- 08 Mar, 2005 40 commits
-
-
Jan Blunck authored
d_drop() must use the dentry->d_lock spinlock. In some cases __d_drop() was used without holding the dentry->d_lock spinlock, too. This could end in a race with __d_lookup(). Signed-off-by:
Jan Blunck <j.blunck@tu-harburg.de> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
David Howells authored
The attached patch fixes a race between kallsyms and insmod/rmmod. The problem is this: (1) The various kallsyms functions poke around in the module list without any locking so that they can be called from the oops handler. (2) Although insmod and rmmod use locks to exclude each other, these have no effect on the kallsyms function. (3) Although rmmod modifies the module state with the machine "stopped", it hasn't removed the metadata from the module metadata list, meaning that as soon as the machine is "restarted", the metadata can be observed by kallsyms. It's not possible to say that an item in that list should be ignored if it's state is marked as inactive - you can't get at the state information because you can't trust the metadata in which it is embedded. Furthermore, list linkage information is embedded in the metadata too, so you can't trust that either... (4) kallsyms may be walking the module list without a lock whilst either insmod or rmmod are busy changing it. insmod probably isn't a problem since nothing is going a way, but rmmod is as it's deleting an entry. (5) Therefore nothing that uses these functions can in any way trust any pointers to "static" data (such as module symbol names or module names) that are returned. (6) On ppc64 the problems are exacerbated since the hypervisor may reschedule bits of the kernel, making operations that appear adjacent occur a long time apart. This patch fixes the race by only linking/unlinking modules into/from the master module list with the machine in the "stopped" state. This means that any "static" information can be trusted as far as the next kernel reschedule on any given CPU without the need to hold any locks. However, I'm not sure how this is affected by preemption. I suspect more work may need to be done in that case, but I'm not entirely sure. This also means that rmmod has to bump the machine into the stopped state twice... but since that shouldn't be a common operation, I don't think that's a problem. I've amended this patch to not get spinlocks whilst in the machine locked state - there's no point as nothing else can be holding spinlocks. Signed-Off-By:David Howells <dhowells@redhat.com> Signed-off-by:
-
Liam Girdwood authored
This is a resend of a patch that has been applied to 2.4. The low power codec functionality has also now been included in ALSA. It checks the codec ID before doing an AC97 register reset. This allows the kernel to support low power codecs that are powered down by a reset command. This patch also fixes some other minor issues. Changes:- - Added AC97_DEFAULT_POWER_OFF to ac97_codec_ids[] - ac97_probe now checks hardwired codec ID's before sending a reset - Added support for WM9713 - Moved the codec specific inits after the mixer setup as some init - tings were being clobbered. - Added extra check so that default_digital_ops doesn't overwrite a valid codec_ops. (SPDIF) Signed-off-by:
Liam Girdwood <liam.girdwood@wolfsonmicro.com> Signed-off-by:
-
Corey Minyard authored
This patch fixes a race between the CMOS clock setting and the NMI code. The NMI code indiscriminatly sets index registers and values in the same place the CMOS clock is set. If you are setting the CMOS clock and an NMI occurs, Bad values could be written to or read from the CMOS RAM, or the NMI operation might not occur correctly. Fixing this requires creating a special lock so the NMI code can know its CPU owns the lock an "do the right thing" in that case. This was discovered and the fix has been tested by a very demanding customer who tests the heck of out the software we deliver. Signed-off-by:
Corey Minyard <minyard@acm.org> Signed-off-by:
-
Prasanna Meda authored
Reorder test_root testing from 3,5,7 to 7,5,3 so that average case becomes good. Even number check is added. Signed-off-by:
Prasanna Meda <pmeda@akamai.com> Signed-off-by:
-
Matthew Wilcox authored
If we chose to "do_more", we would double-count the amount freed in the second and subsequent block groups. Fix it the same way as was done in ext2 a couple of years ago. Signed-off-by:
Matthew Wilcox <matthew@wil.cx> Signed-off-by:
-
Matthew Wilcox authored
Rename variables in ext3/balloc.c as was done in ext2 a couple of years ago. Signed-off-by:
-
Matthew Wilcox authored
