08 Dec, 2006
40 commits
-
It is possible to have tasklets get scheduled before softirqd has had a chance
to spawn on all CPUs. This is totally harmless; after success during action
CPU_UP_PREPARE, action CPU_ONLINE will be called, which immediately wakes
softirqd on the appropriate CPU to process the already pending tasklets. So
there is no danger of having a missed wakeup for any tasklets that were
already pending.In particular, i386 is affected by this during startup, and is visible when
using a very large initrd; during the time it takes for the initrd to be
decompressed, a timer IRQ can come in and schedule RCU callbacks. It is also
possible that resending of a hardware IRQ via a softirq triggers the same bug.Because of different timing conditions, this shows up in all emulators and
virtual machines tested, including Xen, VMware, Virtual PC, and Qemu. It is
also possible to trigger on native hardware with a large enough initrd,
although I don't have a reliable case demonstrating that.Signed-off-by: Zachary Amsden
Cc:
Cc: Ingo Molnar
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
When kernel is compiled with old version of autofs (CONFIG_AUTOFS_FS), and
new (observed at least with 5.x.x) automount deamon is started, kernel
correctly reports incompatible version of kernel and userland daemon, but
then screws things up instead of correct handling of the error:autofs: kernel does not match daemon version
=====================================
[ BUG: bad unlock balance detected! ]
-------------------------------------
automount/4199 is trying to release lock (&type->s_umount_key) at:
[] get_sb_nodev+0x76/0xa4
but there are no more locks to release!other info that might help us debug this:
no locks held by automount/4199.stack backtrace:
[] dump_trace+0x68/0x1b2
[] show_trace_log_lvl+0x18/0x2c
[] show_trace+0xf/0x11
[] dump_stack+0x12/0x14
[] print_unlock_inbalance_bug+0xe7/0xf3
[] lock_release+0x8d/0x164
[] up_write+0x14/0x27
[] get_sb_nodev+0x76/0xa4
[] vfs_kern_mount+0x83/0xf6
[] do_kern_mount+0x2d/0x3e
[] do_mount+0x607/0x67a
[] sys_mount+0x72/0xa4
[] sysenter_past_esp+0x5f/0x99
DWARF2 unwinder stuck at sysenter_past_esp+0x5f/0x99
Leftover inexact backtrace:
=======================and then deadlock comes.
The problem: autofs_fill_super() returns EINVAL to get_sb_nodev(), but
before that, it calls kill_anon_super() to destroy the superblock which
won't be needed. This is however way too soon to call kill_anon_super(),
because get_sb_nodev() has to perform its own cleanup of the superblock
first (deactivate_super(), etc.). The correct time to call
kill_anon_super() is in the autofs_kill_sb() callback, which is called by
deactivate_super() at proper time, when the superblock is ready to be
killed.I can see the same faulty codepath also in autofs4. This patch solves
issues in both filesystems in a same way - it postpones the
kill_anon_super() until the proper time is signalized by deactivate_super()
calling the kill_sb() callback.[raven@themaw.net: update comment]
Signed-off-by: Jiri Kosina
Acked-by: Ian Kent
Cc:
Signed-off-by: Ian Kent
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The i2c and hwmon trees have moved to a new location.
The lm-sensors project moved to a new home as well.
Signed-off-by: Jean Delvare
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add suspend/resume methods to drivers/serial/8250_pnp.c. Tested on a
P4/HT 16550A box, ttyS0 login survives across suspend to ram.[akpm@osdl.org: cleanups]
Signed-off-by: Mike Galbraith
Cc: "Rafael J. Wysocki"
Cc: Pavel Machek
Cc: Russell King
Cc: Adam Belay
Cc: Bjorn Helgaas
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Some people want to use ide_cd for CD-ROM but still dynamically load
ide-scsi for things like tape drives. If you compile in the CD driver this
works out but if you want them modular you need an option to ensure that
whoever loads first the right things happen.This replaces the original draft patch which leaked a scsi host reference
[akpm@osdl.org: add MODULE_PARM_DESC]
Signed-off-by: Alan Cox
Cc: Bartlomiej Zolnierkiewicz
Cc: Jens Axboe
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Make the locking self-test failures (of 'FAILURE' type) easier to debug by
printing more information.Signed-off-by: Ingo Molnar
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Some have reported a chain-table overflow - double its size.
