introduces an inconsistency between esp and esp0 before the task is run the
first time. esp0 is no longer the actual start of the stack, but 8 bytes
off.
This shows itself clearly in a scenario when a ptracer that is set to also
ptrace eventual children traces program1 which then clones thread1. Now
the ptracer wants to modify the registers of thread1. The x86 ptrace
implementation bases it's knowledge about saved user-space registers upon
p->thread.esp0. But this will be a few bytes off causing certain writes to
the kernel stack to overwrite a saved kernel function address making the
kernel when actually running thread1 jump out into user-space. Very
spectacular.
The testcase I've used is:
/* start with strace -f ./a.out */
#include <pthread.h>
#include <stdio.h>
void *do_thread(void *p)
{
for (;;);
}
int main()
{
pthread_t one;
pthread_create(&one, NULL, &do_thread, NULL);
for (;;);
return 0;
}
So, my solution is to instead of just adjusting esp0 that creates an
inconsitent state I adjust where the user-space registers are saved with -8
bytes. This gives us the wanted extra bytes on the start of the stack and
esp0 is now correct. This solves the issues I saw from the original
testcase from Mateusz Berezecki and has survived testing here. I think
this should go into -mm a round or two first however as there might be some
cruft around depending on pt_regs lying on the start of the stack. That
however would have broken with the first change too!
It's actually a 2-line diff but I had to move the comment of why the -8 bytes
are there a few lines up. Thanks to Zwane for helping me with this.
Signed-off-by: Alexander Nyberg <alexn@telia.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch adds i18n support for make *config, allowing users to have the
config process in their own language.
No printk was harmed in the process, don't worry, so all the bug reports,
kernel messages, etc, remain in english, just the user tools to configure
the kernel are internationalized.
Users not interested in translations can just unset the related LANG,
LC_ALL, etc env variables and have the config process in plain english,
something like:
LANG= make menuconfig
is enough for having the whole config process in english. Or just don't
install any translation file.
Translations for brazilian portuguese are being done by a team of
volunteers at:
The current patch doesn't use any optimization to reduce the size of the
generated .mo file, it is possible to use the config option as a key, but
this doesn't prevent the current patch from being used or the translations
done under the current scheme to be in any way lost if we chose to do any
kind of keying.
Thanks to Fabricio Vaccari for starting the pt_BR (brazilian portuguese)
translation effort, Thiago Maciera for helping me with the gconf.cc (QT
frontent) i18n coding and to all the volunteers that are already working on
the first translation, to pt_BR.
I left the question on whether to ship the translations with the stock kernel
sources to be discussed here, please share your suggestions.
Signed-off-by: Arnaldo Carvalho de Melo <acme@conectiva.com.br> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org Signed-off-by: Andrew Morton <akpm@osdl.org>
this matches the API used by other link layer like ethernet or token
ring.
Signed-off-by: Christoph Hellwig <hch@lst.de> Acked-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Signed-off-by: David S. Miller <davem@davemloft.net>
This causes sk->sk_prot to change, which makes the socket
release free the sock into the wrong SLAB cache. Fix this
by introducing sk_prot_creator so that we always remember
where the sock came from.
Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Nathan Scott [Thu, 5 May 2005 20:33:20 +0000 (13:33 -0700)]
[XFS] Use the right offset when ensuring a delayed allocate conversion has covered the offset originally requested. Can cause data corruption when multiple processes are performing writeout on different areas of the same file. Quite difficult to hit though.
SGI Modid: xfs-linux:xfs-kern:22377a
Signed-off-by: Nathan Scott <nathans@sgi.com> Signed-off-by: Christoph Hellwig <hch@sgi.com>
.
Russell King [Thu, 5 May 2005 12:11:00 +0000 (13:11 +0100)]
[PATCH] ARM: Fix kernel stack offset calculations
Various places in the ARM kernel implicitly assumed that kernel
stacks are always 8K due to hard coded constants. Replace these
constants with definitions.
Correct the allowable range of kernel stack pointer values within
the allocation. Arrange for the entire kernel stack to be zeroed,
not just the upper 4K if CONFIG_DEBUG_STACK_USAGE is set.
