07 Feb, 2008
40 commits
-
The early per-cpu handling needs a slight tweak to work when booting
on a non-zero cpu.We got away with this for a long time, but can't any longer as now
even printk() calls functions (cpu_clock() for example) that thus make
early references to per-cpu variables.Signed-off-by: David S. Miller
-
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
9p: fix p9_printfcall export
9p: transport API reorganization
9p: add remove function to trans_virtio
9p: Convert semaphore to spinlock for p9_idpool
9p: fix mmap to be read-only
9p: add support for sticky bit
9p: Fix soft lockup in virtio transport
9p: fix bug in attach-per-user
9p: block-based virtio client
9p: create transport rpc cut-thru
9p: fix bug in p9_clone_stat -
ERROR: "p9_printfcall" [net/9p/9pnet_virtio.ko] undefined!
Signed-off-by: Andrew Morton
Acked-by: Eric Van Hensbergen -
This merges the mux.c (including the connection interface) with trans_fd
in preparation for transport API changes. Ultimately, trans_fd will need
to be rewritten to clean it up and simplify the implementation, but this
reorganization is viewed as the first step.Signed-off-by: Eric Van Hensbergen
-
Request from rusty:
Just cleaning up patches for 2.6.25 merge, and noticed that
net/9p/trans_virtio.c doesn't have a remove function. This will crash when
removing the module (console doesn't have one because it can't really be
removed).Signed-off-by: Eric Van Hensbergen
-
When booting from v9fs, down_interruptible in p9_idpool_get() triggered a BUG
as it was being called with IRQs disabled. A spinlock seems like the right
thing to be using since the idr functions go out of their way not to sleep.This patch eliminates the BUG by converting the semaphore to a spinlock.
Signed-off-by: Anthony Liguori
Acked-by: Eric Van Hensbergen -
v9fs was allowing writable mmap which could lead to kernel BUG() cases.
This sets the mmap function to generic_file_readonly_mmap which (correctly)
returns an error to applications which open mmap for writing.Signed-off-by: Eric Van Hensbergen
-
GDM gets unhappy if /var/gdm doesn't have the sticky bit set. This patch adds
support for the sticky bit in much the same way setuid/setgid is supported.With this patch, I can launch X from a v9fs rootfs (although I quickly run out
of fds in the server once gnome starts up).Signed-off-by: Anthony Liguori
Acked-by: Eric Van Hensbergen -
This fixes a poorly placed spinlock which could result in a
soft lockup condition.Signed-off-by: Eric Van Hensbergen
-
When a new user attached at a directory other than the root, he would end
up in the parent directory of the cwd. This was due to a logic error in
the code which attaches the user at the mount point and walks back to the
cwd. This patch fixes that.Signed-off-by: Eric Van Hensbergen
-
This replaces the console-based virto client with a block-based
client using a single request queue.Signed-off-by: Eric Van Hensbergen
-
Add a new transport function which allows a cut-thru directly to
the transport instead of processing request through the mux if the
cut-thru exists.Signed-off-by: Eric Van Hensbergen
-
This patch fixes a bug in the copying of 9P
stat information where string references
weren't being updated properly.Signed-off-by: Martin Sava
Signed-off-by: Eric Van Hensbergen -
Docbook fatal error, file was moved:
docproc: linux-2.6.24-git15/drivers/base/dmapool.c: No such file or directorySigned-off-by: Randy Dunlap
Signed-off-by: Linus Torvalds -
* git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-x86:
x86: fix deadlock, make pgd_lock irq-safe
virtio: fix trivial build bug
x86: fix mttr trimming
x86: delay CPA self-test and repeat it
x86: fix 64-bit sections
generic: add __FINITDATA
x86: remove suprious ifdefs from pageattr.c
x86: mark the .rodata section also NX
x86: fix iret exception recovery on 64-bit
cpuidle: dubious one-bit signed bitfield in cpuidle.h
x86: fix sparse warnings in powernow-k8.c
x86: fix sparse error in traps_32.c
x86: trivial sparse/checkpatch in quirks.c
x86 ptrace: disallow null cs/ss
MAINTAINERS: RDC R-321x SoC maintainer
brk randomization: introduce CONFIG_COMPAT_BRK
brk: check the lower bound properly
x86: remove X2 workaround
x86: make spurious fault handler aware of large mappings
x86: make traps on entry code be debuggable in user space, 64-bit -
lockdep just caught this one:
=================================
[ INFO: inconsistent lock state ]
2.6.24 #38
---------------------------------
inconsistent {in-softirq-W} -> {softirq-on-W} usage.