Signed-off-by: Ingo Molnar
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
CONFIG_W1_SLAVE_DS2433_CRC can be used directly, there's no reason for the
indirection of defining a different variable in the Makefile.Signed-off-by: Adrian Bunk
Signed-off-by: Evgeniy Polyakov
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Port fix to the off-by-one in find_next_usable_block's memscan from ext2 to
ext4; but it didn't cause a serious problem for ext4 because the additional
ext4_test_allocatable check rescued it from the error.[akpm@osdl.org: build fix]
Signed-off-by: Mingming Cao
Signed-off-by: Hugh Dickins
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
ext4_new_blocks has a nice io_error label for setting -EIO, so goto that in
the one place that doesn't already use it.Signed-off-by: Mingming Cao
Signed-off-by: Hugh Dickins
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The reservations tree is an rb_tree not a list, so it's less confusing to use
rb_entry() than list_entry() - though they're both just container_of().Signed-off-by: Mingming Cao
Signed-off-by: Hugh Dickins
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
rsv_end is the last block within the reservation, so alloc_new_reservation
should accept start_block == rsv_end as success.Signed-off-by: Mingming Cao
Signed-off-by: Hugh Dickins
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
grp_goal 0 is a genuine goal (unlike -1), so ext4_try_to_allocate_with_rsv
should treat it as such.Signed-off-by: Mingming Cao
Signed-off-by: Hugh Dickins
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
ext4_new_blocks should reset the reservation window size to 0 when squeezing
the last blocks out of an almost full filesystem, so the retry doesn't skip
any groups with less than half that free, reporting ENOSPC too soon.Signed-off-by: Mingming Cao
Signed-off-by: Hugh Dickins
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
In the current jbd code, if a buffer on BJ_SyncData list is dirty and not
locked, the buffer is refiled to BJ_Locked list, submitted to the IO and
waited for IO completion.But the fsstress test showed the case that when a buffer was already
submitted to the IO just before the buffer_dirty(bh) check, the buffer was
not waited for IO completion.Following patch solves this problem. If it is assumed that a buffer is
submitted to the IO before the buffer_dirty(bh) check and still being
written to disk, this buffer is refiled to BJ_Locked list.Signed-off-by: Hisashi Hifumi
Cc: Jan Kara
Cc: "Stephen C. Tweedie"
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Always build hweight8/16/32/64() functions into the kernel so that loadable
modules may use them.I didn't remove GENERIC_HWEIGHT since ALPHA_EV67, ia64, and some variants
of UltraSparc(64) provide their own hweight functions.Fixes config/build problems with NTFS=m and JOYSTICK_ANALOG=m.
Kernel: arch/x86_64/boot/bzImage is ready (#19)
Building modules, stage 2.
MODPOST 94 modules
WARNING: "hweight32" [fs/ntfs/ntfs.ko] undefined!
WARNING: "hweight16" [drivers/input/joystick/analog.ko] undefined!
WARNING: "hweight8" [drivers/input/joystick/analog.ko] undefined!
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Remove the carta_random32.h header file. The carta_random32() function was
was put in and removed in favor of random32(). In the removal process, the
header file was forgotten.Signed-off-by: Stephane Eranian
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add kernel .config file to REPORTING-BUGS.
Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
According to the datasheet rs5c372 supports three different methods for
reading register values. Change from method #1 to method #3, since method #3
is the only one that works on Thecus N2100 board with this RTC.Signed-off-by: Riku Voipio
Cc: Alessandro Zummo
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We add a save link for O_DIRECT writes to protect the i_size against the
crashes before we actually finish the I/O. If we hit an -ENOSPC in
aops->prepare_write(), we would do a truncate() to release the blocks which
might have got initialized. Now the truncate would add another save link
for the same inode causing a reiserfs panic for having multiple save links
for the same inode.Signed-off-by: Vladimir V. Saveliev
Signed-off-by: Amit Arora
Signed-off-by: Suzuki K P
Cc: Jeff Mahoney
Cc: Chris Mason
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
At present show_state prints a header the does not match the output of
show_task, as follows:-
sibling
task PC pid father child younger older
init S 00000000 0 1 0 2 (NOTLB)
-This patch corrects the output of show_state so that the header is
aligned with the data, ala:-
free sibling
task PC stack pid father child younger older
init S 00000000 0 1 0 2 (NOTLB)
-Signed-off-by: Chris Caputo
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
In the functions do_proc_dointvec() and do_proc_doulongvec_minmax(),
there seems to be a bug in string length calculation if string contains
negative integer.The console log given below explains the bug. Setting negative values
may not be a right thing to do for "console log level" but then the test
(given below) can be used to demonstrate the bug in the code.# echo "-1 -1 -1 -123456" > /proc/sys/kernel/printk
# cat /proc/sys/kernel/printk
-1 -1 -1 -1234
#
# echo "-1 -1 -1 123456" > /proc/sys/kernel/printk
# cat /proc/sys/kernel/printk
-1 -1 -1 1234
#(akpm: the bug is that 123456 gets truncated)
It works as expected if string contains all +ve integers
# echo "1 2 3 4" > /proc/sys/kernel/printk
# cat /proc/sys/kernel/printk
1 2 3 4
#The patch given below fixes the issue.
Signed-off-by: Praveen BP
Cc: "Eric W. Biederman"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Initialization synclink_gt forgot to unregister pci driver on error path.
Signed-off-by: Akinobu Mita
Cc: Paul Fulghum
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Check the return value of platform_device_register_simple().
Cc: David Brownell
Signed-off-by: Akinobu Mita
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Check register_filesystem() and kern_mount() return values.
Cc: Ingo Molnar
Signed-off-by: Akinobu Mita
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Replace kmalloc+memset with kzalloc
Signed-off-by: Yan Burman
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This a set of fixes mostly to make the driver actually work:
1. Actually select the line for setting parameters and receiver
disable/enable.
2. Select the line for receive and transmit interrupt handling correctly.
3. Report the transmitter empty state correctly.
4. Set the I/O type of ports correctly.
5. Perform polled transmission correctly.
6. Don't fix the console line at ttyS3.
7. Magic SysRq support.
8. Various small bits here and there.Tested with a DECstation 2100 (thanks Flo for making this possible).
[akpm@osdl.org: fix typo]
Signed-off-by: Maciej W. Rozycki
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Provide a common interface for all the subsystems to lock and unlock their
per-subsystem hotcpu mutexes.When CONFIG_HOTPLUG_CPU is not set, these operations would be no-ops.
[akpm@osdl.org: macros -> inlines]
Signed-off-by: Gautham R Shenoy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Pass struct dev pointer to dma_cache_sync()
dma_cache_sync() is ill-designed in that it does not have a struct device
pointer argument which makes proper support for systems that consist of a
mix of coherent and non-coherent DMA devices hard. Change dma_cache_sync
to take a struct device pointer as first argument and fix all its callers
to pass it.Signed-off-by: Ralf Baechle
Cc: James Bottomley
Cc: "David S. Miller"
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
dma_is_consistent() is ill-designed in that it does not have a struct
device pointer argument which makes proper support for systems that consist
of a mix of coherent and non-coherent DMA devices hard. Change
dma_is_consistent to take a struct device pointer as first argument and fix
the sole caller to pass it.Signed-off-by: Ralf Baechle
Cc: James Bottomley
Cc: "David S. Miller"
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
On 32bits SMP platforms, 64bits i_size is protected by a seqcount
(i_size_seqcount).When i_size is read or written, i_size_seqcount is read/written as well, so
it make sense to group these two fields together in the same cache line.This patch moves i_size_seqcount next to i_size, and also moves i_version
to let offsetof(struct inode, i_size) being 0x40 instead of 0x3c (for
32bits platforms).For 64 bits platforms, i_size_seqcount doesnt exist, and the move of a
'long i_version' should not introduce a new hole because of padding.Signed-off-by: Eric Dumazet
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Be more careful about function pointer args:
look for "(...*" instead of just "(".This line in include/linux/input.h fools the current kernel-doc script
into deciding that this is a function pointer:unsigned long ffbit[NBITS(FF_MAX)];
Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We currently insert socket dentries into the global dentry hashtable. This
is suboptimal because there is currently no way these entries can be used
for a lookup(). (/proc/xxx/fd/xxx uses a different mechanism). Inserting
them in dentry hashtable slows dcache lookups.To let __dpath() still work correctly (ie not adding a " (deleted)") after
dentry name, we do :- Right after d_alloc(), pretend they are hashed by clearing the
DCACHE_UNHASHED bit.- Call d_instantiate() instead of d_add() : dentry is not inserted in
hash table.__dpath() & friends work as intended during dentry lifetime.
