- 18 Dec, 2007 40 commits
-
-
Ingo Molnar authored
this is the tale of a full day spent debugging an ancient but elusive bug. after booting up thousands of random .config kernels, i finally happened to generate a .config that produced the following rare bootup failure on 32-bit x86: | ..TIMER: vector=0x31 apic1=0 pin1=2 apic2=-1 pin2=-1 | ..MP-BIOS bug: 8254 timer not connected to IO-APIC | ...trying to set up timer (IRQ0) through the 8259A ... failed. | ...trying to set up timer as Virtual Wire IRQ... failed. | ...trying to set up timer as ExtINT IRQ... failed :(. | Kernel panic - not syncing: IO-APIC + timer doesn't work! Boot with apic=debug | and send a report. Then try booting with the 'noapic' option this bug has been reported many times during the years, but it was never reproduced nor fixed. the bug that i hit was extremely sensitive to .config details. First i did a .config-bisection - suspecting some .config detail. That led to CONFIG_X86_MCE: enabling X86_MCE magically made the bug disappear and the system would boot up just fine. Debugging my way through the MCE code ended up identifying two unlikely candidates: the thing that made a real difference to the hang was that X86_MCE did two printks: Intel machine check architecture supported. Intel machine check reporting enabled on CPU#1. Adding the same printks to a !CONFIG_X86_MCE kernel made the bug go away! this left timing as the main suspect: i experimented with adding various udelay()s to the arch/x86/kernel/io_apic_32.c:check_timer() function, and the race window turned out to be narrower than 30 microseconds (!). That made debugging especially funny, debugging without having printk ability before the bug hits is ... interesting ;-) eventually i started suspecting IRQ activities - those are pretty much the only thing that happen this early during bootup and have the timescale of a few dozen microseconds. Also, check_timer() changes the IRQ hardware in various creative ways, so the main candidate became IRQ0 interaction. i've added a counter to track timer irqs (on which core they arrived, at what exact time, etc.) and found that no timer IRQ would arrive after the bug condition hits - even if we re-enable IRQ0 and re-initialize the i8259A, but that we'd get a small number of timer irqs right around the time when we call the check_timer() function. Eventually i got the following backtrace triggered from debug code in the timer interrupt: ...trying to set up timer as Virtual Wire IRQ... failed. ...trying to set up timer as ExtINT IRQ... Pid: 1, comm: swapper Not tainted (2.6.24-rc5 #57) EIP: 0060:[<c044d57e>] EFLAGS: 00000246 CPU: 0 EIP is at _spin_unlock_irqrestore+0x5/0x1c EAX: c0634178 EBX: 00000000 ECX: c4947d63 EDX: 00000246 ESI: 00000002 EDI: 00010031 EBP: c04e0f2e ESP: f7c41df4 DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068 CR0: 8005003b CR2: ffe04000 CR3: 00630000 CR4: 000006d0 DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000 DR6: ffff0ff0 DR7: 00000400 [<c05f5784>] setup_IO_APIC+0x9c3/0xc5c the spin_unlock() was called from init_8259A(). Wait ... we have an IRQ0 entry while we are in the middle of setting up the local APIC, the i8259A and the PIT?? That is certainly not how it's supposed to work! check_timer() was supposed to be called with irqs turned off - but this eroded away sometime in the past. This code would still work most of the time because this code runs very quickly, but just the right timing conditions are present and IRQ0 hits in this small, ~30 usecs window, timer irqs stop and the system does not boot up. Also, given how early this is during bootup, the hang is very deterministic - but it would only occur on certain machines (and certain configs). The fix was quite simple: disable/restore interrupts properly in this function. With that in place the test-system now boots up just fine. (64-bit x86 io_apic_64.c had the same bug.) Phew! One down, only 1500 other kernel bugs are left ;-) Signed-off-by:
Ingo Molnar <mingo@elte.hu> Signed-off-by:
Thomas Gleixner <tglx@linutronix.de>
-
Steven Rostedt authored
In commit 76d21601 lazy irq disabling was implemented, and the simple irq handler had a masking set to it. Remy Bohmer discovered that some devices in the ARM architecture would trigger the mask, but never unmask it. His patch to do the unmasking was questioned by Russell King about masking simple irqs to begin with. Looking further, it was discovered that the problems Remy was seeing was due to improper use of the simple handler by devices, and he later submitted patches to fix those. But the issue that was uncovered was that the simple handler should never mask. This patch reverts the masking in the simple handler. Signed-off-by:
Steven Rostedt <srostedt@redhat.com> Signed-off-by:
Russell King <rmk+kernel@arm.linux.org.uk>
-
Jan Beulich authored
The patch introducing this left out 64-bit x86 despite it also having extra entries. this solves Xen guest troubles. Signed-off-by:
Jan Beulich <jbeulich@novell.com> Signed-off-by:
-
Masami Hiramatsu authored
Kprobes for x86-64 may cause a kernel crash if it inserted on "iret" instruction. "call absolute" is invalid on x86-64, so we don't need treat it. - Change the processing order as same as x86-32. - Add "iret"(0xcf) case. - Remove next_rip local variable. Signed-off-by:
Masami Hiramatsu <mhiramat@redhat.com> Signed-off-by:
-
Masami Hiramatsu authored
jprobe for x86-64 may cause kernel page fault when the jprobe_return() is called from incorrect function. - Use jprobe_saved_regs instead getting it from stack. (Especially on x86-64, it may get incorrect data, because pt_regs can not be get by using container_of(rsp)) - Change the type of stack pointer to unsigned long *. Signed-off-by:
-
Adrian Bunk authored
This patch fixes the following section mismatches with CONFIG_HOTPLUG=n, CONFIG_HOTPLUG_CPU=y: ... WARNING: vmlinux.o(.text+0x41cd3): Section mismatch: reference to .init.data:tvec_base_done.22610 (between 'timer_cpu_notify' and 'run_timer_softirq') WARNING: vmlinux.o(.text+0x41d67): Section mismatch: reference to .init.data:tvec_base_done.22610 (between 'timer_cpu_notify' and 'run_timer_softirq') ... Signed-off-by:
Adrian Bunk <bunk@kernel.org> Signed-off-by:
-
Kevin Hilman authored
Add unlocked version for use by irq_chip.set_type handlers which may wish to change handler to level or edge handler when IRQ type is changed. The normal set_irq_handler() call cannot be used because it tries to take irq_desc.lock which is already held when the irq_chip.set_type hook is called. Signed-off-by:
Kevin Hilman <khilman@mvista.com> Signed-off-by:
-
Thomas Gleixner authored
Resolve the following regression of a choppy, almost unusable laptop: http://lkml.org/lkml/2007/12/7/299 http://bugzilla.kernel.org/show_bug.cgi?id=9525 A previous version of the code did the reprogramming of the broadcast device in the return from idle code. This was removed, but the logic in tick_handle_oneshot_broadcast() was kept the same. When a broadcast interrupt happens we signal the expiry to all CPUs which have an expired event. If none of the CPUs has an expired event, which can happen in dyntick mode, then we reprogram the broadcast device. We do not reprogram otherwise, but this is only correct if all CPUs, which are in the idle broadcast state have been woken up. The code ignores, that there might be pending not yet expired events on other CPUs, which are in the idle broadcast state. So the delivery of those events can be delayed for quite a time. Change the tick_handle_oneshot_broadcast() function to check for CPUs, which are in broadcast state and are not woken up by the current event, and enforce the rearming of the broadcast device for those CPUs. Signed-off-by:
-
Barry Kasindorf authored
This patch is for controlling the upper 32bits of the event ctrl msrs. This includes the upper 4 bits of the event select and the Guest Only and Host Only bits This patch is necessary to make Event Based Profiling work reliably on a Family 10h processor [akpm@linux-foundation.org: checkpatch.pl fixes] Signed-off-by:
Barry Kasindorf <barry.kasindorf@amd.com> Signed-off-by:
Robert Richter <robert.richter@amd.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/netdev-2.6Linus Torvalds authored
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/netdev-2.6: (23 commits) iwlwifi: fix rf_kill state inconsistent during suspend and resume b43: Fix rfkill radio LED bcm43xx_debugfs sscanf fix libertas: select WIRELESS_EXT iwlwifi3945/4965: fix rate control algo reference leak ieee80211_rate: missed unlock wireless/ipw2200.c: add __dev{init,exit} annotations zd1211rw: Fix alignment problems libertas: add Dan Williams as maintainer sis190 endianness ucc_geth: really fix section mismatch pcnet_cs: add new id ixgb: make sure jumbos stay enabled after reset Net: ibm_newemac, remove SPIN_LOCK_UNLOCKED net: smc911x: shut up compiler warnings ucc_geth: minor whitespace fix drivers/net/s2io.c section fixes drivers/net/sis190.c section fix hamachi endianness fixes e100: free IRQ to remove warningwhenrebooting ...
