27 Jun, 2007
5 commits
-
au_readl() is correct here.
Signed-off-by: Yoichi Yuasa
Signed-off-by: Ralf Baechle -
Signed-off-by: Yoichi Yuasa
Signed-off-by: Ralf Baechle -
Signed-off-by: Yoichi Yuasa
Signed-off-by: Ralf Baechle -
Signed-off-by: Yoichi Yuasa
Signed-off-by: Ralf Baechle -
Reported by Eugene Surovegin .
Signed-off-by: Ralf Baechle
25 Jun, 2007
3 commits
-
The return value of futex_find_get_task() needs to be -ESRCH in case
that the search fails. This was part of the original futex fixes and
got accidentally dropped, when the futex-tidy-up patch was split out.Signed-off-by: Thomas Gleixner
Cc: Ingo Molnar
Cc: Stable Team
Signed-off-by: Linus Torvalds -
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
[NET]: Make skb_seq_read unmap the last fragment
[NET]: Re-enable irqs before pushing pending DMA requests
[TCP] tcp_read_sock: Allow recv_actor() return return negative error value.
[PPP]: Fix osize too small errors when decoding mppe.
[PPP]: Revert 606f585e363527da9feaed79465132c0c661fd9e
[TIPC]: Fix infinite loop in netlink handler
[SKBUFF]: Fix incorrect config #ifdef around skb_copy_secmark
[IPV4]: include sysctl.h from inetdevice.h
[IPV6] NDISC: Fix thinko to control Router Preference support.
[NETFILTER]: nfctnetlink: Don't allow to change helper
[NETFILTER]: nf_conntrack_sip: add missing message types containing RTP info
24 Jun, 2007
32 commits
-
Register %ebx serves as the "global offset table base register" for
position-independent code. For absolute code, %ebx serves as a local
register and has no specified role in the function calling sequence. In
either case, a function must preserve the register value for the caller.acpi_copy_wakeup_routine overrides %ebx without saving it, this may corrupt
the called data.Kevin found that most time the value of Sx is saved in %esi, however
sometimes compiler also uses %ebx. When this happens, suspends fails since
sleep value in ebx is changed by acpi_copy_wakeup_routine.The same funtion in X86_64 doesn't have this problem.
Signed-off-by: Zhang Rui
Looks-okay-to: Pavel Machek
Signed-off-by: Rafael J. Wysocki
Cc: Len Brown
Acked-by: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
One of error path in ext4_read_inode() leaks bh since brelse is forgoten.
Signed-off-by: Kirill Korotaev
Acked-by: Vasily Averin
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
One of error path in ext3_read_inode() leaks bh since brelse is forgoten.
Signed-off-by: Kirill Korotaev
Acked-by: Vasily Averin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Removing a watched file will oops if audit is disabled (auditctl -e 0).
