Alexander Nyberg [Wed, 10 Aug 2005 17:11:36 +0000 (10:11 -0700)]
[PATCH] ns558 list handling fix
Need to use list_for_entry_safe(), as we're removing items during the
traversal. list_for_each_entry() uses the first ptr also as an iterator, if
you kfree() it slab takes it, might poison it and then you try to use it to
iterate to the next object in list.
Fix the p-persistence CSMA algorithm which in simplex mode was starting
with a slottime delay before doing anything else as if there was carrier
collision resulting in bad performance on simplex links.
Signed-off-by: Ralf Baechle DL5RB <ralf@linux-mips.org> Acked-by: Jeff Garzik <jgarzik@pobox.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ben Dooks [Wed, 10 Aug 2005 15:45:14 +0000 (16:45 +0100)]
[PATCH] ARM: 2849/1: S3C24XX - USB host update (2848/1)
Patch from Ben Dooks
Rename the s3c2410_report_oc() to s3c2410_usb_report_oc()
as this is an usb specific function.
Change port power on the usb-simtec implementation to only
power up the output if both are set, as per the usb 1.1
specification
Signed-off-by: Ben Dooks <ben-linux@fluff.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Russell King [Wed, 10 Aug 2005 15:18:35 +0000 (16:18 +0100)]
[ARM] Control v6 'global' bit via Linux PTE entries
Unfortunately, we can't use the "user" bit in the page tables to
control whether a page table entry is "global" or "asid" specific,
since the vector page is mapped as "user" accessible but is not
process specific.
Therefore, give direct control of the ARMv6 "nG" (not global)
bit to the mm layers.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Markus Lidel [Tue, 9 Aug 2005 21:30:57 +0000 (14:30 -0700)]
[PATCH] I2O: added pci_request_regions() before using the controller
Added pci_request_regions() before using the controller to avoid duplicate
usage of the I2O controller when the dpt_i2o driver and I2O subsystem is
loaded at the same time.
Signed-off-by: Markus Lidel <Markus.Lidel@shadowconnect.com> Cc: James Bottomley <James.Bottomley@SteelEye.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Markus Lidel [Tue, 9 Aug 2005 17:08:03 +0000 (10:08 -0700)]
[PATCH] i2o: remove new configuration API
Remove new configuration API from i2o_config
The API-patch is still available from the I2O website (which is mentioned in
the kernel config now). It is removed because it creates a new binary
sysfs-attribute, which doesn't have the limitiation of 4k. Expect for the
Adaptec controllers, which has a limitation in the hardware this attribute
doesn't make sense anywhere else. Until the sysfs API provides an attribute
which doesn't buffer (like firmware) and let access to at least 64k blocks i
provide a separate patch...
(akpm: basically, this API was introduced post-2.6.12 and Markus wants to pull
it out before 2.6.13).
Signed-off-by: Markus Lidel <Markus.Lidel@shadowconnect.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Jackson [Tue, 9 Aug 2005 17:07:59 +0000 (10:07 -0700)]
[PATCH] cpuset release ABBA deadlock fix
Fix possible cpuset_sem ABBA deadlock if 'notify_on_release' set.
For a particular usage pattern, creating and destroying cpusets fairly
frequently using notify_on_release, on a very large system, this deadlock
can be seen every few days. If you are not using the cpuset
notify_on_release feature, you will never see this deadlock.
The existing code, on task exit (or cpuset deletion) did:
get cpuset_sem
if cpuset marked notify_on_release and is ready to release:
compute cpuset path relative to /dev/cpuset mount point
call_usermodehelper() forks /sbin/cpuset_release_agent with path
drop cpuset_sem
Unfortunately, the fork in call_usermodehelper can allocate memory, and
allocating memory can require cpuset_sem, if the mems_generation values
changed in the interim. This results in an ABBA deadlock, trying to obtain
cpuset_sem when it is already held by the current task.
To fix this, I put the cpuset path (which must be computed while holding
cpuset_sem) in a temporary buffer, to be used in the call_usermodehelper
call of /sbin/cpuset_release_agent only _after_ dropping cpuset_sem.
