Skip to contentSkip to conversation

2.6.16-rc1-mm2

PATCH In review Andrew Morton 23 participants 45 messages Jan 20, 2006 – Feb 24, 2006 1 message not in the archive

Subsystems (7)
lore
Share & tools
lore
#0 lore raw

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.16-rc1/2.6.16-rc1-mm2/

- This kernel has a big ACPI update

- reiser3 should be safe(r) to use.

Known problems:

- You'll probably see something like this

Memory: 4017084k/6291456k available (2896k kernel code, 176452k reserved, 1868k data, 208k init)

BUG: sleeping function called from invalid context at kernel/mutex.c:84         in_atomic():1, irqs_disabled():0
Call Trace: <ffffffff8012374e>{__might_sleep+177} <ffffffff803cd7da>{mutex_lock+26}

<ffffffff8016b533>{kmem_cache_create+161} <ffffffff8063b0fb>{free_all_boo

in early boot. Please ignore.

- drivers/i2c/busses/scx200_acb.c doesn't compile on architectures which don't have asm/msr.h.

Boilerplate:

- -mm kernel commit activity can be reviewed by subscribing to the mm-commits mailing list.

echo "subscribe mm-commits" | mail [email protected]

- If you hit a bug in -mm and it's not obvious which patch caused it, it is most valuable if you can perform a bisection search to identify which patch introduced the bug. Instructions for this process are at

http://www.zip.com.au/~akpm/linux/patches/stuff/bisecting-mm-trees.txt

But beware that this process takes some time (around ten rebuilds and reboots), so consider reporting the bug first and if we cannot immediately identify the faulty patch, then perform the bisection search.

- When reporting bugs, please try to Cc: the relevant maintainer and mailing list on any email.

Changes since 2.6.16-rc1-mm1:

linus.patch
git-acpi.patch
git-agpgart.patch
git-alsa.patch
git-audit.patch
git-blktrace.patch
git-block.patch
git-cfq.patch
git-cpufreq.patch
git-ia64.patch
git-infiniband.patch
git-kbuild.patch
git-libata-all.patch
git-netdev-all.patch
git-ntfs.patch
git-ocfs2.patch
git-powerpc.patch
git-sym2.patch
git-pcmcia.patch
git-sas-jg.patch
git-watchdog.patch