swapper/1 [HC0[0]:SC0[0]:HE1:SE1] takes:
(pgd_lock){-+..}, at: [] mm_init+0x1da/0x250
{in-softirq-W} state was registered at:
[] 0xffffffffffffffff
irq event stamp: 394559
hardirqs last enabled at (394559): [] get_page_from_freelist+0x30a/0x4c0
hardirqs last disabled at (394558): [] get_page_from_freelist+0x125/0x4c0
softirqs last enabled at (393952): [] __do_softirq+0xce/0xe0
softirqs last disabled at (393945): [] call_softirq+0x1c/0x30other info that might help us debug this:
no locks held by swapper/1.stack backtrace:
Pid: 1, comm: swapper Not tainted 2.6.24 #38Call Trace:
[] print_usage_bug+0x18b/0x190
[] mark_lock+0x53d/0x560
[] __lock_acquire+0x3ca/0xed0
[] lock_acquire+0xa8/0xe0
[] ? mm_init+0x1da/0x250
[] _spin_lock+0x30/0x70
[] mm_init+0x1da/0x250
[] mm_alloc+0x39/0x50
[] bprm_mm_init+0x2a/0x1a0
[] do_execve+0x7b/0x220
[] sys_execve+0x46/0x70
[] kernel_execve+0x64/0xd0
[] ? _stext+0x1e/0x20
[] init_post+0x9a/0xf0
[] ? trace_hardirqs_on_thunk+0x35/0x3a
[] ? trace_hardirqs_on+0xba/0xd0
[] ? child_rip+0xa/0x12
[] ? restore_args+0x0/0x44
[] ? child_rip+0x0/0x12turns out that pgd_lock has been used on 64-bit x86 in an irq-unsafe
way for almost two years, since commit 8c914cb704a11460e.Signed-off-by: Ingo Molnar
Signed-off-by: Thomas Gleixner -
fix build bug:
drivers/virtio/virtio_balloon.c: In function 'fill_balloon':
drivers/virtio/virtio_balloon.c:98: error: implicit declaration of function 'msleep'Acked-by: Rusty Russell
Signed-off-by: Ingo Molnar -
Pavel Emelyanov reported that his networking card did not work
and bisected it down to:"
The commit093af8d7f0ba3c6be1485973508584ef081e9f93
x86_32: trim memory by updating e820broke my e1000 card: on loading driver says that
e1000: probe of 0000:04:03.0 failed with error -5
and the interface doesn't appear.
"on a 32-bit kernel, base will overflow when try to do PAGE_SHIFT,
and highest_addr will always less 4G.So use pfn instead of address to avoid the overflow when more than
4g RAM is installed on a 32-bit kernel.Many thanks to Pavel Emelyanov for reporting and testing it.
Bisected-by: Pavel Emelyanov
Signed-off-by: Yinghai Lu
Tested-by: Pavel Emelyanov
Signed-off-by: Ingo Molnar -
delay the CPA self-test so that any impact (corruption) of
user-space pagetables can be triggered. Repeat the test
every 30 seconds.this would have prevented the bug fixed by 8cb2a7c1e95e472b5,
at its source.Signed-off-by: Ingo Molnar
-
fix 64-bit section warnings.
Signed-off-by: Ingo Molnar
-
Signed-off-by: Ingo Molnar
-
The .rodata section really should just be read only; the config option
is there to make breaking up the 2Mb page an option (so people whos machines
give more performance for the 2Mb case can opt to do so).
But when the page gets split anyway, this is no longer an issue, so
clean up the code and remove the ifdefsSigned-off-by: Arjan van de Ven
Signed-off-by: Ingo Molnar -
The .rodata section shouldn't just be read-only,
but also non-executable. This is free since we've broken
up the 2MB page already anyway.also update test_nx to check for this.
Signed-off-by: Arjan van de Ven
Signed-off-by: Ingo Molnar -
This change broke recovery of exceptions in iret:
commit 72fe4858544292ad64600765cb78bc02298c6b1c
Author: Glauber de Oliveira Costax86: replace privileged instructions with paravirt macros
The ENTRY(native_iret) macro adds alignment padding before the iretq
instruction, so "iret_label" no longer points exactly at the instruction.
It was sloppy to leave the old "iret_label" label behind when replacing
its nearby use. Removing it would have revealed the other use of the
label later in the file, and upon noticing that use, anyone exercising
the minimum of attention to detail expected of anyone touching this
subtle code would realize it needed to change as well.Signed-off-by: Roland McGrath
Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
fix these sparse warnings:
CHECK arch/x86/kernel/acpi/cstate.c
include/linux/cpuidle.h:82:17: error: dubious one-bit signed bitfield
CHECK arch/x86/kernel/acpi/processor.c
include/linux/cpuidle.h:82:17: error: dubious one-bit signed bitfield
CHECK arch/x86/kernel/cpu/cpufreq/powernow-k7.c
include/linux/cpuidle.h:82:17: error: dubious one-bit signed bitfield
CHECK arch/x86/kernel/cpu/cpufreq/powernow-k8.c
include/linux/cpuidle.h:82:17: error: dubious one-bit signed bitfield
CHECK arch/x86/kernel/cpu/cpufreq/longhaul.c
include/linux/cpuidle.h:82:17: error: dubious one-bit signed bitfield
CHECK arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
include/linux/cpuidle.h:82:17: error: dubious one-bit signed bitfieldSigned-off-by: Harvey Harrison
Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
arch/x86/kernel/cpu/cpufreq/powernow-k8.c:830:7: warning: symbol 'hi' shadows an earlier one
arch/x86/kernel/cpu/cpufreq/powernow-k8.c:824:6: originally declared here
arch/x86/kernel/cpu/cpufreq/powernow-k8.c:830:15: warning: symbol 'lo' shadows an earlier one
arch/x86/kernel/cpu/cpufreq/powernow-k8.c:824:14: originally declared hereSigned-off-by: Harvey Harrison
Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
This was being used to ensure the proper alignment of the FXSAVE/FXRSTOR data.