So the new logic is:
get cpuset_sem
if cpuset marked notify_on_release and is ready to release:
compute cpuset path relative to /dev/cpuset mount point
stash path in kmalloc'd buffer
drop cpuset_sem
call_usermodehelper() forks /sbin/cpuset_release_agent with path
free path
The sharp eyed reader might notice that this patch does not contain any
calls to kmalloc. The existing code in the check_for_release() routine was
already kmalloc'ing a buffer to hold the cpuset path. In the old code, it
just held the buffer for a few lines, over the cpuset_release_agent() call
that in turn invoked call_usermodehelper(). In the new code, with the
application of this patch, it returns that buffer via the new char
**ppathbuf parameter, for later use and freeing in cpuset_release_agent(),
which is called after cpuset_sem is dropped. Whereas the old code has just
one call to cpuset_release_agent(), right in the check_for_release()
routine, the new code has three calls to cpuset_release_agent(), from the
various places that a cpuset can be released.
This patch has been build and booted on SN2, and passed a stress test that
previously hit the deadlock within a few seconds.
Signed-off-by: Paul Jackson <pj@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] i8xx_tco.c: arm watchdog only when started
i8xx_tco.c v0.08: only "arm" the watchdog when the watchdog has been
started. (Kernel Bug 4251: system reset when battery is read and i8xx_tco
driver loaded)
Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
default_idle() is looking at new variable can_do_pal_halt. However,
that variable did not get cleared upon "nohalt" boot option. Result
is that "nohalt" option is ignored until perfmon is exercised.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
Kumar Gala [Mon, 8 Aug 2005 21:49:17 +0000 (14:49 -0700)]
[PATCH] ppc32: Fix MPC834x USB memory map offsets
The memory mappings for MPC8349 USB MPH and DR modules were reversed.
Signed-off-by: Li Yang <LeoLi@freescale.com> Signed-off-by: Jiang Bo <Tanya.jiang@freescale.com> Signed-off-by: Kumar Gala <kumar.gala@freescale.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[SCSI] fix target scanning oops with fc transport class
We have some nasty issues with 2.6.12-rc6. Any request to scan on
the lpfc or qla2xxx FC adapters will oops. What is happening is the
system is defaulting to non-transport registered targets, which
inherit the parent of the scan. On this second scan, performed by
the attribute, the parent becomes the shost instead of the rport.
The slave functions in the 2 FC adapters use starget_to_rport()
routines, which incorrectly map the shost as an rport pointer.
Additionally, this pointed out other weaknesses:
- If the target structure is torn down outside of the transport,
we have no method for it to be regenerated at the proper parent.
- We have race conditions on the target being allocated by both
the midlayer scan (parent=shost) and by the fc transport
(parent=rport).
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
John McCutchan [Mon, 8 Aug 2005 17:52:16 +0000 (13:52 -0400)]
[PATCH] fsnotify_name/inoderemove
The patch below unhooks fsnotify from vfs_unlink & vfs_rmdir. It
introduces two new fsnotify calls, that are hooked in at the dcache
level. This not only more closely matches how the VFS layer works, it
also avoids the problem with locking and inode lifetimes.
The two functions are
- fsnotify_nameremove -- called when a directory entry is going away.
It notifies the PARENT of the deletion. This is called from
d_delete().
- inoderemove -- called when the files inode itself is going away. It
notifies the inode that is being deleted. This is called from
dentry_iput().
Signed-off-by: John McCutchan <ttb@tentacle.dhs.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Michael Krufky [Mon, 8 Aug 2005 16:22:43 +0000 (09:22 -0700)]
[PATCH] dvb: lgdt330x frontend: some bug fixes & add lgdt3303 support
- Structural changes within lgdt330x driver, framework now supports
both chips... tested OK on lgdt3302 and lgdt3303.
- Add LG/TUA6034 dvb_pll_desc for ATSC with LG TDVS-H062F & DViCO FusionHDTV5.
- Fixed LGDT330X signal strength: For now, always set it to 0.
- Corrected LGDT330X boundary condition error in read_snr: dB calculation.