-
git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-devLinus Torvalds authored
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev: libata: fix ATAPI draining libata: update atapi_eh_request_sense() such that lbam/lbah contains buffer size libata-acpi: implement _GTF command filtering libata-acpi: improve _GTF execution error handling and reporting libata-acpi: improve ACPI disabling libata-acpi: implement dev->gtf_cache and evaluate _GTF right after _STM during resume libata-acpi: implement and use ata_acpi_init_gtm() libata-acpi: add new hooks ata_acpi_dissociate() and ata_acpi_on_disable() libata: ata_dev_disable() should be called from EH context libata: add more opcodes to ata.h libata: update ata_*_printk() macros such that level can be a variable libata-acpi: adjust constness in ata_acpi_gtm/stm() parameters sata_mv: improve warnings about Highpoint RocketRAID 23xx cards libata: add ST3160023AS / 3.42 to NCQ blacklist libata: clear link->eh_info.serror from ata_std_postreset() sata_sil: fix spurious IRQ handling
-
Eric W. Biederman authored
Fix: sysctl table check failed: /net/ax25/ax0/ax25_default_mode .3.9.1.2 Unknown sysctl binary path Pid: 2936, comm: kissattach Not tainted 2.6.24-rc5 #1 [<c012ca6a>] set_fail+0x3b/0x43 [<c012ce7a>] sysctl_check_table+0x408/0x456 [<c012ce8e>] sysctl_check_table+0x41c/0x456 [<c012ce8e>] sysctl_check_table+0x41c/0x456 ... Signed-off-by:
Eric W. Biederman <ebiederm@xmission.com> Cc: Bernard Pidoux <pidoux@ccr.jussieu.fr> Cc: "David S. Miller" <davem@davemloft.net> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Christoph Lameter authored
This ensures that the quicklists are drained. Otherwise draining may only occur when the processor reaches an idle state. Fixes fatal leakage of pgd_t's on 2.6.22 and later. Signed-off-by:
Christoph Lameter <clameter@sgi.com> Reported-by:
Dhaval Giani <dhaval@linux.vnet.ibm.com> Cc: <stable@kernel.org> Signed-off-by:
-
Christoph Lameter authored
Remove a recently added useless masking of GFP_ZERO. GFP_ZERO is already masked out in new_slab() (See how it calls allocate_slab). No need to do it twice. This reverts the SLUB parts of 7fd27255. Cc: Matt Mackall <mpm@selenic.com> Reviewed-by:
Pekka Enberg <penberg@cs.helsinki.fi> Signed-off-by:
-
Jan Kara authored
Fix compilation warning about discarded const. Signed-off-by:
Jan Kara <jack@suse.cz> Signed-off-by:
-
Nishanth Aravamudan authored
The hugetlb documentation has gotten a bit out of sync with the current code. Updated the sysctl file to refer to Documentation/vm/hugetlbpage.txt. Update that file to contain the current state of affairs (with the newer named sysctl in place). Signed-off-by:
Nishanth Aravamudan <nacc@us.ibm.com> Acked-by:
Adam Litke <agl@us.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Cc: Dave Hansen <haveblue@us.ibm.com> Cc: David Gibson <david@gibson.dropbear.id.au> Signed-off-by:
-
Nishanth Aravamudan authored
This reverts commit 54f9f80d ("hugetlb: Add hugetlb_dynamic_pool sysctl") Given the new sysctl nr_overcommit_hugepages, the boolean dynamic pool sysctl is not needed, as its semantics can be expressed by 0 in the overcommit sysctl (no dynamic pool) and non-0 in the overcommit sysctl (pool enabled). (Needed in 2.6.24 since it reverts a post-2.6.23 userspace-visible change) Signed-off-by:
-
Nishanth Aravamudan authored
