Sujith [Fri, 13 Mar 2009 03:26:11 +0000 (08:56 +0530)]
ath9k: Fix bug in handling single stream stations
AP mode currently sets up the dual stream capability
for all stations. This patch fixes it by checking if the
associated station supports dual stream MCS rates (8-15).
We would disregard any MCS rates above 15, since Atheros
HW supports only 0..15 rates currently, and can't receive
at rates > 15 anyway.
Signed-off-by: Sujith <Sujith.Manoharan@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Remove the PCI retry timeout code as that was just taken from ipw2100
due to historical reasons but in reality its a no-op, additionally its
simply incorrect as each PCI devices has its own custom PCI configuration
space on PCI config space >= 0x40. Not to mention we were trying to write
0 to a place that already has 0 on it.
Cc: Matthew Garrett <mjg59@srcf.ucam.org> Cc: Ben Cahill <ben.m.cahill@intel.com> Cc: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com> Tested-by: Adel Gadllah <adel.gadllah@gmail.com> Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jouni Malinen [Thu, 12 Mar 2009 19:53:23 +0000 (21:53 +0200)]
ath9k: Add support for multiple virtual AP interfaces
This patch fixes the TSF offset calculation for staggered Beacon frames
and sets ATH_BCBUF back to the earlier value 4 to enable multi-BSS
configurations of up to four BSSes.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Helmut Schaa [Thu, 12 Mar 2009 13:04:34 +0000 (14:04 +0100)]
mac80211: start pending scan after probe/auth/assoc timed out
If a scan is queued in STA mode while the interface is in state direct
probe, authenticate or associate the scan is delayed until the interface
enters disabled or associated state. But in case of direct probe-,
authentication- or association- timeout sta_work will not be scheduled
anymore (without external trigger) and thus the pending scan is not
executed and prevents a new scan from being triggered (-EBUSY).
Fix this by queueing the sta work again after direct probe-, authentication-
and association- timeout.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Thu, 12 Mar 2009 10:16:48 +0000 (11:16 +0100)]
mac80211: reduce max number of queues
No hw/driver actually supports more than four queues right now,
and we allocate a number of things per queue which means we
waste a bit of memory. Reduce the maximum number to four to
accurately reflect what we do (and need for QoS). Even if we
had hardware supporting more queues we couldn't take advantage
of that right now anyway.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reinette Chatre [Thu, 12 Mar 2009 16:20:40 +0000 (09:20 -0700)]
cfg80211/nl80211: remove usage of CONFIG_NL80211
The scan capability added to cfg80211/nl80211 introduced a
dependency on nl80211 by cfg80211. We can thus no longer have
just cfg80211 without nl80211. Specifically, cfg80211_scan_done()
calls nl80211_send_scan_aborted() or nl80211_send_scan_done().
Now we remove the option for user to select nl80211. It will always
be compiled if user selects cfg80211.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Don't call ieee80211_sta_find_ibss() directly, like it's done in STA
mode, so that the commit() call is more harmless respectively has
less site-effects.
Signed-off-by: Alina Friedrichsen <x-alina@gmx.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Grr-target: Eric Anholt <eric@anholt.net>
Grr-target: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Fri, 27 Mar 2009 23:50:49 +0000 (16:50 -0700)]
Merge branch 'drm-intel-next' of git://git.kernel.org/pub/scm/linux/kernel/git/anholt/drm-intel
* 'drm-intel-next' of git://git.kernel.org/pub/scm/linux/kernel/git/anholt/drm-intel: (25 commits)
drm/i915: Fix LVDS dither setting
drm/i915: Check for dev->primary->master before dereference.
drm/i915: TV detection fix
drm/i915: TV mode_set sync up with 2D driver
drm/i915: Fix TV get_modes to return modes count
drm/i915: Sync crt hotplug detection with intel video driver
drm/i915: Sync mode_valid/mode_set with intel video driver
drm/i915: TV modes' parameters sync up with 2D driver
agp/intel: Add support for new intel chipset.
i915/drm: Remove two redundant agp_chipset_flushes
drm/i915: Display fence register state in debugfs i915_gem_fence_regs node.
drm/i915: Add information on pinning and fencing to the i915 list debug.
drm/i915: Consolidate gem object list dumping
drm/i915: Convert i915 proc files to seq_file and move to debugfs.
drm: Convert proc files to seq_file and introduce debugfs
drm/i915: Fix lock order reversal in GEM relocation entry copying.
drm/i915: Fix lock order reversal with cliprects and cmdbuf in non-DRI2 paths.
drm/i915: Fix lock order reversal in shmem pread path.
drm/i915: Fix lock order reversal in shmem pwrite path.
drm/i915: Make GEM object's page lists refcounted instead of get/free.