- At dismantle time, once dput() must clear the dentry, setting again
DCACHE_UNHASHED bit inside the custom d_delete() function provided by
socket code, so that dput() can just kill_it.Signed-off-by: Eric Dumazet
Cc: Al Viro
Acked-by: "David S. Miller"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Some dentries don't need to be globally visible in dentry hashtable.
(pipes & sockets)Such dentries dont need to wait for a RCU grace period at delete time.
Being able to free them permits a better CPU cache use (hot cache)This patch combined with (dont insert pipe dentries into dentry_hashtable)
reduced time of { pipe(p); close(p[0]); close(p[1]);} on my UP machine (1.6
GHz Pentium-M) from 3.23 us to 2.86 us (But this patch does not depend on
other patches, only bench results)Signed-off-by: Eric Dumazet
Cc: Al Viro
Cc: Maneesh Soni
Cc: "Paul E. McKenney"
Cc: Dipankar Sarma
Acked-by: David Miller
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We currently insert pipe dentries into the global dentry hashtable. This
is suboptimal because there is currently no way these entries can be used
for a lookup(). (/proc/xxx/fd/xxx uses a different mechanism). Inserting
them in dentry hashtable slows dcache lookups.To let __dpath() still work correctly (ie not adding a " (deleted)") after
dentry name, we do :- Right after d_alloc(), pretend they are hashed by clearing the
DCACHE_UNHASHED bit.- Call d_instantiate() instead of d_add() : dentry is not inserted in
hash table.__dpath() & friends work as intended during dentry lifetime.
- At dismantle time, once dput() must clear the dentry, setting again
DCACHE_UNHASHED bit inside the custom d_delete() function provided by
pipe code, so that dput() can just kill_it.This patch, combined with (avoid RCU for never hashed dentries) reduced
time of { pipe(p); close(p[0]); close(p[1]);} on my UP machine (1.6GHz
Pentium-M) from 3.23 us to 2.86 us (But this patch does not depend on other
patches, only bench results)Signed-off-by: Eric Dumazet
Acked-by: David Miller
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
On some workloads, (for example when lot of close() syscalls are done), RCU
qlen can be quite large, and RCU heads are no longer in cpu cache when
rcu_do_batch() is called.This patch adds a prefetch() in rcu_do_batch() to give CPU a hint to bring
back cache lines containing 'struct rcu_head's.Most list manipulations macros include prefetch(), but not open coded ones
(at least with current C compilers :) )I got a nice speedup on a trivial benchmark (3.48 us per iteration instead
of 3.95 us on a 1.6 GHz Pentium-M)while (1) { pipe(p); close(fd[0]); close(fd[1]);}
Signed-off-by: Eric Dumazet
Cc: "Paul E. McKenney"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Remove the videodev chapter from the kernel-api book. It's done much better
in the videobook kernel-doc.Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Correct lots of typos, kernel-doc warnings, & kernel-doc usage in fusion and
i2o drivers.Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add Fusion and I2O message-based device interfaces to kernel-api book.
Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