hugetlb: introduce nr_overcommit_hugepages sysctl While examining the code to support /proc/sys/vm/hugetlb_dynamic_pool, I became convinced that having a boolean sysctl was insufficient: 1) To support per-node control of hugepages, I have previously submitted patches to add a sysfs attribute related to nr_hugepages. However, with a boolean global value and per-mount quota enforcement constraining the dynamic pool, adding corresponding control of the dynamic pool on a per-node basis seems inconsistent to me. 2) Administration of the hugetlb dynamic pool with multiple hugetlbfs mount points is, arguably, more arduous than it needs to be. Each quota would need to be set separately, and the sum would need to be monitored. To ease the administration, and to help make the way for per-node control of the static & dynamic hugepage pool, I added a separate sysctl, nr_overcommit_hugepages. This value serves as a high watermark for the overall hugepage pool, while nr_hugepages serves as a low watermark. The boolean sysctl can then be removed, as the condition nr_overcommit_hugepages > 0 indicates the same administrative setting as hugetlb_dynamic_pool == 1 Quotas still serve as local enforcement of the size of the pool on a per-mount basis. A few caveats: 1) There is a race whereby the global surplus huge page counter is incremented before a hugepage has allocated. Another process could then try grow the pool, and fail to convert a surplus huge page to a normal huge page and instead allocate a fresh huge page. I believe this is benign, as no memory is leaked (the actual pages are still tracked correctly) and the counters won't go out of sync. 2) Shrinking the static pool while a surplus is in effect will allow the number of surplus huge pages to exceed the overcommit value. As long as this condition holds, however, no more surplus huge pages will be allowed on the system until one of the two sysctls are increased sufficiently, or the surplus huge pages go out of use and are freed. Successfully tested on x86_64 with the current libhugetlbfs snapshot, modified to use the new sysctl. Signed-off-by:
-
Eric Sandeen authored
ecryptfs in 2.6.24-rc3 wasn't surviving fsx for me at all, dying after 4 ops. Generally, encountering problems with stale data and improperly zeroed pages. An extending truncate + write for example would expose stale data. With the changes below I got to a million ops and beyond with all mmap ops disabled - mmap still needs work. (A version of this patch on a RHEL5 kernel ran for over 110 million fsx ops) I added a few comments as well, to the best of my understanding as I read through the code. Signed-off-by:
Eric Sandeen <sandeen@redhat.com> Acked-by:
Michael Halcrow <mhalcrow@us.ibm.com> Signed-off-by:
-
Nathan Lynch authored
bloat-o-meter assumes that a '.' anywhere in a symbol's name means that it is static and prepends 'static.' to the first part of the symbol name, discarding the portion of the name that follows the '.'. However, the names of function entry points begin with '.' in the ppc64 ABI. This causes all function text size changes to be accounted to a single 'static.' entry in the output when comparing ppc64 kernels. Change getsizes() to ignore the first character of the symbol name when searching for '.'. Signed-off-by:
Nathan Lynch <ntl@pobox.com> Cc: Matt Mackall <mpm@selenic.com> Signed-off-by:
-
Shannon Nelson authored
We can't use the device in a dev_err() after a kzalloc failure or after the kfree, so simplify it to the pdev that was originally passed in. Cc: Eric Sesterhenn <snakebyte@gmx.de> Signed-off-by:
Shannon Nelson <shannon.nelson@intel.com> Signed-off-by:
-
Shannon Nelson authored
A few fixups from Andrew's code comments. - removed "static inline" forward-declares - changed use of min() to min_t() - removed some unnecessary NULL initializations - removed a couple of BUG() calls Fixes this: drivers/dma/ioat_dma.c: In function `ioat1_tx_submit': drivers/dma/ioat_dma.c:177: sorry, unimplemented: inlining failed in call to '__ioat1_dma_memcpy_issue_pending': function body not available drivers/dma/ioat_dma.c:268: sorry, unimplemented: called from here Signed-off-by:
-
Eric Sandeen authored
eCryptfs wasn't setting s_blocksize in it's superblock; just pick it up from the lower FS. Having an s_blocksize of 0 made things like "filefrag" which call FIGETBSZ unhappy. Signed-off-by:
-
Mel Gorman authored