...
Anton Vorontsov [Fri, 27 Mar 2009 23:00:03 +0000 (16:00 -0700)]
ucc_geth: Fix three oopses in PHY {de,}initialization code
When there are no free snums, UCC ethernet should gracefully fail, but
currently it oopses this way:
# ifconfig eth0 up
fill_init_enet_entries: Can not get SNUM.
ucc_geth_startup: Can not fill p_init_enet_param_shadow.
eth0: Cannot configure net device, aborting.
Unable to handle kernel paging request for data at address 0x00000190
Faulting instruction address: 0xc0294c88
Oops: Kernel access of bad area, sig: 11 [#1]
[...]
NIP [c0294c88] mutex_lock+0x0/0x1c
LR [c01b6be8] phy_stop+0x20/0x70
Call Trace:
[efb25da0] [efb2eb60] 0xefb2eb60 (unreliable)
[efb25db0] [c01b2058] ucc_geth_stop+0x2c/0x8c
[efb25dd0] [c01b4194] ucc_geth_open+0x48/0x27c
[efb25df0] [c020eec0] dev_open+0xc0/0x118
[...]
This is because the ucc_geth_stop() routine assumes that ugeth->phydev
is always initialized by the ucc_geth_open(), while it is not in case
of errors.
If we add a check to the ucc_geth_stop(), then another oops pops up:
Unable to handle kernel paging request for data at address 0x00000004
Faulting instruction address: 0xc01b46a4
Oops: Kernel access of bad area, sig: 11 [#1]
[...]
NIP [c01b46a4] adjust_link+0x20/0x1b4
LR [c01b770c] phy_state_machine+0xdc/0x44c
Call Trace:
[ef83bf10] [c021b388] linkwatch_schedule_work+0x74/0xf8 (unreliable)
[ef83bf40] [c01b770c] phy_state_machine+0xdc/0x44c
[ef83bf60] [c004c13c] run_workqueue+0xb8/0x148
[ef83bf90] [c004c870] worker_thread+0x70/0xd0
[ef83bfd0] [c00505fc] kthread+0x48/0x84
[ef83bff0] [c000f464] kernel_thread+0x4c/0x68
[...]
That one happens because ucc_geth_stop() does not call phy_disconnect()
and so phylib state machine is running without any idea that a MAC has
just died.
Also, when device tree specifies fixed-link, and CONFIG_FIXED_PHY
is disabled, we'll get this oops:
0:01 not found
eth2: Could not attach to PHY
eth2: Cannot initialize PHY, aborting.
Unable to handle kernel paging request for data at address 0x00000190
Faulting instruction address: 0xc02967d0
Oops: Kernel access of bad area, sig: 11 [#1]
[...]
NIP [c02967d0] mutex_lock+0x0/0x1c
LR [c01b6bcc] phy_stop+0x20/0x70
Call Trace:
[ef82be50] [efb6bb60] 0xefb6bb60 (unreliable)
[ef82be60] [c01b2058] ucc_geth_stop+0x2c/0x8c
[ef82be80] [c01b4194] ucc_geth_open+0x48/0x27c
[ef82bea0] [c0210a04] dev_open+0xc0/0x118
[ef82bec0] [c020f85c] dev_change_flags+0x84/0x1ac
[ef82bee0] [c037b768] ic_open_devs+0x168/0x2bc
[ef82bf20] [c037ca98] ip_auto_config+0x90/0x28c
[ef82bf60] [c0001b9c] do_one_initcall+0x34/0x1a0
[ef82bfd0] [c035e240] do_initcalls+0x38/0x58
[ef82bfe0] [c035e2c4] kernel_init+0x30/0x90
[ef82bff0] [c000f464] kernel_thread+0x4c/0x68
[...]
And again, ucc_geth_stop() assumes that ugeth->phydev is there, while
it isn't.