To reproduce:
- auditctl -e 1
- touch /tmp/foo
- auditctl -w /tmp/foo
- auditctl -e 0
- rm /tmp/foo (or mv)Signed-off-by: Tony Jones
Cc: Al Viro
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Yan Zheng pointed out that ext2_remount lacks checking if -o xip should be
enabled or not. This patch checks for presence of direct_access on the
backing block device and if the blocksize meets the requirements.Signed-off-by: Carsten Otte
Cc: Yan Zheng
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Background:
When a userspace application wants to know about machine check events, it
opens /dev/mcelog and does a read(). Usually, we found that this interface
works well, but in some cases, when the system was taking large numbers of
machine check exceptions, the read() would hang. The system would output a
soft-lockup warning, and the daemon reading from /dev/mcelog would suck up
as much of a single CPU as it could spinning in system space.Description:
This patch fixes this bug. In particular, there was a "continue" inside a
timeout loop that presumably was intended to break out of the outer loop,
but instead caused the inner loop to continue. This patch also makes the
condition for the break-out a little more evident by changing a
!time_before to a time_after_eq.Result:
The read() no longer hangs in this test case.Testing:
On my system, I could replicate the bug with the following command:
# for i in `seq 15000`; do ./inject_sbe.sh; done
where inject_sbe.sh contains commands to inject a single-bit error into the
next memory write transaction.Patch:
This patch is against git f1518a088bde6aea49e7c472ed6ab96178fcba3e.Signed-off-by: Joshua Wise
Signed-off-by: Tim Hockin
Cc: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Update to checkpatch.pl v0.06. Of note:
- do { and else handled correctly as control structures for { matching
- trailing whitespace correctly tripped when line otherwise empty
- support for const, including const foo * const bar
- multiline macros defining values correctly reportedThis version of checkpatch.pl can be found at the following URL:
http://www.kernel.org/pub/linux/kernel/people/apw/checkpatch/checkpatch.pl-0.06
Full Changelog:
Andy Whitcroft (14):
Version: 0.06
cleanup the Type regular expression declarations
fix up block counting
end of line counts as a space for ++ and --
do { needs the same checks as if, for et al
handle "const foo * const a" as a valid type
add spacing checks following ;
complete whitespace lines should trip trailing whitespace check
else is also a block control structure
badly formatted else can trip function declaration
detect and report trailing statements after else
types need to be terminated by a boundary
multiline macros defining values should be surrounded by parentheses
soften the wording of the Signed-off-by: warningsSigned-off-by: Andy Whitcroft
Cc: "Randy.Dunlap"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
MAX_ORDER is the first order that is not possible.
Use MAX_ORDER - 1 to calculate the larges possible object size in slab.h
Signed-off-by: Christoph Lameter
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
These should have been documented from the beginning. Fix it.
Signed-off-by: Dave Hansen
Cc: Al Viro
Cc: Christoph Hellwig
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Encourage developers to avoid the volatile type class in kernel code.
Signed-off-by: Jonathan Corbet
Signed-off-by: Jesper Juhl
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add asm-um/paravirt.h so that i386 headers that get pulled into UML
don't cause build failures when they want asm/paravirt.h.Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Get UML to use the generic bug support rather than arch specific one.
If I insert an artificial bug right before loading init, I get this:
Kernel panic - not syncing: Kernel mode signal 4
EIP: 0023:[] CPU: 0 Not tainted ESP: 002b:f7fd4fbc EFLAGS: 00000246
Not tainted
EAX: 00000000 EBX: 00007870 ECX: 00000013 EDX: 00007870
ESI: 0000786d EDI: 00000011 EBP: f7fd4fd8 DS: 002b ES: 002b
08273bec: [] show_regs+0x104/0x106
08273c08: [] panic_exit+0x2c/0x4b
08273c18: [] notifier_call_chain+0x32/0x5b
08273c38: [] __atomic_notifier_call_chain+0x30/0x32
08273c54: [] atomic_notifier_call_chain+0x2f/0x31
08273c70: [] panic+0x75/0x131
08273c94: [] relay_signal+0x87/0x95
08273cb0: [] sig_handler_common_skas+0x9e/0x120
08273cd8: [] sig_handler+0x28/0x4f
08273cec: [] handle_signal+0x53/0x89
08273d0c: [] hard_handler+0x18/0x28
08273d1c: [] transitions+0xf7d598b8/0xfffffff0With this patch in place, this is how it looks:
BUG: failure at init/main.c:779/init_post()!