In some cases the IO subsystem is able to merge requests if the pages are adjacent in physical memory. This was achieved in the allocator by having expand() return pages in physically contiguous order in situations were a large buddy was split. However, list-based anti-fragmentation changed the order pages were returned in to avoid searching in buffered_rmqueue() for a page of the appropriate migrate type. This patch restores behaviour of rmqueue_bulk() preserving the physical order of pages returned by the allocator without incurring increased search costs for anti-fragmentation. Signed-off-by:
Mel Gorman <mel@csn.ul.ie> Cc: James Bottomley <James.Bottomley@steeleye.com> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Mark Lord <mlord@pobox.com Signed-off-by:
-
Adam Jackson authored
These types define the size of data read from /dev/apm_bios. They should not be hidden behind #ifdef __KERNEL__. This is killing my xserver compile, apm_event_t is used in the xserver source. Signed-off-by:
-
Adrian Bunk authored
cpufreq_stats_free_table() mustn't be __cpuexit since it's called by the __cpuinit cpufreq_stat_cpu_callback(). This patch fixes the following section mismatch reported by Chris Clayton: WARNING: vmlinux.o(.init.text+0x143dd): Section mismatch: reference to .exit.text:cpufreq_stats_free_table (between 'cpufreq_stat_cpu_callback' and 'cpufreq_stats_init') Signed-off-by:
Dave Jones <davej@codemonkey.org.uk> Cc: Russell King <rmk@arm.linux.org.uk> Signed-off-by:
-
Julia Lawall authored
The error handling code should undo the ioremap as well. The problem was detected using the following semantic match (http://www.emn.fr/x-info/coccinelle/) // <smpl> @@ type T,T1,T2; identifier E; statement S; expression x1,x2; constant C; int ret; @@ T E; ... * E = ioremap(...); if (E == NULL) S ... when != iounmap(E) when != if (E != NULL) { ... iounmap(E); ...} when != x1 = (T1)E if (...) { ... when != iounmap(E) when != if (E != NULL) { ... iounmap(E); ...} when != x2 = (T2)E ( * return; | * return C; | * return ret; ) } // </smpl> Signed-off-by:
Julia Lawall <julia@diku.dk> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Olaf Hering <olaf@aepfle.de> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Signed-off-by:
-
Dave Young authored
In kobject_register, the kobject reference is get in kobject_init, and then kobject_add. If kobject_add fail, it will only cleanup the reference got by itself. Signed-off-by:
Dave Young <hidave.darkstar@gmail.com> Reviewed-by:
-
WANG Cong authored
Improve the error handling for mm/sparse.c::sparse_add_one_section(). And I see no reason to check 'usemap' until holding the 'pgdat_resize_lock'. [geoffrey.levand@am.sony.com: sparse_index_init() returns -EEXIST] Cc: Christoph Lameter <clameter@sgi.com> Acked-by:
Dave Hansen <haveblue@us.ibm.com> Cc: Rik van Riel <riel@redhat.com> Acked-by:
Yasunori Goto <y-goto@jp.fujitsu.com> Cc: Andy Whitcroft <apw@shadowen.org> Signed-off-by:
WANG Cong <xiyou.wangcong@gmail.com> Signed-off-by:
Geoff Levand <geoffrey.levand@am.sony.com> Signed-off-by:
-
WANG Cong authored
Since sparse_index_alloc() can return NULL on memory allocation failure, we must deal with the failure condition when calling it. Signed-off-by:
-
Dave Jones authored
Ingo hit some BUG_ONs that were probably caused by these missing unlocks causing an unbalance. He couldn't reproduce the bug reliably, so it's unknown that it's definitly fixing the problem he hit, but it's a fairly good chance, and this fixes an obvious bug. [ Dave: "Ingo followed up that he hit some lockdep related output with this applied, so it may not be right. I'll look at it after xmas if no-one has it figured out before then." Akpm: "It looks pretty correct to me though." ] Signed-off-by:Dave Jones <davej@redhat.com> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by:
-
Ivan Kokshaysky authored