This patch fixes all three oopses simply by rearranging some code:
- In ucc_geth_open(): move init_phy() call to the beginning, so
that we only call ucc_geth_stop() with a PHY attached;
- Move phy_disconnect() call from ucc_geth_close() to
ucc_geth_stop(), so that we'll always disconnect the PHY.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Dmitri Vorobiev [Fri, 27 Mar 2009 22:55:36 +0000 (15:55 -0700)]
net: Add missing include into include/linux/netdevice.h
The inline function skb_gro_mac_header defined in include/linux/netdevice.h
makes use of page_address(). Depending on configuration options, the latter
is either defined as a macro or is declared as a function in another header
file, namely include/linux/mm.h. However, include/linux/netdevice.h does not
include include/linux/mm.h.
On MIPS, this has produced the following build error:
CC kernel/sysctl_check.o
In file included from include/linux/icmpv6.h:173,
from include/linux/ipv6.h:208,
from include/net/ip_vs.h:26,
from kernel/sysctl_check.c:6:
include/linux/netdevice.h: In function 'skb_gro_mac_header':
include/linux/netdevice.h:1132: error: implicit declaration of function
'page_address'
include/linux/netdevice.h:1133: warning: pointer/integer type mismatch
in conditional expression
make[1]: *** [kernel/sysctl_check.o] Error 1
make: *** [kernel] Error 2
The patch adds the missing include and fixes the build error.
Signed-off-by: Dmitri Vorobiev <dmitri.vorobiev@movial.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Chris Wilson [Fri, 6 Mar 2009 23:27:52 +0000 (23:27 +0000)]
drm/i915: Check for dev->primary->master before dereference.
I've hit the occasional oops inside i915_wait_ring() with an indication of
a NULL derefence of dev->primary->master. Adding a NULL check is
consistent with the other potential users of dev->primary->master.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net>
Zhenyu Wang [Wed, 4 Mar 2009 11:36:01 +0000 (19:36 +0800)]
drm/i915: Fix TV get_modes to return modes count
The get_modes hook must return the number of modes added. This also fixes
TV mode's clock calculation int overflow issue, and use 0.01 precision for
mode refresh validation.
Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net>
i915/drm: Remove two redundant agp_chipset_flushes
agp_chipset_flush() is for flushing the intel GMCH write cache via the
IFP, these two uses are for when we're getting the object into the cpu
READ domain, and thus should not be needed. This confused me when I was
getting my head around the code.
With thanks to airlied for helping me check my mental picture of how the
flushes and clflushes are supposed to be used.
Signed-off-by: Owain G. Ainsworth <oga@openbsd.org> Signed-off-by: Eric Anholt <eric@anholt.net>
Eric Anholt [Wed, 18 Feb 2009 07:53:41 +0000 (23:53 -0800)]
drm/i915: Add information on pinning and fencing to the i915 list debug.
This was inspired by a patch by Chris Wilson, though none of it applied in any
way due to the debugfs work and I decided to change the formatting of the
new information anyway.
Ben Gamari [Wed, 18 Feb 2009 01:08:49 +0000 (20:08 -0500)]
drm: Convert proc files to seq_file and introduce debugfs
The old mechanism to formatting proc files is extremely ugly. The
seq_file API was designed specifically for cases like this and greatly
simplifies the process.
Also, most of the files in /proc really don't belong there. This patch
introduces the infrastructure for putting these into debugfs and exposes
all of the proc files in debugfs as well.
Signed-off-by: Ben Gamari <bgamari@gmail.com> Signed-off-by: Eric Anholt <eric@anholt.net>
Linus Torvalds [Fri, 27 Mar 2009 21:48:34 +0000 (14:48 -0700)]
Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-quota-2.6
* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-quota-2.6: (27 commits)
ext2: Zero our b_size in ext2_quota_read()
trivial: fix typos/grammar errors in fs/Kconfig
quota: Coding style fixes
quota: Remove superfluous inlines
quota: Remove uppercase aliases for quota functions.
nfsd: Use lowercase names of quota functions
jfs: Use lowercase names of quota functions
udf: Use lowercase names of quota functions
ufs: Use lowercase names of quota functions
reiserfs: Use lowercase names of quota functions
ext4: Use lowercase names of quota functions
ext3: Use lowercase names of quota functions
ext2: Use lowercase names of quota functions
ramfs: Remove quota call
vfs: Use lowercase names of quota functions
quota: Remove dqbuf_t and other cleanups
quota: Remove NODQUOT macro
quota: Make global quota locks cacheline aligned
quota: Move quota files into separate directory
ext4: quota reservation for delayed allocation
...