Signed-off-by: Russell King <rmk@arm.linux.org.uk>
Roman Kagan [Wed, 13 Apr 2005 17:40:17 +0000 (21:40 +0400)]
[PATCH] drivers/base/bus.c: fix iteration in driver_detach()
With 2.6.11 and 2.6.12-rc2 (and perhaps a few versions before) usb
drivers for multi-interface devices, which do
usb_driver_release_interface() in their disconnect(), make rmmod hang.
It turns out to be due to a bug in drivers/base/bus.c:driver_detach(),
that iterates over the list of attached devices with
list_for_each_safe() under an assumption that device_release_driver()
only releases the current device, while it may also call
device_release_driver() for other devices on the same list.
The following patch fixes it. Please consider applying.
Signed-off-by: Roman Kagan <rkagan@mail.ru> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Dean Nelson [Tue, 22 Mar 2005 23:00:00 +0000 (16:00 -0700)]
[IA64-SGI] move nodepda pointer out of pda
Remove the p_nodepda and p_subnodepda pointers from the pda_s structure.
And then define a new per-cpu pointer to the nodepda and export it so
that it can be accessed by kernel modules.
Signed-off-by: Dean Nelson <dcn@sgi.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
Al Viro [Wed, 4 May 2005 04:40:12 +0000 (05:40 +0100)]
[PATCH] asm/signal.h unification
New file - asm-generic/signal.h. Contains declarations of
__sighandler_t, __sigrestore_t, SIG_DFL, SIG_IGN, SIG_ERR and default
definitions of SIG_BLOCK, SIG_UNBLOCK and SIG_SETMASK.
asm-*/signal.h switched to including it. The only exception is
asm-parisc/signal.h that wants its own declaration of __sighandler_t;
that one is left as-is.
asm-ppc64/signal.h required one more thing - unlike everybody else it
used __sigrestorer_t instead of usual __sigrestore_t. PPC64 switched to
common spelling.
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Al Viro [Wed, 4 May 2005 04:39:52 +0000 (05:39 +0100)]
[PATCH] ISA DMA Kconfig fixes - part 4 (irda)
* net/irda/irda_device.c::irda_setup_dma() made conditional on
ISA_DMA_API (it uses helpers in question and irda is usable on
platforms that don't have them at all - think of USB IRDA, for
example).
* irda drivers that depend on ISA DMA marked as dependent on
ISA_DMA_API
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Al Viro [Wed, 4 May 2005 04:39:32 +0000 (05:39 +0100)]
[PATCH] ISA_DMA Kconfig fixes - part 2 (parport_pc)
Part of parport_pc that uses ISA DMA helpers made conditional on
CONFIG_ISA_DMA_API. As the result, driver got usable for boxen that do
not have ISA DMA stuff and have normal PCI parport card stuck into
them - these never use DMA anyway.
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Al Viro [Wed, 4 May 2005 04:39:22 +0000 (05:39 +0100)]
[PATCH] ISA DMA Kconfig fixes - part 1
A bunch of drivers use ISA DMA helpers or their equivalents for
platforms that have ISA with different DMA controller (a lot of ARM
boxen). Currently there is no way to put such dependency in Kconfig -
CONFIG_ISA is not it (e.g. it is not set on platforms that have no ISA
slots, but have on-board devices that pretend to be ISA ones).
New symbol added - ISA_DMA_API. Set when we have functional
enable_dma()/set_dma_mode()/etc. set of helpers. Next patches in the
series will add missing dependencies for drivers that need them.
I'm very carefully staying the hell out of the recurring flamefest on
what exactly CONFIG_ISA would mean in ideal world - added symbol has a
well-defined meaning and for now I really want to treat it as completely
independent from the mess around CONFIG_ISA.
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Dely Sy [Fri, 29 Apr 2005 01:08:53 +0000 (18:08 -0700)]
[PATCH] PCI Hotplug: fix pciehp regression
I fogot to remove the code that freed the memory in cleanup_slots().
Here is the new patch, which I have also taken care of the comment
by Eike to remove the cast in hotplug_slot->private.
Signed-off-by: Dely Sy <dely.l.sy@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
[PATCH] PCI Hotplug ibmphp_pci.c: Fix masking out needed information too early
here is the patch that fixes the bug introduced by my previous patch which
already went into 2.6.12-rc2 and is likely to cause trouble is someone hits
one the else case here by accident.