This would create a sparse error in the _correct_ cases, hiding further
warnings. Use BUILD_BUG_ON instead.Signed-off-by: Harvey Harrison
Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
arch/x86/kernel/quirks.c:384:3: warning: returning void-valued expression
arch/x86/kernel/quirks.c:387:3: warning: returning void-valued expression
arch/x86/kernel/quirks.c:390:3: warning: returning void-valued expression
arch/x86/kernel/quirks.c:393:3: warning: returning void-valued expressionSigned-off-by: Harvey Harrison
Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
In my revamp of the x86 ptrace code for setting register values,
I accidentally omitted a check that was there in the old code.
Allowing %cs to be 0 causes a bad crash in recovery from iret failure.
This patch fixes that regression against 2.6.24, and adds a comment
that should help prevent this subtlety from being overlooked again.Signed-off-by: Roland McGrath
Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
Signed-off-by: Florian Fainelli
Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
based on similar patch from: Pavel Machek
Introduce CONFIG_COMPAT_BRK. If disabled then the kernel is free
(but not obliged to) randomize the brk area.Heap randomization breaks ancient binaries, so we keep COMPAT_BRK
enabled by default.Signed-off-by: Ingo Molnar
-
There is a check in sys_brk(), that tries to make sure that we do not
underflow the area that is dedicated to brk heap.The check is however wrong, as it assumes that brk area starts immediately
after the end of the code (+bss), which is wrong for example in
environments with randomized brk start. The proper way is to check whether
the address is not below the start_brk address.Signed-off-by: Jiri Kosina
Signed-off-by: Ingo Molnar -
With the spurious handler fix, the X2 does not lock up anymore.
Signed-off-by: Ingo Molnar
-
In very rare cases, on certain CPUs, we could end up in the spurious
fault handler and ignore a large pud/pmd mapping. The resulting pte
pointer points into the mapped physical space and dereferencing it
will fault recursively.Make the code aware of large mappings and do the permission check
on the pmd/pud entry, when a large pud/pmd mapping is detected.Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar -
Unify the x86-64 behavior for 32-bit processes that set
bogus %cs/%ss values (the only ones that can fault in iret)
match what the native i386 behavior is. (do not kill the task
via do_exit but generate a SIGSEGV signal)[ tglx@linutronix.de: build fix ]
Signed-off-by: Roland McGrath
Signed-off-by: Ingo Molnar -
* 'async-tx-for-linus' of git://lost.foo-projects.org/~dwillia2/git/iop:
async_tx: allow architecture specific async_tx_find_channel implementations
async_tx: replace 'int_en' with operation preparation flags
async_tx: kill tx_set_src and tx_set_dest methods
async_tx: kill ASYNC_TX_ASSUME_COHERENT
iop-adma: use LIST_HEAD instead of LIST_HEAD_INIT
async_tx: use LIST_HEAD instead of LIST_HEAD_INIT
async_tx: fix compile breakage, mark do_async_xor __always_inline -
Signed-off-by: Daniel Walker
Signed-off-by: Linus Torvalds -
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/selinux-2.6:
SELinux: Remove security_get_policycaps()
security: allow Kconfig to set default mmap_min_addr protection -
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
ata_piix.c:piix_init_one() must be __devinit
sata_via.c: Remove missleading comment.
libata-core: unblacklist HITACHI drives
sata_nv: fix ATAPI issues with memory over 4GB (v7)
ata: drivers/ata/sata_mv.c needs dmapool.h
libata: kill now unused n_iter and fix sata_fsl
ahci: fix CAP.NP and PI handling
sata_mv: Support SoC controllers
Rename: linux/pata_platform.h to linux/ata_platform.h -
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (35 commits)
virtio net: fix oops on interface-up
Fix PHY Lib support for gianfar and ucc_geth
forcedeth: preserve registers
forcedeth: phy status fix
forcedeth: restart tx/rx
ipvs: Make wrr "no available servers" error message rate-limited
[PPPOL2TP]: Label unused warning when CONFIG_PROC_FS is not set.
[NET_SCHED]: cls_flow: support classification based on VLAN tag
[VLAN]: Constify skb argument to vlan_get_tag()
[NET_SCHED]: cls_flow: fix key mask validity check
[NET_SCHED]: em_meta: fix compile warning
b43: Fix DMA for 30/32-bit DMA engines
b43: fix build with CONFIG_SSB_PCIHOST=n
mac80211: Is not EXPERIMENTAL anymore
iwl3945-base.c: fix off-by-one errors
b43legacy: fix DMA slot resource leakage
b43legacy: drop packets we are not able to encrypt
b43legacy: fix suspend/resume
b43legacy: fix PIO crash
Generic HDLC - use random_ether_addr()
...