Linus Torvalds [Fri, 27 Mar 2009 21:48:07 +0000 (14:48 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
dlm: fix length calculation in compat code
dlm: ignore cancel on granted lock
dlm: clear defunct cancel state
dlm: replace idr with hash table for connections
dlm: comment typo fixes
dlm: use ipv6_addr_copy
dlm: Change rwlock which is only used in write mode to a spinlock
Eric Anholt [Wed, 11 Mar 2009 19:30:04 +0000 (12:30 -0700)]
drm/i915: Fix lock order reversal with cliprects and cmdbuf in non-DRI2 paths.
This introduces allocation in the batch submission path that wasn't there
previously, but these are compatibility paths so we care about simplicity
more than performance.
kernel.org bug #12419.
Signed-off-by: Eric Anholt <eric@anholt.net> Reviewed-by: Keith Packard <keithp@keithp.com> Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Eric Anholt [Mon, 9 Mar 2009 20:42:30 +0000 (13:42 -0700)]
drm/i915: Fix lock order reversal in shmem pwrite path.
Like the GTT pwrite path fix, this uses an optimistic path and a
fallback to get_user_pages. Note that this means we have to stop using
vfs_write and roll it ourselves.
Signed-off-by: Eric Anholt <eric@anholt.net> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Eric Anholt [Thu, 19 Mar 2009 21:10:50 +0000 (14:10 -0700)]
drm/i915: Make GEM object's page lists refcounted instead of get/free.
We've wanted this for a few consumers that touch the pages directly (such as
the following commit), which have been doing the refcounting outside of
get/put pages.
Signed-off-by: Eric Anholt <eric@anholt.net> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Eric Anholt [Mon, 9 Mar 2009 16:42:23 +0000 (09:42 -0700)]
drm/i915: Fix lock order reversal in GTT pwrite path.
Since the pagefault path determines that the lock order we use has to be
mmap_sem -> struct_mutex, we can't allow page faults to occur while the
struct_mutex is held. To fix this in pwrite, we first try optimistically to
see if we can copy from user without faulting. If it fails, fall back to
using get_user_pages to pin the user's memory, and map those pages
atomically when copying it to the GPU.
Signed-off-by: Eric Anholt <eric@anholt.net> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Ma Ling [Wed, 18 Mar 2009 12:13:27 +0000 (20:13 +0800)]
drm/i915: Use a different PLL timing search function on G4X.
This improves the PLL timings according to the suggestion of the hardware
engineers. This results in some outputs being able to sync that weren't
able to before.
This is part of fixing fd.o bug #17508.
Signed-off-by: Ma Ling <ling.ma@intel.com>
[anholt: cleaned up a couple of redundant comments] Signed-off-by: Eric Anholt <eric@anholt.net>
Eric Anholt [Thu, 12 Mar 2009 23:27:11 +0000 (16:27 -0700)]
drm/i915: Change DCC tiling detection case to cover only mobile parts.
Later spec investigation has revealed that every 9xx mobile part has
had this register in this format. Also, no non-mobile parts have been shown
to have this register. So make all mobile use the same code, and all
non-mobile use the hack 965 detection.
DVB: firedtv: FireDTV S2 problems with tuning solved
Signed-off-by: Beat Michel Liechti <bml303@gmail.com>
Tuning was broken on FireDTV S2 (and presumably FloppyDTV S2) because a
wrong opcode was sent. The box only gave "not implemented" responses.
Changing the opcode to _TUNE_QPSK2 fixes this for good.
Cc: stable@kernel.org Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Nick Piggin [Wed, 11 Mar 2009 20:17:36 +0000 (13:17 -0700)]
fs: avoid I_NEW inodes
To be on the safe side, it should be less fragile to exclude I_NEW inodes
from inode list scans by default (unless there is an important reason to
have them).
Normally they will get excluded (eg. by zero refcount or writecount etc),
however it is a bit fragile for list walkers to know exactly what parts of
the inode state is set up and valid to test when in I_NEW. So along these
lines, move I_NEW checks upward as well (sometimes taking I_FREEING etc
checks with them too -- this shouldn't be a problem should it?)