Kernel panic - not syncing: BUG!EIP: 0023:[] CPU: 0 Not tainted ESP: 002b:f7f0dfbc EFLAGS: 00000246
Not tainted
EAX: 00000000 EBX: 000069db ECX: 00000013 EDX: 000069db
ESI: 000069d8 EDI: 00000011 EBP: f7f0dfd8 DS: 002b ES: 002b
098efedc: [] show_regs+0x104/0x106
098efef8: [] panic_exit+0x2c/0x4b
098eff08: [] notifier_call_chain+0x32/0x5b
098eff28: [] __atomic_notifier_call_chain+0x30/0x32
098eff44: [] atomic_notifier_call_chain+0x2f/0x31
098eff60: [] panic+0x75/0x131
098eff84: [] init_post+0xcd/0xe8
098eff9c: [] kernel_init+0x8e/0x9a
098effb4: [] run_kernel_thread+0x41/0x53
098effe0: [] new_thread_handler+0x62/0x8b
098efffc: [] 0xa55a5a5a[ jdike - added BUG_TABLE to linker script ]
Signed-off-by: Nick Piggin
Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The stallion driver oopses while initializing ISA cards due to an
uninitialized variable. This patch changes the initialisation order to
match the PCI code path.Signed-off-by: Ingo Korb
Acked-by: Jiri Slaby
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Hopefully this fixes http://bugzilla.kernel.org/show_bug.cgi?id=8635
The struct in6_addr passed to csum_ipv6_magic() is 4 byte aligned, so we
can't use the regular 64-bit loads. Since the cost of handling of 4 byte
and 1 byte aligned 64-bit data is roughly the same, this code can cope with
any src/dst [mis]alignment.Signed-off-by: Ivan Kokshaysky
Cc: Richard Henderson
Cc: Dustin Marquess
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This version brings a some new tests, and a host of changes to fix
false positives, of particular note:- detect 'var ++;' and 'var --;' as a bad combination
- multistatement #defines are now checked based on statement count
- multistatement #defines with initialisation correctly reported
- checks the location of the inline keywords
- EXPORT_SYMBOL for variables are now understood
- typedefs are loosened to handle sparse etcThis version of checkpatch.pl can be found at the following URL:
http://www.shadowen.org/~apw/public/checkpatch/checkpatch.pl-0.05
Full Changelog:
Andy Whitcroft (18):
Version: 0.05
macro definition checks should be for a single statement
avoid assignements only in if conditionals
declarations of function pointers need no space
multiline macros which are purely initialisation cannot be wrapped
EXPORT_SYMBOL can also directly follow a variable definition
check on the location of the inline keyword
EXPORT_SYMBOL needs to allow for attributes
ensure we do not find C99 // in strings
handle malformed #include lines
accept the {0,} form
typedefs are sensible for defining function pointer parameters
ensure { handling correctly handles nested switch() statements
trailing whitespace checks are not anchored
typedefs for sparse bitwise annotations make sense
update the type matcher to include sparse annotations
clean up indent and spacingSigned-off-by: Andy Whitcroft
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The intervals of domains that do not have SD_BALANCE_NEWIDLE must be
considered for the calculation of the time of the next balance. Otherwise
we may defer rebalancing forever.Siddha also spotted that the conversion of the balance interval
to jiffies is missing. Fix that to.From: Srivatsa Vaddagiri
also continue the loop if !(sd->flags & SD_LOAD_BALANCE).
Tested-by: Paul E. McKenney
It did in fact trigger under all three of mainline, CFS, and -rt including CFS
-- see below for a couple of emails from last Friday giving results for these
three on the AMD box (where it happened) and on a single-quad NUMA-Q system
(where it did not, at least not with such severity).Signed-off-by: Christoph Lameter
Signed-off-by: Ingo Molnar
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If slabs are allocated or freed from a large set of call sites (typical for
the kmalloc area) then we may create more output than fits into a single
PAGE and sysfs only gives us one page. The output should be truncated.
This patch fixes the checks to do the truncation properly.Signed-off-by: Christoph Lameter
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
When binding the driver, check the ID register for a valid identity, in case
the SM501 is not functioning correctly.Signed-off-by: Ben Dooks
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add documentation for the SM501 in Documentation/SM501.txt outlining the SM501
driver.Signed-off-by: Ben Dooks
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Ensure that the M1XCLK and MCLK are sourced from the same PLL (and refuse to
bind the driver if they are not).Update the PCI to safe initialisation values, as 72MHz is the maximum clock
for 33MHz PCI bus mastering.Signed-off-by: Ben Dooks
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The order of the set and mask operation in sm501_init_reg() was setting and
then masking the bits set. Correct the order so that we do not end up with
288MHz SDRAM clocks on certain systems.Signed-off-by: Ben Dooks
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This init sequence of setting the SDRAM clock before the bus clock is
recommend by Silicon Motion to stop problems with writes not sticking into
registers.Signed-off-by: Vincent Sanders
Signed-off-by: Ben Dooks
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch adds support for suspending the core (mfd driver) of the SM501.