External trees

-x86_64-fix-mce-exception-stack-for-boot-cpu.patch -scsi_transport_spi-build-fix.patch -synclink_gt-fix-size-of-register-value-storage.patch -dsp_spos_scb_lib-assignment-fix.patch -sem2mutex-drivers-hwmon.patch
-sem2mutex-i2c-2.patch -sem2mutex-arch-ia64-ia32-sys_ia32c.patch -sem2mutex-arch-ia64-kernel-perfmonc.patch -ia64-eliminate-softlockup-warning.patch -RT_CACHE_STAT_INC-warning-fix.patch
-net-fix-1.patch -gregkh-pci-pci-msi-vector-targeting-abstractions-fix.patch -sem2mutex-drivers-pcmcia.patch
-sem2mutex-drivers-usb.patch -usb-iomega-umini-is-unusual.patch
-net2280-warning-fix.patch -auerswald-support-more-tk-devices.patch -libusual-fix-warning-on-64bit-boxes.patch -mm-dirty_exceeded-speedup.patch -mm-dirty_exceeded-speedup-fix.patch -mm-migration-page-refcounting-fix.patch -mm-migration-page-refcounting-fix-warning-fix.patch -mm-migration-page-refcounting-fix-warning-fix-2.patch -mm-migration-page-refcounting-fix-2.patch -simplify-migrate_page_add.patch -zone-reclaim-resurrect-may_swap.patch -zone-reclaim-reclaim-logic.patch -zone-reclaim-reclaim-logic-tidy.patch -zone-reclaim-reclaim-logic-tweaks.patch -zone-reclaim-proc-override.patch
-sem2mutex-mm-slabc.patch -numa-policies-in-the-slab-allocator-v2.patch -mm-optimize-numa-policy-handling-in-slab-allocator.patch -uml-add-__raw_writel-definition.patch -uml-move-ldt-creation.patch -uml-move-libc-dependent-utility-procedures.patch -uml-move-libc-dependent-time-code.patch -uml-change-interface-to-boot_timer_handler.patch -uml-move-headers-to-arch-um-include.patch -uml-move-libc-dependent-skas-memory-mapping-code.patch -uml-move-libc-dependent-skas-process-handling.patch -uml-eliminate-some-globals.patch -uml-implement-soft-interrupts.patch -uml-use-setjmp-longjmp-instead-of-sigsetjmp-siglongjmp.patch -uml-tt-mode-softint-fixes.patch -uml-remove-leftover-from-patch-revertal.patch -uml-make-daemon-transport-behave-properly.patch -uml-networking-clear-transport-specific-structure.patch -uml-fix-spinlock-recursion-and-sleep-inside-spinlock-in-error-path.patch -uml-sigio-code-reduce-spinlock-hold-time.patch -uml-avoid-malloc-to-sleep-in-atomic-sections.patch -uml-arch-kconfig-menu-cleanups.patch -uml-allow-again-to-move-backing-file-and-to-override-saved-location.patch -uml-ubd-code-fix-a-bit-of-whitespace.patch -prevent-trident-driver-from-grabbing-pcnet32-hardware.patch -sem2mutex-drivers-macintosh-windfarm_corec.patch -elevator=as-back-compatibility.patch -v9fs-add-readpage-support.patch -fix-sched_setscheduler-semantics.patch -add-missing-syscall-declarations.patch -hfs-cleanup-hfsplus-prints.patch -hfs-cleanup-hfs-prints.patch
-hfs-add-hfsx-support.patch
-hfs-set-correct-ctime.patch -hfs-set-correct-create-date-for-links.patch -hfs-set-type-creator-for-symlinks.patch -edac-atomic-scrub-operations.patch -edac-drivers-for-amd-76x-and-intel-e750x-e752x.patch -edac-drivers-for-intel-i82860-i82875.patch -edac-drivers-for-radisys-82600.patch -edac-core-edac-support-code.patch -edac-core-edac-support-code-fix.patch -edac-with-sysfs-interface-added.patch -edac-with-sysfs-interface-added-tidy.patch -edac-swsusp-fixes.patch -edac-change-default-also-handle-pulled-hardware.patch -nfsd-check-error-status-from-nfsd_sync_dir.patch -nfsd-remove-inline-from-a-couple-of-large-nfs-functions.patch -svcrpc-save-and-restore-the-daddr-field-when-request-deferred.patch -nfsd4-misc-lock-fixes.patch -nfsd4-fix-nfsd4_lock-cleanup-on-failure.patch -nfsd4-rename-lk_stateowner.patch -nfsd4-remove-release_state_owner.patch -nfsd4-fix-check_for_locks.patch -nfsd4-operation-debugging.patch -svcrpc-gss-handle-the-gss_s_continue.patch -svcrpc-gss-server-context-init-failure-handling.patch -svcrpc-gss-svc-context-creation-error-handling.patch -nfsd4-fix-open-of-recovery-directory.patch -nfsd4-recovery-lookup-dir-check.patch -nfsd4-handle-replays-of-failed-open-reclaims.patch -nfsd4-no-replays-on-unconfirmed-owners.patch -nfsd4-nfs4statec-miscellaneous-goto-removals.patch -nfsd4-simplify-process-open1-logic.patch -nfsd4-dont-create-on-open-that-fails-due-to-err_grace.patch -nfsd4-fix-open_downgrade.patch -nfsd4-fix-bug-in-rdattr_error-return.patch -nfsd4-clean-up-settattr-code.patch -nfsd-vfsc-endianness-fixes.patch -nfsd4_truncate-bogus-return-value.patch -nfserr_serverfault-returned-host-endian.patch -nfsd4_lock-returns-bogus-values-to-clients.patch -knfsd-fix-some-more-errno-nfserr-confusion-in-vfsc.patch -knfsd-provide-missing-nfsv2-part-of-patch-for-checking-vfs_getattr.patch -exportfs-add-find_acceptable_alias-helper.patch -vfa-at-functions-core.patch
-vfs-at-functions-i386.patch
-vfs-at-functions-x86_64.patch -generic-sys_rt_sigsuspend.patch -generic-sys_rt_sigsuspend-asmlinkage-fix.patch -handle-tif_restore_sigmask-for-frv.patch -handle-tif_restore_sigmask-for-i386.patch -tif_restore_sigmask-support-for-arch-powerpc.patch -uml-add-tif_restore_sigmask-support.patch -uml-use-generic-sys_rt_sigsuspend.patch -add-pselect-ppoll-system-call-implementation.patch -add-pselect-ppoll-system-call-implementation-rename-types.patch -add-pselect-ppoll-system-call-implementation-tidy.patch -add-pselect-ppoll-system-call-implementation-fix.patch -add-pselect-ppoll-system-calls-on-i386.patch