Signed-off-by: Nick Piggin <npiggin@suse.de> Acked-by: Jan Kara <jack@suse.cz> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Merge code for single and multiple-instance mounts
new_pts_mount() (including the get_sb_nodev()), shares a lot of code
with init_pts_mount(). The only difference between them is the 'test-super'
function passed into sget().
Move all common code into devpts_get_sb() and remove the new_pts_mount() and
init_pts_mount() functions,
Changelog[v3]:
[Serge Hallyn]: Remove unnecessary printk()s
Changelog[v2]:
(Christoph Hellwig): Merge code in 'do_pts_mount()' into devpts_get_sb()
With mknod_ptmx() moved to devpts_get_sb(), init_pts_mount() becomes
a wrapper around get_init_pts_sb(). Remove get_init_pts_sb() and
fold code into init_pts_mount().
We create 'ptmx' node in both single-instance and multiple-instance
mounts. So devpts_get_sb() can call mknod_ptmx() once rather than
have both modes calling mknod_ptmx() separately.
Parse mount options just once and copy them to super block
Since all the mount option parsing is done in devpts, we could do it
just once and pass it around in devpts functions and eventually store
it in the super block.
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
On remount, devpts fs only needs to parse the mount options. Users cannot
directly create/dirty files in /dev/pts so the MS_RDONLY flag and
shrinking the dcache does not really apply to devpts.
So effectively on remount, devpts only parses the mount options and updates
these options in its super block. As such, we could replace do_remount_sb()
call with a direct parse_mount_options().
Doing so enables subsequent patches to avoid parsing the mount options twice
and simplify the code.
simple_set_mnt() is defined as returning 'int' but always returns 0.
Callers assume simple_set_mnt() never fails and don't properly cleanup if
it were to _ever_ fail. For instance, get_sb_single() and get_sb_nodev()
should:
up_write(sb->s_unmount);
deactivate_super(sb);
if simple_set_mnt() fails.
Since simple_set_mnt() never fails, would be cleaner if it did not
return anything.
[akpm@linux-foundation.org: fix build] Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Acked-by: Serge Hallyn <serue@us.ibm.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Christoph Hellwig <hch@lst.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Jan Engelhardt [Wed, 11 Feb 2009 21:14:54 +0000 (13:14 -0800)]
fs: make struct dentry->d_op const
This change will allow for tagging many dentry_operations const in the
source tree.
Signed-off-by: Jan Engelhardt <jengelh@medozas.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
We should drop the ->s_umount mutex if an error occurs after the
sget()/grab_super() call. This was introduced when adding support
for multiple instances of devpts and noticed during a code review/reorg.
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
vfs: Update fs.h to use inline functions when no file locking set
This avoids various issues which might give rise to compiler warnings
about missing functions and/or unused variable with the previous
macros. This also fixes a bug where one of the macros was returning
0, but it should have been void.
Reported-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Steven Whitehouse <swhiteho@redhat.com> Tested-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Duane Griffin [Thu, 8 Jan 2009 22:43:51 +0000 (22:43 +0000)]
ufs: copy symlink data into the correct union member
Copy symlink data into the union member it is accessed through. Although
this shouldn't make a difference to behaviour it makes the code easier
to follow and grep through. It may also prevent problems if the
struct/union definitions change in the future.
Signed-off-by: Duane Griffin <duaneg@dghda.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Duane Griffin [Thu, 8 Jan 2009 22:43:49 +0000 (22:43 +0000)]
ufs: don't truncate longer ufs2 fast symlinks
ufs2 fast symlinks can be twice as long as ufs ones, however the code
was using the ufs size in various places. Fix that so ufs2 symlinks over
60 characters aren't truncated.
Note that we copy the entire area instead of using the maxsymlinklen field
from the superblock. This way we will be more robust against corruption (of
the superblock).
While we are at it, use memcpy instead of open-coding it with for loops.
Signed-off-by: Duane Griffin <duaneg@dghda.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Duane Griffin [Thu, 8 Jan 2009 22:43:48 +0000 (22:43 +0000)]
ufs: validate maximum fast symlink size from superblock
The maximum fast symlink size is set in the superblock of certain types
of UFS filesystem. Before using it we need to check that it isn't longer
than the available space we have in the inode.