Signed-off-by: Mac Michaels <wmichaels1@earthlink.net> Signed-off-by: Michael Krufky <mkrufky@m1k.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Harald Welte [Mon, 8 Aug 2005 09:50:55 +0000 (11:50 +0200)]
[PATCH] don't try to do any NAT on untracked connections
With the introduction of 'rustynat' in 2.6.11, the old tricks of preventing
NAT of 'untracked' connections (e.g. NOTRACK target in 'raw' table) are no
longer sufficient.
The ip_conntrack_untracked.status |= IPS_NAT_DONE_MASK effectively
prevents iteration of the 'nat' table, but doesn't prevent nat_packet()
to be executed. Since nr_manips is gone in 'rustynat', nat_packet() now
implicitly thinks that it has to do NAT on the packet.
This patch fixes that problem by explicitly checking for
ip_conntrack_untracked in ip_nat_fn().
Signed-off-by: Harald Welte <laforge@netfilter.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ivan Kokshaysky [Mon, 8 Aug 2005 08:55:54 +0000 (12:55 +0400)]
[PATCH] VIA VT8235 PCI quirk
Like many other southbridges from different manufacturers, VIA VT8235
chip has two non-standard BARs for power management and SMBus registers
(see the datasheet at http://www.via.com.tw).
This new quirk routine fixes boot problem with 2.6.13-rc2/rc6 kernels on
Targa Visionary 811 Athlon64 laptop, as reported by Mikael Pettersson
<mikpe@csd.uu.se>.
Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Antonino Daplas [Mon, 8 Aug 2005 06:22:43 +0000 (14:22 +0800)]
[PATCH] nvidiafb: Fix initial display corruption on certain laptops
Reported by:Vincent Fortier (Bugzilla Bug 4768)
"At boot time the screen appears moved to the mid right portion of the actual
video pannel making the end of the line appears at the left edge... It simply
looks like moved half way to the right"
His particular hardware has a display with an unusual dimension (1920x1200) but
unfortunately has no EDID block. None of the entries in the global mode
database is correct for this particular display, and it particularly has
difficulty scaling up 640x480 (the default startup mode of nvidiafb) to
1920x1200 which causes the above described problem.
1, Add 1920x1200 to the global mode database.
2. Let nvidiafb base the startup mode from the flatpanel dimensions only if the
EDID block is absent, no boot mode parameter is specified by the user, and
a flatpanel/LCD display is attached.
Arnd Bergmann [Mon, 8 Aug 2005 04:12:42 +0000 (14:12 +1000)]
[PATCH] ppc64: add default config for BPA
This adds a bpa_defconfig file and make target. The config settings
are made for the current version of the Cell Processor Based Blade,
so there are not too many drivers enabled. A few more drivers might
get added in the future though.
James Bottomley [Mon, 8 Aug 2005 16:51:38 +0000 (11:51 -0500)]
[SCSI] Bug 4940 Repeatable Kernel Panic on Adaptec 2015S I20 device on bootup
From: "Salyzyn, Mark" <mark_salyzyn@adaptec.com>
Prevent driver from loading if another driver (i2o) has already claimed
the resources associated with the card. Discussion associated with this
bug can be referenced at http://bugzilla.kernel.org/show_bug.cgi?id=4940
where it was agreed to use pci_request_regions in both the dpt_i2o and
the i2o driver to prevent both drivers loading on the same adapter(s).
Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Rejections fixed up and
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[PATCH] Move the fix to align node_end_pfns to a proper location
Move the fix to align node_end_pfns to a proper location. The earlier fix
made the node_remap_start_vaddr to get misaligned causing remap_numa_kva to
barf again :-/
Miklos Szeredi [Sun, 7 Aug 2005 16:42:25 +0000 (09:42 -0700)]
[PATCH] namespace.c: fix bind mount from foreign namespace
I'm resending this patch, because I still believe it's the correct fix.
Tested before/after applying the patch with a test application
available from:
http://www.inf.bme.hu/~mszeredi/nstest.c
Bind mount from a foreign namespace results in an un-removable mount.
The reason is that mnt->mnt_namespace is copied from the old mount in
clone_mnt(). Because of this check_mnt() in sys_umount() will fail.