Merged

+x86_64-compat_sys_futimesat-fix.patch

x86_64 fix

+config_isa-does-not-make-sense-for-config_ppc_pseries.patch

pSeries lacks ISA.

+prototypes-for-at-functions-typo-fix.patch
+prototypes-for-at-functions-typo-fix-fix.patch

syscalls.h additions.

+knfsd-restore-recently-broken-acl-functionality-to-nfs-server.patch

knfsd fix

+config_doublefault-kconfig-fix.patch

Kconfig cleanup

+hdspm-printk-warning-fixes.patch
+pcxhr-printk-warning-fix.patch

Warning fixes

+git-audit-fixup.patch

Fix reject due to git-audit.

-sem2mutex-audit_netlink_sem-fix.patch

Folded into sem2mutex-audit_netlink_sem.patch

+gregkh-driver-drivers-base-proper-prototypes.patch +gregkh-driver-empty_release_functions_are_broken.patch -gregkh-driver-aoe-type-cleanups.patch-added-to-mm-tree.patch -gregkh-driver-aoe-skb_check-cleanup.patch +gregkh-driver-aoe-update-driver-compatibility-string.patch

Driver tree updates

+drm-ati-use-null-instead-of-0.patch +ati_pcigart-simplify-page_count-manipulations.patch

DRM cleanups

+gregkh-i2c-hwmon-f71805f-add-documentation.patch +gregkh-i2c-hwmon-f71805f-new-driver.patch +gregkh-i2c-hwmon-it87-probe-i2c-0x2d-only.patch -gregkh-i2c-hwmon-f71805f-new-driver.patch -gregkh-i2c-hwmon-f71805f-add-documentation.patch +gregkh-i2c-i2c-scx200_acb-01-whitespace.patch +gregkh-i2c-i2c-scx200_acb-02-debug.patch +gregkh-i2c-i2c-scx200_acb-03-refactor.patch +gregkh-i2c-i2c-scx200_acb-04-lock_kernel.patch +gregkh-i2c-i2c-scx200_acb-05-cs5535.patch +gregkh-i2c-i2c-scx200_acb-06-poll.patch +gregkh-i2c-i2c-scx200_acb-07-docs.patch +gregkh-i2c-hwmon-sensor-attr-array-2.patch +gregkh-i2c-hwmon-w83792d-use-attr-arrays.patch +gregkh-i2c-hwmon-w83792d-drop-useless-macros.patch +gregkh-i2c-i2c-speedup-block-transfers.patch +gregkh-i2c-i2c-convert-semaphores-to-mutexes-2.patch +gregkh-i2c-i2c-convert-semaphores-to-mutexes-3.patch +gregkh-i2c-hwmon-convert-semaphores-to-mutexes.patch +gregkh-i2c-hwmon-f71805f-convert-semaphore-to-mutex.patch +gregkh-i2c-hwmon-w83627hf-add-w83687thf-support.patch

I2C tree updates

+sem2mutex-input-layer-3.patch

More mutex conversions

+m25p80-printk-warning-fix.patch

Warning fix

+drivers-mtd-small-cleanups.patch

MTD cleanups

-git-netdev-all-revert-e1000-changes.patch

e1000 got fixed.

