07 Nov, 2007
6 commits
-
Normally io priorities follow the CPU nice, unless a specific scheduling
class has been set. Once that is set, there's no way to reset the
behaviour to 'none' so that it follows CPU nice again.Currently passing in 0 as the ioprio class/value will return -1/EINVAL,
change that to allow resetting of a set scheduling class.Signed-off-by: Jens Axboe
-
In theory, if the queue was idle long enough, cfq_idle_class_timer may have
a false (and very long) timeout because jiffies can wrap into the past wrt
->last_end_request.Signed-off-by: Oleg Nesterov
Signed-off-by: Jens Axboe -
After the fresh boot:
ionice -c3 -p $$
echo cfq >> /sys/block/XXX/queue/scheduler
dd if=/dev/XXX of=/dev/null bs=512 count=1Now dd hangs in D state and the queue is completely stalled for approximately
INITIAL_JIFFIES + CFQ_IDLE_GRACE jiffies. This is because cfq_init_queue()
forgets to initialize cfq_data->last_end_request.(I guess this patch is not complete, overflow is still possible)
Signed-off-by: Oleg Nesterov
Signed-off-by: Jens Axboe -
Spotted by Nick , hopefully can explain the second trace in
http://bugzilla.kernel.org/show_bug.cgi?id=9180.If ->async_idle_cfqq != NULL cfq_put_async_queues() puts it IOPRIO_BE_NR times
in a loop. Fix this.Signed-off-by: Oleg Nesterov
Signed-off-by: Jens Axboe -
Reported by Al Viro.
This fixes it:
[AC]FLAGS -> KBUILD_[AC]FLAGS conversion in Makefile-i386.
Signed-off-by: Jeff Dike
Cc: Al Viro
Cc: Sam Ravnborg
Signed-off-by: Linus Torvalds
06 Nov, 2007
34 commits
-
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
9p: add missing end-of-options record for trans_fd
9p: return NULL when trans not found
9p: use copy of the options value instead of original
9p: fix memory leak in v9fs_get_sb -
The list of options that the fd transport accepts is missing end-of-options
marker. This patch adds it.Signed-off-by: Latchesar Ionkov
Acked-by: Eric Van Hensbergen -
v9fs_match_trans function returns arbitrary transport module instead of NULL
when the requested transport is not registered. This patch modifies the
function to return NULL in that case.Signed-off-by: Latchesar Ionkov
Acked-by: Eric Van Hensbergen -
v9fs_parse_options function uses strsep which modifies the value of the
v9ses->options field. That modified value is later passed to the function
that creates the transport potentially making the transport creation
function to fail.This patch creates a copy of v9ses->option field that v9fs_parse_options
function uses instead of the original value.Signed-off-by: Latchesar Ionkov
Acked-by: Eric Van Hensbergen -
This patch fixes a memory leak in v9fs_get_sb.
Signed-off-by: Latchesar Ionkov
Acked-by: Eric Van Hensbergen -
* 'drm-patches' of master.kernel.org:/pub/scm/linux/kernel/git/airlied/drm-2.6:
drm: DRM: fix memset size error
drm: remove remnants of DRM_COPY_FROM/TO_USER_IOCTL
drm: remove second forward decleration of drm device struct. -
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
libata: handle broken cable reporting
pata_hpt37x: Fix outstanding bug reports on the HPT374 and 37x cable detect
ata_piix: Add additional PCI identifier for 40 wire short cable
pata_serverworks: Fix problem with some drive combinations
libata: Don't disable dipm with SET FEATURES
libata and bogus LBA48 drives -
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
phylib: Silence driver registration
phylib: Add ID for Marvell 88E1240
82596: free nonexistent resource fix
SUNHME: Fix missing NETIF_F_VLAN_CHALLENGED on PCI happy meals -
The size passing to memset is wrong.
Signed-off-by: Li Zefan
Signed-off-by: Dave Airlie -
This is a bug in the savage driver since I introduced these changes.
Signed-off-by: Dave Airlie
-
Signed-off-by: Dave Airlie
-
Keeping the list in sync with the old IDE driver
Signed-off-by: Alan Cox
Signed-off-by: Linus Torvalds -
Signed-off-by: Li Zefan
Cc: Thomas Gleixner
Cc: john stultz
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix system call defines for system call 180 and 181 to match the underlying
system call table function entries. System call 180 calls sys_pread64, and
181 calls sys_pwrite64, so make the definitions match.Signed-off-by: Greg Ungerer
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Release the crypt_stat hash mutex on allocation error. Check for error
conditions when doing crypto hash calls.Signed-off-by: Michael Halcrow
Reported-by: Kazuki Ohta
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The extent_offset is getting incremented twice per loop iteration through any
given page. It should only be getting incremented once. This bug should only
impact hosts with >4K page sizes.Signed-off-by: Michael Halcrow
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Remove panic from phonedev. See
http://bugzilla.kernel.org/show_bug.cgi?id=9266
for details (phonedev panics if unregistering device not registered).