The solution is to set mnt->mnt_namespace to current->namespace in
clone_mnt(). clone_mnt() is either called from do_loopback() or
copy_tree(). copy_tree() is called from do_loopback() or
copy_namespace().
When called (directly or indirectly) from do_loopback(), always
current->namspace is being modified: check_mnt(nd->mnt). So setting
mnt->mnt_namespace to current->namspace is the right thing to do.
When called from copy_namespace(), the setting of mnt_namespace is
irrelevant, since mnt_namespace is reset later in that function for
all copied mounts.
Jamie said:
This patch is correct. The old code was buggy for more fundamental and
serious reason: it broke the invariant that a tree of vfsmnts all have the
same value of mnt_namespace (and the same for the mnt_list list).
Olaf Hering [Sun, 7 Aug 2005 16:42:24 +0000 (09:42 -0700)]
[PATCH] remove linux/pagemap.h from linux/swap.h
sparc can not include linux/pagemap.h because of the following circular
dependency:
asm-sparc/pgtable include linux/swap.h
linux/swap.h include now linux/pagemap.h
linux/pagemap.h include linux/mm.h
linux/mm.h include asm/pgtable.h
It needs to have the swp_entry_t type fully visible in pgtable.h,
we can't work around this using macros.
Signed-off-by: Olaf Hering <olh@suse.de> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Tom Duffy [Sun, 7 Aug 2005 16:42:16 +0000 (09:42 -0700)]
[PATCH] Make visws compile again
In file included from linux-2.6.13-rc5/arch/i386/kernel/timers/timer_pit.c:20:
linux-2.6.13-rc5/include/asm-i386/mach-visws/do_timer.h: In function `do_timer_overflow':
linux-2.6.13-rc5/include/asm-i386/mach-visws/do_timer.h:32: error: `i8259A_lock' undeclared (first use in this function)
linux-2.6.13-rc5/include/asm-i386/mach-visws/do_timer.h:32: error: (Each undeclared identifier is reported only once
linux-2.6.13-rc5/include/asm-i386/mach-visws/do_timer.h:32: error: for each function it appears in.)
make[3]: *** [arch/i386/kernel/timers/timer_pit.o] Error 1
make[2]: *** [arch/i386/kernel/timers] Error 2
make[1]: *** [arch/i386/kernel] Error 2
make: *** [_all] Error 2
Signed-off-by: Tom Duffy <thomas.duffy.99@alumni.brown.edu> Cc: Andrey Panin <pazke@orbita1.ru> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ivan Kokshaysky [Sun, 7 Aug 2005 09:49:59 +0000 (13:49 +0400)]
[PATCH] Fix restore of 64-bit PCI BAR's
For 64-bit BAR[i] only pci_dev->resource[i] is valid, ->resource[i+1]
slot is unused and contains zeroes in all fields.
So when we update a PCI BAR, all we need is just to check that we're
going to update a _valid_ resource.
Also make sure to write high bits - use "x >> 16 >> 16" (rather than the
simpler ">> 32") to avoid warnings on 32-bit architectures where we're
not going to have any high bits.
Russell King [Sun, 7 Aug 2005 13:20:26 +0000 (14:20 +0100)]
[PATCH] ARM: Make sa1100fb_display_dma_period() an inline function
This function produces a warning when CPU_FREQ=n. Since it's a very
simple calculation, make it inline instead of adding preprocessor
directives around it.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
When the sparse mem changes and the kexec changes
were merged into setup.c they came in, in the wrong order.
This patch changes the order so we don't run sparse_init
which uses the bootmem allocator until we all of the
reserve_bootmem calls has been made.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] i386 visws: Add machine_shutdown and emergency_restart
Another x86 subarchitecture bit I missed. This adds both
machine_emergency_restart missed in my reboot fixes and
machine_shutdown needed for kexec support.
Linus Torvalds [Sat, 6 Aug 2005 16:44:37 +0000 (09:44 -0700)]
ppc: Export __handle_mm_fault for MOL
When we did the handle_mm_fault cleanup and get_user_page() race fixes,
handle_mm_fault turned into an inline function that called the real
__handle_mm_fault() code. The export needed for MOL on ppc wasn't
updated to match the new world order, though.
