If gfar_poll will be called before the
resources get allocated again gfar_clean_tx_ring will call
dev_kfree_skb_any(NULL).
* topic/memdup_user:
ALSA: sound/pci: use memdup_user()
ALSA: sound/usb: use memdup_user()
ALSA: sound/isa: use memdup_user()
ALSA: sound/core: use memdup_user()
Merge branch 'master' of git://git.alsa-project.org/alsa-kernel into for-linus
* 'master' of git://git.alsa-project.org/alsa-kernel:
[ALSA] intel8x0: add one retry to the ac97_clock measurement routine
[ALSA] intel8x0: fix wrong conditions in ac97_clock measure routine
[ALSA] intel8x0: do not use zero value from PICB register
[ALSA] intel8x0: an attempt to make ac97_clock measurement more reliable
[ALSA] pcm-midlevel: Add more strict buffer position checks based on jiffies
[ALSA] hda_intel: fix unexpected ring buffer positions
In certain cases symlinks can appear to have zero size if a lookup
on the inode occurs within a certain (very short) time after the
symlink has been created. The symlink is correctly created on disk
but appears to have zero size when stat()ed. This patch closes the
race and prevents incorrect sizes appearing.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
cfq-iosched: don't delay queue kick for a merged request
"Zhang, Yanmin" <yanmin_zhang@linux.intel.com> reports that commit b029195dda0129b427c6e579a3bb3ae752da3a93 introduced a regression
of about 50% with sequential threaded read workloads. The test
case is:
tiotest -k0 -k1 -k3 -f 80 -t 32
which starts 32 threads each reading a 80MB file. Twiddle the kick
queue logic so that we do start IO immediately, if it appears to be
a fully merged request. We can't really detect that, so just check
if the request is bigger than a page or not. The assumption is that
since single bio issues will first queue a single request with just
one page attached and then later do merges on that, if we already
have more than a page worth of data in the request, then the request
is most likely good to go.
Verified that this doesn't cause a regression with the test case that
commit b029195dda0129b427c6e579a3bb3ae752da3a93 was fixing. It does not,
we still see maximum sized requests for the queue-then-merge cases.
buffer: switch do_emergency_thaw() away from pdflush_operation()
This is (again) a preparatory patch similar to commit a2a9537ac0b37a5da6fbe7e1e9cb06c524d2a9c4. It open codes a simple
async way of executing do_thaw_all() out of context, so we can get
rid of pdflush.
Sachin Sant [Tue, 14 Apr 2009 14:35:55 +0000 (14:35 +0000)]
powerpc: pseries/dtl.c should include asm/firmware.h
A randconfig build on powerpc failed with:
dtl.c: In function 'dtl_init':
dtl.c:238: error: implicit declaration of function 'firmware_has_feature'
dtl.c:238: error: 'FW_FEATURE_SPLPAR' undeclared (first use in this function)
- We need firmware.h for these definitions.
Signed-off-by: Sachin Sant <sachinp@in.ibm.com> Signed-off-by: Jeremy Kerr <jk@ozlabs.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
Paul Mackerras [Mon, 13 Apr 2009 14:09:09 +0000 (14:09 +0000)]
powerpc: Fix data-corrupting bug in __futex_atomic_op
Richard Henderson pointed out that the powerpc __futex_atomic_op has a
bug: it will write the wrong value if the stwcx. fails and it has to
retry the lwarx/stwcx. loop, since 'oparg' will have been overwritten
by the result from the first time around the loop. This happens
because it uses the same register for 'oparg' (an input) as it uses
for the result.
This fixes it by using separate registers for 'oparg' and 'ret'.
Cc: stable@kernel.org Signed-off-by: Paul Mackerras <paulus@samba.org>
Mike Mason [Fri, 10 Apr 2009 08:57:03 +0000 (08:57 +0000)]
powerpc/pseries: Set error_state to pci_channel_io_normal in eeh_report_reset()
While adding native EEH support to Emulex and Qlogic drivers, it was
discovered that dev->error_state was set to pci_io_channel_normal too
late in the recovery process. These drivers rely on error_state to
determine if they can access the device in their slot_reset callback,
thus error_state needs to be set to pci_io_channel_normal in
eeh_report_reset(). Below is a detailed explanation (courtesy of Richard
Lary) as to why this is necessary.
Background:
PCI MMIO or DMA accesses to a frozen slot generate additional EEH
errors. If the number of additional EEH errors exceeds EEH_MAX_FAILS the
adapter will be shutdown. To avoid triggering excessive EEH errors and
an undesirable adapter shutdown, some drivers use the
pci_channel_offline(dev) wrapper function to return a Boolean value
based on the value of pci_dev->error_state to determine if PCI MMIO or
DMA accesses are safe. If the wrapper returns TRUE, drivers must not
make PCI MMIO or DMA access to their hardware.
The pci_dev structure member error_state reflects one of three values,
1) pci_channel_io_normal, 2) pci_channel_io_frozen, 3)
pci_channel_io_perm_failure. Function pci_channel_offline(dev) returns
TRUE if error_state is pci_channel_io_frozen or pci_channel_io_perm_failure.