Signed-off-by: Matti Linnanvuori
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
commit 4ae3f847e49e3787eca91bced31f8fd328d50496 ("md: raid5: fix
clearing of biofill operations") did not get applied correctly,
presumably due to substantial similarities between handle_stripe5 and
handle_stripe6.This patch moves the chunk of new code from handle_stripe6 (where it isn't
needed (yet)) to handle_stripe5.Signed-off-by: Neil Brown
Cc: "Dan Williams"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Make virtcons_probe() __devinit.
Fixes this section warning:WARNING: vmlinux.o(.text+0x14c10b): Section mismatch: reference to .init.text:hvc_alloc (between 'virtcons_probe' and 'ac_register_board')
Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Change the name of this data to use a name (suffix) that is whitelisted
by MODPOST so that the section warning is fixed (not generated).WARNING: vmlinux.o(.data+0x1b140): Section mismatch: reference to .init.text:m48t59_rtc_probe (between 'm48t59_rtc_platdrv' and 'm48t59_nvram_attr')
Signed-off-by: Randy Dunlap
Acked-by: Alessandro Zummo
Cc: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
KERNEL_DEFINES needs whitespace trimmed, otherwise the whitespace crunching
done by make fools the patsubst which is used to remove KERNEL_DEFINES from
USER_CFLAGS.Signed-off-by: Jeff Dike
Cc: Sam Ravnborg
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix an incompatible-pointer warning.
Signed-off-by: WANG Cong
Signed-off-by: Jeff Dike
Cc: Jens Axboe
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix 'and' typo (PT_WRITE_OK is defined 2)
Signed-off-by: Roel Kluin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch fixes an off-by-one error spotted by the Coverity checker.
Signed-off-by: Adrian Bunk
Acked-by: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
A tiny vestige of arm26 has appeared: remove it again.
(akpm: someone (tm) needs to remove include/asm-arm26/ too)
Signed-off-by: Hugh Dickins
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
When a bitmap is empty bitmap_scnlistprintf() would leave the buffer
uninitialized. Set it to an empty string in this case.I didn't see any in normal kernel callers hitting this, but some custom
debug code of mine did.Signed-off-by: Andi Kleen
Acked-by: Paul Jackson
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The Build with randconfig fails with following error with the
2.6.24-rc4-git9include/linux/kallsyms.h:56: error: `NULL' undeclared (first use in this
function)
include/linux/kallsyms.h:56: error: (Each undeclared identifier is
reported only once
include/linux/kallsyms.h:56: error: for each function it appears in.)
make[2]: *** [arch/powerpc/platforms/cell/spu_callbacks.o] Error 1
make[1]: *** [arch/powerpc/platforms/cell] Error 2
make: *** [arch/powerpc/platforms] Error 2Signed-off-by: Kamalesh Babulal
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Grant Grundler was asking for more detail about correct usage of local
atomic operations and suggested adding the resulting summary to
local_ops.txt."Please add a bit more detail. If DaveM is correct (he normally is), then
there must be limits on how the local_t can be used in the kernel process
and interrupt contexts. I'd like those rules spelled out very clearly
since it's easy to get wrong and tracking down such a bug is quite painful."Signed-off-by: Mathieu Desnoyers
Signed-off-by: Grant Grundler
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
drivers/serial/8250_early.c:80: warning: conflicting types for built-in function `putc'
Signed-off-by: Yinghai Lu
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
max_phys_segments and max_sectors were swapped
Signed-off-by: Vasily Averin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Let's make immediately obvious from where sysctl comes from and messages
itself more noticeable.Signed-off-by: Alexey Dobriyan
Acked-by: "Eric W. Biederman"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Roel Kluin
Acked-by: Karsten Keil
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
One or two ancient drives predated the cable spec and didn't sent the
valid bits for the field. I had hoped to leave this out of libata as a
piece of historical annoyance but a recent CD drive shows the same bug so
we have to import support for it.Same concept as Bartlomiej's changes old IDE except that as we have
centralised blacklists we can avoid keeping another private table of stuffSigned-off-by: Alan Cox
Signed-off-by: Jeff Garzik -
- Read frequency correctly
- Correct cable detect handling
- Fix wrong filter testSigned-off-by: Alan Cox
Signed-off-by: Jeff Garzik