This fixes some of the alpha-specific build problems, except a) modpost warning about COMMON symbol "saved_config" and b) nasty final link failure with gcc-4.x, -Os and scsi-disk driver configured built-in (due to jump table in .rodata referencing discarded .exit.text). - build failure with gcc-4.2.x: fix up casts in cia_io* routines to avoid warnings ('discards qualifiers from pointer target type'), which are failures, thanks to -Werror; - modpost warnings: add missing __init qualifier for titan and marvel; for non-generic build, move machine vectors from .data to .data.init.refok section; - unbreak CPU-specific optimization: rearrange cpuflags-y assignments so that extended -mcpu value (ev56, pca56, ev67) overrides basic one (ev5, ev6) and not vice versa. Signed-off-by:Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Richard Henderson <rth@twiddle.net> Signed-off-by:
-
Andries E. Brouwer authored
As it turns out, the kernel divides by EXT3_INODES_PER_GROUP(s) when mounting an ext3 filesystem. If that number is zero, a crash follows. Below a patch. This crash was reported by Joeri de Ruiter, Carst Tankink and Pim Vullers. Cc: <linux-ext4@vger.kernel.org> Acked-by:
Alan Cox <alan@redhat.com> Signed-off-by:
-
Uwe Kleine-König authored
This was introduced in 4af8e944Signed-off-by:
Uwe Kleine-König <Uwe.Kleine-Koenig@digi.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by:
-
Geoff Levand authored
SPARSEMEM_VMEMMAP needs to be a selectable config option to support building the kernel both with and without sparsemem vmemmap support. This selection is desirable for platforms which could be configured one way for platform specific builds and the other for multi-platform builds. Signed-off-by:
Miguel Botón <mboton@gmail.com> Signed-off-by:
-
Sheela authored
Share net is not supported, Rusty is an "idiot" . Signed-off-by:
Sheela Sequeira <sheela.sequeira@gmail.com> Reviewed-by:
James Morris <jmorris@namei.org> Acked-by:
Rusty Russell <rusty@rustcorp.com.au> Signed-off-by:
-
Eric Sandeen authored
ecryptfs_destroy_mount_crypt_stat() checks whether each auth_tok->global_auth_tok_key is nonzero and if so puts that key. However, in some early mount error paths nothing has initialized the pointer, and we try to key_put() garbage. Running the bad cipher tests in the testsuite exposes this, and it's happy with the following change. Signed-off-by:
-
Eric Dumazet authored
While auditing proc_doulongvec_ms_jiffies_minmax() usage in kernel, I found a bug in drivers/parport/procfs.c, incorrectly using sizeof(int) instead of sizeof(unsigned long) Only 64bit arches are affected by this old bug. Signed-off-by:
Eric Dumazet <dada1@cosmosbay.com> Signed-off-by:
-
David Brownell authored
Reorder at32_rtc_probe() so that it's safe (no oopsing) to fire the IRQ handler the instant that it's registered. (Bug noted via "Debug shared IRQ handlers" kernel debug option.) Signed-off-by:
David Brownell <dbrownell@users.sourceforge.net> Cc: Alessandro Zummo <a.zummo@towertech.it> Cc: <hcegtvedt@atmel.com> Signed-off-by:
-
Ivan Kokshaysky authored
First of all, thanks to Bob Tracy <rct@frus.com> and Michael Cree <mcree@orcon.net.nz> for testing. Especially to Bob, as he has done titanic multi-day git-bisect work that finally helped to reproduce and nail down the bug (http://bugzilla.kernel.org/show_bug.cgi?id=9457). [ev6-]stxncpy.S: it's t12, not t2 register that is supposed to contain the last byte offset upon return. As a result of wrong register use (which was my fault back in 2003, IIRC), under some circumstances extra terminating zero bytes were added to destination string. This particularly led to incorrect DEVPATH strings generated in uevent and therefore to udev problems. strncpy.S: unrelated bug I found while testing the above fix - destination is not properly zero-padded then a byte count exceeds source length. Actually this is addition to strncpy fix from last year. Signed-off-by:
-