Turn it into a GPL export while at it, since this is all about internal
interfaces and MOL is GPL'd anwyay.
Linus Torvalds [Sat, 6 Aug 2005 16:39:57 +0000 (09:39 -0700)]
Add fakey 'deflateBound()' function to the in-kernel zlib routines
It's not the real deflateBound() in newer zlib libraries, partly because
the upcoming usage of it won't have the "stream" available, so we can't
have the same interfaces anyway.
David S. Miller [Sat, 6 Aug 2005 13:35:48 +0000 (06:35 -0700)]
[TG3]: Save initial PCI state before registering the netdevice.
Else on SMP systems it is possible for hotplug to execute,
invoke tg3_open(), and end up loading the uninitialized
PCI register save area into the card.
Signed-off-by: David S. Miller <davem@davemloft.net>
Marcel Holtmann [Sat, 6 Aug 2005 10:36:51 +0000 (12:36 +0200)]
[Bluetooth] Remove unused functions and cleanup symbol exports
This patch removes the unused bt_dump() function and it also removes
its BT_DMP macro. It also unexports the hci_dev_get(), hci_send_cmd()
and hci_si_event() functions.
Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Marcel Holtmann [Sat, 6 Aug 2005 10:36:42 +0000 (12:36 +0200)]
[Bluetooth] Revert session reference counting fix
The fix for the reference counting problem of the signal DLC introduced
a race condition which leads to an oops. The reason for it is not fully
understood by now and so revert this fix, because the reference counting
problem is not crashing the RFCOMM layer and its appearance it rare.
Marcel Holtmann [Sat, 6 Aug 2005 10:36:36 +0000 (12:36 +0200)]
[Bluetooth] Send HCI_Reset for Kensington dongle
The Kensington Bluetooth USB adapter is based on a Broadcom chip
with the HID proxy support. To initialize these kind of devices
correctly it is necessary to send HCI_Reset as the first command.
The original resize implementation was incorrect in the following
points.
* actual allocation size of tag_index was shorter than real_max_size,
but assumed to be of the same size, possibly causing memory access
beyond the allocated area.
* bits in tag_map between max_deptn and real_max_depth were
initialized to 1's, making the tags permanently reserved.
In an attempt to fix above two bugs, I had removed allocation optimization
in init_tag_map and real_max_size. Tag map/index were allocated and freed
immediately during resize.
Unfortunately, I wasn't considering that tag map/index can be resized
dynamically with tags beyond new_depth active. This led to accessing
freed area after shrinking tags and led to the following bug reporting
thread on linux-scsi.
To fix the problem, I've revived real_max_depth without allocation
optimization in init_tag_map, and Andrew Vasquez confirmed that the
problem was fixed. As Jens is not going to be available for a week, he
asked me to make sure that this patch reaches you.
David Gibson [Fri, 5 Aug 2005 18:59:35 +0000 (11:59 -0700)]
[PATCH] Fix hugepage crash on failing mmap()
This patch fixes a crash in the hugepage code. unmap_hugepage_area() was
assuming that (due to prefault) PTEs must exist for all the area in
question. However, this may not be the case, if mmap() encounters an error
before the prefault and calls unmap_region() to clean up any partial
mapping.
Depending on the hugepage configuration, this crash can be triggered by an
unpriveleged user.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
James Bottomley [Fri, 5 Aug 2005 18:59:34 +0000 (11:59 -0700)]
[PATCH] fix voyager compile after machine_emergency_restart breakage
[PATCH] i386: Implement machine_emergency_reboot
introduced this new function into arch/i386/reboot.c. However,
subarchitectures are entitled to implement their own copies of reboot.c
from which this new function is now missing.
It looks like visws will also need a similar fixup
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Marcel Selhorst [Fri, 5 Aug 2005 18:59:33 +0000 (11:59 -0700)]
[PATCH] tpm_infineon: Support for new TPM 1.2 and PNPACPI
This patch includes support for the new Infineon Trusted Platform Module
SLB 9635 TT 1.2 and does further include ACPI-support for both chip
versions (SLD 9630 TT 1.1 and SLB9635 TT 1.2). Since the ioports and
configuration registers are not correctly set on some machines, the
configuration is now done via PNPACPI, which reads out the correct values
out of the DSDT-table. Note that you have to have CONFIG_PNP,
CONFIG_ACPI_BUS and CONFIG_PNPACPI enabled to run this driver (assuming
that mainboards including a TPM do have the need for ACPI anyway).