The EEH driver sets pci_dev->error_state to pci_channel_io_frozen at the
point where the PCI slot is frozen. Currently, the EEH driver restores
dev->error_state to pci_channel_io_normal in eeh_report_resume() before
calling the driver's resume callback. However, when the EEH driver calls
the driver's slot_reset callback() from eeh_report_reset(), it
incorrectly indicates the error state is still pci_channel_io_frozen.
Waiting until eeh_report_resume() to restore dev->error_state to
pci_channel_io_normal is too late for Emulex and QLogic FC drivers and
any other drivers which are designed to use common code paths in these
two cases: i) those called after the driver's slot_reset callback() and
ii) those called after the PCI slot is frozen but before the driver's
slot_reset callback is called. Case i) all driver paths executed to
reinitialize the hardware after a reset and case ii) all code paths
executed by driver kernel threads that run asynchronous to the main
driver thread, such as interrupt handlers and worker threads to process
driver work queues.
Emulex and QLogic FC drivers are designed with common code paths which
require that pci_channel_offline(dev) reflect the true state of the
hardware. The state transitions that the hardware takes from Normal
Operations to Slot Frozen to Reset to Normal Operations are documented
in the Power Architecture™ Platform Requirements+ (PAPR+) in Table 75.
PE State Control.
PAPR defines the following 3 states:
0 -- Not reset, Not EEH stopped, MMIO load/store allowed, DMA allowed
(Normal Operations)
1 -- Reset, Not EEH stopped, MMIO load/store disabled, DMA disabled
2 -- Not reset, EEH stopped, MMIO load/store disabled, DMA disabled
(Slot Frozen)
An EEH error places the slot in state 2 (Frozen) and the adapter driver
is notified that an EEH error was detected. If the adapter driver
returns PCI_ERS_RESULT_NEED_RESET, the EEH driver calls
eeh_reset_device() to place the slot into state 1 (Reset) and
eeh_reset_device completes by placing the slot into State 0 (Normal
Operations). Upon return from eeh_reset_device(), the EEH driver calls
eeh_report_reset, which then calls the adapter's slot_reset callback. At
the time the adapter's slot_reset callback is called, the true state of
the hardware is Normal Operations and should be accurately reflected by
setting dev->error_state to pci_channel_io_normal.
The current implementation of EEH driver does not do so and requires
this change to correct this deficiency.
Signed-off-by: Mike Mason <mmlnx@us.ibm.com> Acked-by: Linas Vepstas <linasvepstas@gmail.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
ixgbe: update real_num_tx_queues on changing num_rx_queues
Move the update of real_num_tx_queues from
ixgbe_acquire_msix_vectors() to ixgbe_set_num_queues(), to ensure it
be always in sync with num_tx_queues.
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Don't do the num_tx_queues based masking on calculating tx queue
index.
1) num_tx_queues is not always power-of-2, because it also depends on
the online cpu numbers. So the masking could be a performance bug
on a 6 cpu system.
2) queue_mapping will be limited by real_num_tx_queues=num_tx_queues
in the generic netdev function set_cur_queue_map(). So the bound
limiting here is not necessary.
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ben Hutchings [Wed, 15 Apr 2009 02:48:34 +0000 (19:48 -0700)]
sfc: Use correct macro to set event bitfield
falcon_sim_phy_event() used EFX_OWORD_FIELD, which operates on
bitfields in 128-bit values, on an event, which is a 64-bit value.
This should be harmless - these macros always use little-endian
ordering, so it would read and write back the following 8 bytes
unchanged - but it is obviously wrong.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ben Hutchings [Wed, 15 Apr 2009 02:47:46 +0000 (19:47 -0700)]
sfc: Match calls to netif_napi_add() and netif_napi_del()
sfc could call netif_napi_add() multiple times for the same
napi_struct, corrupting the list of napi_structs for the associated
device and leading to a busy-loop on device removal. Move the call to
netif_napi_add() and add a call to netif_napi_del() in the obvious
places.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
e1000/e1000e compile report a possible unused variable, fix
that for now. Shortly after this a small refactor and bug
fix will follow in the same code.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Reported-by: Subrata Modak <subrata@linux.vnet.ibm.com> Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michal Schmidt [Tue, 14 Apr 2009 22:16:55 +0000 (15:16 -0700)]
skge: fix occasional BUG during MTU change
The BUG_ON(skge->tx_ring.to_use != skge->tx_ring.to_clean) in skge_up()
was sometimes observed when setting MTU.
skge_down() disables the TX queue, but then reenables it by mistake via
skge_tx_clean().
Fix it by moving the waking of the queue from skge_tx_clean() to the
other caller. And to make sure start_xmit is not in progress on another
CPU, skge_down() should call netif_tx_disable().
The bug was reported to me by Jiri Jilek whose Debian system sometimes
failed to boot. He tested the patch and the bug did not happen anymore.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com> Acked-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Tue, 14 Apr 2009 22:11:06 +0000 (15:11 -0700)]
gro: Restore correct value to gso_size
Since everybody has been focusing on baremetal GRO performance
no one noticed when I added a bug that zapped gso_size for all
GRO packets. This only gets picked up when you forward the skb
out of an interface.
Thanks to Mark Wagner for noticing this bug when testing kvm.
Reported-by: Mark Wagner <mwagner@redhat.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-beck.git / log