Signed-off-by: Ben Dooks
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Recently my console UTF-8 patch went mainline. Here is an additional patch
that fixes two nasty issues and improves a third one, namely:1. My patch changed the behavior if a glyph is not found in the Unicode
mapping table. Previously for Unicode values less than 256 or 512 the
kernel tried to display the glyph from that position of the glyph table,
which could lead to a different accented letter being displayed. I
removed this fallback possibility and changed it to display the
replacement symbol.As Behdad pointed out, some fonts (e.g. sun12x22 from the kbd package)
lack Unicode mapping information, hence all you get is lots of question
marks. Though theoretically it's actually a user-space bug (the font
should be fixed), Behdad and I both believe that it'd be good to work
around in the kernel by re-introducing the fallback solution for ASCII
characters only. This sounds a quite reasonable decision, since all fonts
ship the ASCII characters in the first 128 positions. This way users
won't be surprised by lots of question marks just because s/he issued a
not-so-perfectly parameterized setfont command. As this fallback is only
re-introduced for code points below 128, you still won't see an accented
letter replaced by another, but at least you'll always get the English
letters right.2. My patch introduced "question mark with inverted color attributes" as a
last resort fallback glyph. Though it perfectly works on VGA console, on
framebuffer you may end up with question marks that are highlighed but
shouldn't be, and normal characters that are accidentally highlighed.
This is caused by missing FLUSHes when changing the color attribute.3. I've updated the table of double-width character based on Markus's
updated version. Only ten new code poings (one interval) is added.Signed-off-by: Egmont Koblinger
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
When a namespace is unshared, a refcount on the previous nsproxy is
abusively taken, leading to a memory leak of nsproxy objects.Signed-off-by: Cedric Le Goater
Cc: Badari Pulavarty
Cc: Herbert Poetzl
Cc: Oleg Nesterov
Cc: "Eric W. Biederman"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Jean Delvare
Cc: Rudolf Marek
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Having walked through the entire skbuff, skb_seq_read would leave the
last fragment mapped. As a consequence, the unwary caller would leak
kmaps, and proceed with preempt_count off by one. The only (kind of
non-intuitive) workaround is to use skb_seq_read_abort.This patch makes sure skb_seq_read always unmaps frag_data after
having cycled through the skb's paged part.Signed-off-by: Olaf Kirch
Signed-off-by: David S. Miller -
This moves the local_irq_enable() call in net_rx_action() to before
calling the CONFIG_NET_DMA's dma_async_memcpy_issue_pending() rather
than after. This shortens the irq disabled window and allows for DMA
drivers that need to do their own irq hold.Signed-off-by: Shannon Nelson
Signed-off-by: David S. Miller -
tcp_read_sock() currently assumes that the recv_actor() only returns
number of bytes copied. For network splice receive, we may have to
return an error in some cases. So allow the actor to return a negative
error value.Signed-off-by: Jens Axboe
Signed-off-by: David S. Miller -
The mppe_decompress() function required a buffer that is 1 byte too
small when receiving a message of mru size. This fixes buffer
allocation to prevent this from occurring.Signed-off-by: Konstantin Sharlaimov
Signed-off-by: David S. Miller -
This can cause packet buffer overflows in certain cases,
the real bug will be fixed differently in a followon
changeset.Signed-off-by: David S. Miller
-
The tipc netlink config handler uses the nlmsg_pid from the
request header as destination for its reply. If the application
initialized nlmsg_pid to 0, the reply is looped back to the kernel,
causing hangup. Fix: use nlmsg_pid of the skb that triggered the
request.Signed-off-by: Florian Westphal
Signed-off-by: David S. Miller