Signed-off-by: Duane Griffin <duaneg@dghda.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Add a switch for the various i_mode fmt cases, and remove the comment
about writeability of devices nodes - that part is handled in
inode_permission and comment on (briefly) there.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Make sure that comments describe what's going on and not how, and always
use __d_instantiate instead of two separate branches, one with
d_instantiate and one with __d_instantiate.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Due to a different size of ino_t ustat needs a compat handler, but
currently only x86 and mips provide one. Add a generic compat_sys_ustat
and switch all architectures over to it. Instead of doing various
user copy hacks compat_sys_ustat just reimplements sys_ustat as
it's trivial. This was suggested by Arnd Bergmann.
Found by Eric Sandeen when running xfstests/017 on ppc64, which causes
stack smashing warnings on RHEL/Fedora due to the too large amount of
data writen by the syscall.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Jes Sorensen [Fri, 27 Mar 2009 15:55:41 +0000 (16:55 +0100)]
[IA64] Fix kstat_this_cpu build breakage
arch/ia64/kernel/irq_ia64.c: In function 'ia64_handle_irq':
arch/ia64/kernel/irq_ia64.c:498: error: 'struct kernel_stat' has no member named 'irqs'
arch/ia64/kernel/irq_ia64.c:500: error: 'struct kernel_stat' has no member named 'irqs'
arch/ia64/kernel/irq_ia64.c: In function 'ia64_process_pending_intr':
arch/ia64/kernel/irq_ia64.c:556: error: 'struct kernel_stat' has no member named 'irqs'
arch/ia64/kernel/irq_ia64.c:558: error: 'struct kernel_stat' has no member named 'irqs'
David S. Miller [Fri, 27 Mar 2009 08:09:17 +0000 (01:09 -0700)]
sparc64: Fix MM refcount check in smp_flush_tlb_pending().
As explained by Benjamin Herrenschmidt:
> CPU 0 is running the context, task->mm == task->active_mm == your
> context. The CPU is in userspace happily churning things.
>
> CPU 1 used to run it, not anymore, it's now running fancyfsd which
> is a kernel thread, but current->active_mm still points to that
> same context.
>
> Because there's only one "real" user, mm_users is 1 (but mm_count is
> elevated, it's just that the presence on CPU 1 as active_mm has no
> effect on mm_count().
>
> At this point, fancyfsd decides to invalidate a mapping currently mapped
> by that context, for example because a networked file has changed
> remotely or something like that, using unmap_mapping_ranges().
>
> So CPU 1 goes into the zapping code, which eventually ends up calling
> flush_tlb_pending(). Your test will succeed, as current->active_mm is
> indeed the target mm for the flush, and mm_users is indeed 1. So you
> will -not- send an IPI to the other CPU, and CPU 0 will continue happily
> accessing the pages that should have been unmapped.
To fix this problem, check ->mm instead of ->active_mm, and this
means:
> So if you test current->mm, you effectively account for mm_users == 1,
> so the only way the mm can be active on another processor is as a lazy
> mm for a kernel thread. So your test should work properly as long
> as you don't have a HW that will do speculative TLB reloads into the
> TLB on that other CPU (and even if you do, you flush-on-switch-in should
> get rid of any crap here).
And therefore we should be OK.
Signed-off-by: David S. Miller <davem@davemloft.net>
Divy Le Ray [Thu, 26 Mar 2009 16:39:29 +0000 (16:39 +0000)]
cxgb3: map entire Rx page, feed map+offset to Rx ring.
DMA mapping can be expensive in the presence of iommus.
Reduce the Rx iommu activity by mapping an entire page, and provide the H/W
the mapped address + offset of the current page chunk.
Reserve bits at the end of the page to track mapping references, so the page
can be unmapped.
Signed-off-by: Divy Le Ray <divy@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Divy Le Ray [Thu, 26 Mar 2009 16:39:14 +0000 (16:39 +0000)]
cxgb3: sge setup fixes
Enable timestamps, update delayed ack threshold for iSCSI/iWARP traffic
Remove the len flag in Tx requests. It might corrupt offload trace packets.
Update SGE context setup to avoid potential H/W misprogrammation.
Signed-off-by: Divy Le Ray <divy@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / mv-sheeva.git / log