Using the &= operation before the if statement destroys the information the
if asks for so we always go into the else branch.
[PATCH] PCI: fix up word-aligned 16-bit PCI config access through sysfs
This patch adds the possibility to do word-aligned 16-bit atomic PCI
configuration space accesses via the sysfs PCI interface. As a result, problems
with Emulex LFPC on IBM PowerPC64 are fixed.
Pavel Machek [Tue, 5 Apr 2005 21:49:49 +0000 (23:49 +0200)]
[PATCH] PCI: fix stale PCI pm docs
This fixes u32 vs. pm_message_t confusion in documentation, and
removes references to no-longer-existing (*save_state), too. With
exception of USB (I hope David will fix/apply my patch), this should
fix last piece of this confusion... famous last words.
Signed-off-by: Pavel Machek <pavel@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
[PATCH] PCI: 'is_enabled' flag should be set/cleared when the device is actually enabled/disabled
I think 'is_enabled' flag in pci_dev structure should be set/cleared
when the device actually enabled/disabled. Especially about
pci_enable_device(), it can be failed. By this change, we will also
get the possibility of refering 'is_enabled' flag from the functions
called through pci_enable_device()/pci_disable_device().
Lonnie Mendez [Tue, 3 May 2005 22:02:20 +0000 (17:02 -0500)]
[PATCH] USB cypress_m8: update kernel driver with current source
Fixed problem where setting or retreiving the serial config would fail
with EPIPE. Removed CRTS toggling so the driver behaves more like other
usbserial adapters. Issued new interval of 1ms instead of the default
bInterval. As a result, transfer speed has been substantially
increased. From avg. 850bps to avg. 3300bps. Also added new module
parameter 'interval' to tweak the interval in case this change causes
problems for someone. Cleaned up code and formatting issues so source
is more readable. Replaced the C++ style comments. Various other code
cleanups.
Ian Abbott [Fri, 29 Apr 2005 15:06:14 +0000 (16:06 +0100)]
[PATCH] USB: VID/PID updates for ftdi_sio driver
Some VID/PID updates for the ftdi_sio driver:
* The "Gude Analog- und Digitalsysteme GmbH" entries were missing from
the "combined" table.
* Replaced FTDI_8U232AM_ALT_ALT_PID with 3 PIDs for devices from
4n-galaxy.de.
* Removed redundant FTDI_RM_VID and renamed FTDI_RMCANVIEW_PID to
FTDI_RM_CANVIEW_PID.
* Added VID/PID for serial converter in Mobility Electronics EasiDock
USB 200 (mentioned by Gregory Schmitt).
* Added PID for Active Robots USB comms board (mentioned by John Koch).
On Thu, Apr 14, 2005 at 11:06:21PM +0400, Sergey Vlasov wrote:
> http://thread.gmane.org/gmane.linux.usb.devel/32977
>
> (see "[PATCH] N/3 cdc acm errors").
>
> You also need this driver core fix:
>
> http://thread.gmane.org/gmane.linux.usb.devel/33132
I reproduced the same oops while trying to execute at+mode=99, it would
be nice to get these fix merged since I believe it's still needed to
connect the laptop over gprs (something I didn't test yet).
This further patch will allow you to connect via usbnet, Greg could you
apply? Thanks!
Signed-off-by: Andrea Arcangeli <andrea@suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Sat, 9 Apr 2005 16:00:29 +0000 (09:00 -0700)]
[PATCH] USB: ehci power fixes
Miscellaneous updates for EHCI.
- Mostly updates the power switching on EHCI controllers. One routine
centralizes the "power on/off all ports" logic, and the capability to
do that is reported more correctly.
- Courtesy Colin Leroy, a patch to always power up ports after resumes
which didn't keep a USB device suspended. The reset-everything logic
powers down those ports (on some hardware) so something needs to turn
them back on.
- Minor tweaks/bugfixes for the debug port support.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Herbert Xu [Tue, 3 May 2005 23:27:10 +0000 (16:27 -0700)]
[IPSEC]: Store idev entries
I found a bug that stopped IPsec/IPv6 from working. About
a month ago IPv6 started using rt6i_idev->dev on the cached socket dst
entries. If the cached socket dst entry is IPsec, then rt6i_idev will
be NULL.