journal_commit_transaction() is 720 lines long. This patch pulls about 55 of them out into their own function, removes a goto and cleans up the control flow a little. Signed-off-by:
-
Alex Tomas authored
during truncate ext3 calls journal_forget() for freed blocks, but before these blocks go to the transaction and jbd reserves space in log for them (->t_outstanding_credits). also, journal_forget() removes these blocks from the transaction, but doesn't correct log space reservation. for example, removal of 500MB file reserves 136 blocks, but only 10 blocks go to the log. a commit is expensive and correct reservation allows us to avoid needless commits. here is the patch. tested on UP. Signed-off-by:
Alex Tomas <alex@clusterfs.com> Signed-off-by:
-
Alex Tomas authored
Dynamically allocate the holding array for kjournald write patching rather than allocating it on the stack. Signed-off-by:
-
Alex Tomas authored
fix against credits leak in journal_release_buffer() The idea is to charge a buffer in journal_dirty_metadata(), not in journal_get_*_access()). Each buffer has flag call journal_dirty_metadata() sets on the buffer. Signed-off-by:
-
Roman Zippel authored
The vt_struct and vc_data are always allocated together, so there is no need for a separate vt_struct structure. Signed-off-by:
Roman Zippel <zippel@linux-m68k.org> Signed-off-by:
-
Roman Zippel authored
Remove the macros in console_macros.h and so make the structure references explicit. Signed-off-by:
-
Roman Zippel authored
This removes as far as possible unneccessary vc_cons lookups by using a pointer to the vc_data structure instead of the index. The hidden currcons argument in console_macros.h is temporarily replaced with a hidden vc pointer. Signed-off-by:
-
Peter Tiedemann authored
ctc network driver changes: - Properly initialize ccw array. This fixes the stray oopses after an online/offline cycle. - Correct check for already existing channel. - Add missing kfrees. Signed-off-by:
Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by:
-
Ursula Braun-Krahl authored
With Frank Pavlic <pavlic@de.ibm.com> qeth network driver changes: - Using layer 2 mode IPv6 multicast addresses has not been registered at the OSA card (??? FIXME) - We have to allow QETH_IP_THREAD in qeth_set_online too when running in layer 2 mode otherwise multicast addresses won't be registered at the OSA card. - fake_ll improvements for dhcpcd. - Strip ethernet header of VLAN packets before calling skb_pull. This fixes multicast VLAN traffic stalls. - Fix unused variable warning. Signed-off-by:
Jeff Garzik <jgarzik@pobox.com> Signed-off-by:
-
Ursula Braun-Krahl authored
iucv changes: - Initialize iucv with subsys_initcall to make sure that it is there before either vmlogrdr or netiucv start using it. Signed-off-by:
-
Eric Rossman authored
z90crypt device driver changes: - Correct the condition for which the reader task is scheduled to run. Signed-off-by:
-
Cornelia Huck authored
Common i/o layer changes: - If a device driver tries to start I/O while the common I/O layer wants to start path verification, don't reject the I/O with -EBUSY (which might prompt the driver to retry the next tick) but seemingly accept the I/O and deliver a fake irb with deferred cc 1 after path verification has finished (prompting the driver to retry the I/O). This prevents the device driver from doing useless retries while cio is still busy with path verification. Signed-off-by:
-
Cornelia Huck authored
Common i/o layer changes: - Update scsw information before checking activity control bits in the offline processing. - Clear irb structure after cio initiated I/O completed. - Modify cdev private structure only while holding the lock. - Update scsw information before checking whether we can start path verification. - Only generate a notoper event if the device is not already in the not operation state, otherwise we end up with two unregister calls. Signed-off-by:
-
Martin Schwidefsky authored
Add key management system calls. Signed-off-by:
-
Martin Schwidefsky authored
Make s390 compile and work with gcc4. Signed-off-by:
-
Martin Schwidefsky authored
s390 core changes: - Add -msoft-float to CFLAGS. - Remove experimantal tag from cpu hotplug. - Allow more than 4GB swap on a single device for 64 bit. - Fix race in machine_restart to make sure all cpus entered stopped state before reipl. - Cleanup: use for_each_online_cpu macro where possible. - Add argument brackets to __FD_SET/__FD_CLEAR/__FD_ZERO. - Reset cpu_present in smp startup to avoid long delays if only one cpu is defined. - Regenerate default configuration. Signed-off-by:
-
Paolo \'Blaisorblade\' Giarrusso authored
That var is used only once, use its value directly. Signed-off-by:
Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by:
-
Daniel Dickman authored
Signed-off-by:
Daniel Dickman <didickman@yahoo.com> Signed-off-by:
-
Barry K. Nathan authored
On i386, SOFTWARE_SUSPEND requires the CPU to have PSE support, but DEBUG_PAGEALLOC disables PSE. Thus, allowing both options to be enabled simultaneously makes no sense. This patch disables DEBUG_PAGEALLOC if SOFTWARE_SUSPEND is enabled; it also displays a comment to briefly explain why DEBUG_PAGEALLOC is missing in that case. I have tested this patch against oldconfig and menuconfig on 2.6.11-bk2. Signed-off-by:
Barry K. Nathan <barryn@pobox.com> Acked-by:
Pavel Machek <pavel@ucw.cz> Signed-off-by:
-
Pavel Machek authored
swsusp low-level assembly requires PSE (4mb pages for kernel) and DEBUG_PAGEALLOC disables that capability. Tell people what went wrong, some people seen same problem on VIA cpus... Signed-off-by:
Pavel Machek <pavel@suse.cz> Signed-off-by:
-
Pavel Machek authored
We got quite long list of machines and tricks needed to get them working. Signed-off-by:
-
Pavel Machek authored
This is patch from Rafael, it eliminates order-5 (or worse) allocations during suspend. I did few style/whitespace modifications. It was tested by me, Rafael, and Stefan from SuSE. Signed-off-by:
Rafael J. Wysocki <rjw@sisk.pl> Signed-off-by:
Adrian Bunk <bunk@stusta.de> Signed-off-by:
-
Pat Gefre authored
Signed-off-by:
Patrick Gefre <pfg@sgi.com> Signed-off-by:
-
Pat Gefre authored
2.6 Altix console patch to ignore input during early booting Signed-off-by:
-
Andries E. Brouwer authored
There are several cases where __init function pointers are stored in a general purpose struct. For example, a SCSI template may contain a __init detect function. Have not yet thought of an elegant way to avoid this. One such case is the mtrr code, where struct mtrr_ops has an init field pointing at __init functions. Unless I overlook something, this case may be easy to settle, since the .init field is never used. The patch below comments out the declaration and initialisation of the .init field of struct mtrr_ops, and puts #if 0 ... #endif around the centaur_mcr_init() and cyrix_arr_init() code. Simultaneously a number of variables are made static. Signed-off-by:
-
Christoph Lameter authored
This patch removes the check for the existence of multiple HPET timers. It allows the use of HPET with only a single timer for system time if HPET_EMULATE_RTC is not set. Signed-off-by:
Christoph Lameter <christoph@lameter.com> Signed-off-by:
Shai Fultheim <Shai@Scalex86.org> Acked-by:
"Pallipadi, Venkatesh" <venkatesh.pallipadi@intel.com> Signed-off-by:
-
Lee Revell authored
Clean up the logic in do_timer_interrupt() Signed-off-by:
-
Rik van Riel authored
Because Xen is compiled with -Wall -Werror, has inherited processor.h from Linux and Fedora is now built with gcc4, I discovered this bug. The few callers I verified all call cpuid with unsigned ints, but the function is defined with signed ints. This trivial patch fixes that. Signed-off-by:
Rik van Riel <riel@redhat.com> Signed-off-by:
-
Thomas Gleixner authored
Convert the initializers of hw_interrupt_type structures to C99 initializers. Signed-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
Dave Hansen authored
When CONFIG_HIGHMEM=y, but ZONE_NORMAL isn't quite full, there is, of course, no actual memory at *high_memory. This isn't a problem with normal virt<->phys translations because it's never dereferenced, but CONFIG_NONLINEAR is a bit more finicky. So, don't do __va() in non-existent addresses. Signed-off-by:
Dave Hansen <haveblue@us.ibm.com> Signed-off-by:
-
kianusch@sk-tech.net authored
Those CPU's are found mostly in embedded systems ... one of the most prominent Hardware using GEODE CPU is probably soekris net4801 (http://www.soekris.com). This patch has been on my homepage (http://www.sk-tech.net/support/soekris.html) for quite a time - but I've been asked several time to have it included in the main kernel. Signed-off-by:
-
Andrew Morton authored
ioremap() has special-case handling for the IS region, but inunmap() does not. So iounmap() generates a warning when a caller correctly performs an ioremap()/iounmap() sequence. Fix that by teaching iounmap() about the IS address range. Signed-off-by:
-
Henrik Brix Andersen authored
The current SCx200 drivers use a fixed base address of 0x9000 for the Configuration Block, but some systems (at least the Soekris net4801) uses a base address of 0x6000. This patch first tries the fixed address then - if no configuration block could be found - tries the address written to the Configuration Block Address Scratchpad register by the BIOS. Signed-off-by:
Henrik Brix Andersen <brix@gentoo.org> Signed-off-by:
-