Petr Vandrovec [Fri, 5 Aug 2005 13:50:07 +0000 (15:50 +0200)]
[PATCH] rtc: msleep() cannot be used from interrupt
Since the beginning of July my Opteron box was randomly crashing and
being rebooted by hardware watchdog. Today it finally did it in front
of me, and this patch will hopefully fix it.
The problem is that at the end of June (the 28th, to be exact: commit 47f176fdaf8924bc83fddcf9658f2fd3ef60d573, "[PATCH] Using msleep()
instead of HZ") rtc_get_rtc_time was converted to use msleep() instead
of busy waiting. But rtc_get_rtc_time is used by hpet_rtc_interrupt,
and scheduling is not allowed during interrupt. So I'm reverting this
part of original change, replacing msleep() back with busy loop.
The original code was busy waiting for up to 20ms, but on my hardware in
the worst case update-in-progress bit was asserted for at most 363
passes through loop (on 2GHz dual Opteron), much less than even one
jiffie, not even talking about 20ms. So I changed code to just wait
only as long as necessary. Otherwise when RTC was set to generate
8192Hz timer, it stopped doing anything for 20ms (160 pulses were
skipped!) from time to time, and this is rather suboptimal as far as I
can tell.
Signed-off-by: Petr Vandrovec <vandrove@vc.cvut.cz> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Simon Derr [Fri, 5 Aug 2005 02:52:03 +0000 (19:52 -0700)]
[PATCH] __vm_enough_memory() signedness fix
We have found what seems to be a small bug in __vm_enough_memory() when
sysctl_overcommit_memory is set to OVERCOMMIT_NEVER.
When this bug occurs the systems fails to boot, with /sbin/init whining
about fork() returning ENOMEM.
We hunted down the problem to this:
The deferred update mecanism used in vm_acct_memory(), on a SMP system,
allows the vm_committed_space counter to have a negative value.
This should not be a problem since this counter is known to be inaccurate.
But in __vm_enough_memory() this counter is compared to the `allowed'
variable, which is an unsigned long. This comparison is broken since it
will consider the negative values of vm_committed_space to be huge positive
values, resulting in a memory allocation failure.
Herbert Xu [Fri, 5 Aug 2005 02:52:02 +0000 (19:52 -0700)]
[PATCH] tcp: fix TSO cwnd caching bug
tcp_write_xmit caches the cwnd value indirectly in cwnd_quota. When
tcp_transmit_skb reduces the cwnd because of tcp_enter_cwr, the cached
value becomes invalid.
This patch ensures that the cwnd value is always reread after each
tcp_transmit_skb call.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Cc: "David S. Miller" <davem@davemloft.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
David S. Miller [Fri, 5 Aug 2005 02:52:01 +0000 (19:52 -0700)]
[PATCH] tcp: fix TSO sizing bugs
MSS changes can be lost since we preemptively initialize the tso_segs count
for an SKB before we %100 commit to sending it out.
So, by the time we send it out, the tso_size information can be stale due
to PMTU events. This mucks up all of the logic in our send engine, and can
even result in the BUG() triggering in tcp_tso_should_defer().
Another problem we have is that we're storing the tp->mss_cache, not the
SACK block normalized MSS, as the tso_size. That's wrong too.
Signed-off-by: David S. Miller <davem@davemloft.net> Cc: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
John McCutchan [Fri, 5 Aug 2005 01:12:54 +0000 (21:12 -0400)]
[PATCH] Clean up inotify delete race fix
This avoids the whole #ifdef mess by just getting a copy of
dentry->d_inode before d_delete is called - that makes the codepaths the
same for the INOTIFY/DNOTIFY cases as for the regular no-notify case.
I've been running this under a Gnome session for the last 10 minutes.
Inotify is being used extensively.
Signed-off-by: John McCutchan <ttb@tentacle.dhs.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
projects / linux-2.6 / log