Since we want to look at the rt6i_idev of the original route in this
case, the easiest fix is to store rt6i_idev in the IPsec dst entry just
as we do for a number of other IPv6 route attributes. Unfortunately
this means that we need some new code to handle the references to
rt6i_idev. That's why this patch is bigger than it would otherwise be.
I've also done the same thing for IPv4 since it is conceivable that
once these idev attributes start getting used for accounting, we
probably need to dereference them for IPv4 IPsec entries too.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
[PKT_SCHED]: netetm: adjust parent qlen when duplicating
Fix qlen underrun when doing duplication with netem. If netem is used
as leaf discipline, then the parent needs to be tweaked when packets
are duplicated.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org> Signed-off-by: David S. Miller <davem@davemloft.net>
[PKT_SCHED]: netetm: make qdisc friendly to outer disciplines
Netem currently dumps packets into the queue when timer expires. This
patch makes work by self-clocking (more like TBF). It fixes a bug
when 0 delay is requested (only doing loss or duplication).
Signed-off-by: Stephen Hemminger <shemminger@osdl.org> Signed-off-by: David S. Miller <davem@davemloft.net>
[PKT_SCHED]: netetm: trap infinite loop hange on qlen underflow
Due to bugs in netem (fixed by later patches), it is possible to get qdisc
qlen to go negative. If this happens the CPU ends up spinning forever
in qdisc_run(). So add a BUG_ON() to trap it.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Some network drivers call netif_stop_queue() when detecting loss of
carrier. This leads to packets being queued up at the qdisc level for
an unbound period of time. In order to prevent this effect, the core
networking stack will now cease to queue packets for any device, that
is operationally down (i.e. the queue is flushed and disabled).
Signed-off-by: Tommy S. Christensen <tommy.christensen@tpack.net> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 3 May 2005 23:15:59 +0000 (16:15 -0700)]
[XFRM/RTNETLINK]: Decrement qlen properly in {xfrm_,rt}netlink_rcv().
If we free up a partially processed packet because it's
skb->len dropped to zero, we need to decrement qlen because
we are dropping out of the top-level loop so it will do
the decrement for us.
Spotted by Herbert Xu.
Signed-off-by: David S. Miller <davem@davemloft.net>
- pfm_context_load(): change return value from EINVAL to EBUSY
when context is already loaded.
- pfm_check_task_state(): pass test if context state is MASKED.
It is safe to give access on PFM_CTX_MASKED because the PMU
state (PMD) is stable and saved in software state.
This helps multiplexing programs such as the example given
in libpfm-3.1.
Signed-off-by: stephane eranian <eranian@hpl.hp.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
The pmu_active test is based on the values of PSR.up. THIS IS THE PROBLEM as
it does not take into account the lazy restore logic which is as follow (simplified):
context switch out:
save PMDs
clear psr.up
release ownership
context switch in:
if (ctx->last_cpu == smp_processor_id() && ctx->cpu_activation == cpu_activation) {
set psr.up
return
}
restore PMD
restore PMC
ctx->last_cpu = smp_processor_id();
ctx->activation = ++cpu_activation;
set psr.up
The key here is that on context switch out, we clear psr.up and on context switch in
we check if nobody else used the PMU on that processor since last time we came. In
that case, we assume the PMD/PMC are ours and we simply reactivate.
The Caliper problem is that between the moment we context switch out and the moment we
come back, nobody effectively used the PMU BUT the processor went idle. Normally this
would have no incidence but PAL_HALT does alter the PMU registers. In default_idle(),
the test on psr.up is not strong enough to cover this case and we go into PAL which
trashed the PMU resgisters. When we come back we falsely assume that this is our state
yet it is corrupted. Very nasty indeed.
To avoid the problem it is necessary to forbid going to PAL_HALT as soon as perfmon
installs some valid state in the PMU registers. This happens with an application
attaches a context to a thread or CPU. It is not enough to check the psr/dcr bits.
Hence I propose the attached patch. It adds a callback in process.c to modify the
condition to enter PAL on idle. Basically, now it is conditional to pal_halt=1 AND
perfmon saying it is okay.
git://bbs.cooldavid.org / net-next-2.6.git / log