+kbuild-menu-hide-empty-netdevices-menu-when-net-is-disabled.patch +tweak-orinoco_cs-debugging-message.patch

netdev fixlets

-gregkh-pci-pci-msi-vector-targeting-abstractions.patch -gregkh-pci-pci-per-platform-ia64_-first-last-_device_vector-definitions.patch +gregkh-pci-powerpc-pci-hotplug-remove-rpaphp_find_bus.patch +gregkh-pci-powerpc-pci-hotplug-remove-rpaphp_fixup_new_pci_devices.patch +gregkh-pci-powerpc-pci-hotplug-merge-config_pci_adapter.patch +gregkh-pci-powerpc-pci-hotplug-remove-remove_bus_device.patch +gregkh-pci-powerpc-pci-hotplug-de-convolute-rpaphp_unconfig_pci_adap.patch +gregkh-pci-powerpc-pci-hotplug-merge-rpaphp_enable_pci_slot.patch +gregkh-pci-powerpc-pci-hotplug-cleanup-add-prefix.patch +gregkh-pci-powerpc-pci-hotplug-minor-cleanup-forward-decls.patch +gregkh-pci-powerpc-pci-hotplug-shuffle-error-checking-to-better-location.patch +gregkh-pci-pci-cyblafb-remove-pci_module_init-return-really.patch +gregkh-pci-msi-vector-targeting-abstractions.patch +gregkh-pci-per-platform-ia64_-first-last-_device_vector-definitions.patch +gregkh-pci-altix-msi-support.patch

PCI tree updates

+git-pcmcia-orinoco_cs-fix.patch

git-pcmcia fix

+megaraid-unused-variable.patch

Warning fix

+drivers-scsi-aic7xxx-possible-cleanups.patch +module_alias_blockchardev_major-for-drivers-scsi.patch

SCSI cleanups and fix.

+gregkh-usb-usb-remove-misc-devfs-droppings.patch +gregkh-usb-usb-net2280-warning-fix.patch +gregkh-usb-add-might_sleep-to-usb_unlink_urb.patch +gregkh-usb-usb-add-new-pl2303-device-ids.patch +gregkh-usb-usb-cp2101-add-new-device-ids.patch +gregkh-usb-usb-arm26-fix-compilation-of-drivers-usb-core-message.c.patch +gregkh-usb-usbatm-trivial-modifications.patch +gregkh-usb-usbatm-add-flags-field.patch +gregkh-usb-usbatm-remove-.owner.patch +gregkh-usb-usbatm-kzalloc-conversion.patch +gregkh-usb-usbatm-xusbatm-rewrite.patch +gregkh-usb-usbatm-shutdown-open-connections-when-disconnected.patch +gregkh-usb-usbatm-return-correct-error-code-when-out-of-memory.patch +gregkh-usb-usbatm-use-dev_kfree_skb_any-rather-than-dev_kfree_skb.patch +gregkh-usb-usbatm-measure-buffer-size-in-bytes-force-valid-sizes.patch +gregkh-usb-usbatm-allow-isochronous-transfer.patch +gregkh-usb-usbatm-handle-urbs-containing-partial-cells.patch +gregkh-usb-usbatm-bump-version-numbers.patch +gregkh-usb-usbatm-eilseq-workaround.patch +gregkh-usb-usbatm-semaphore-to-mutex-conversion.patch +gregkh-usb-ueagle-add-iso-support.patch +gregkh-usb-ueagle-cosmetic.patch +gregkh-usb-ueagle-cmv-name-bug.patch +gregkh-usb-usb-add-new-auerswald-device-ids.patch +gregkh-usb-usb-libusual-fix-warning-on-64bit-boxes.patch +gregkh-usb-usb-core-and-hcds-don-t-put_device-while-atomic.patch

USB tree updates

+usb-yealink-printk-warning-fixes.patch
+usb-usbip-warning-fixes.patch

Warning fixes

+x86_64-defconfig-update.patch
+x86_64-config-unwind-info.patch
+x86_64-vsyscall-patch-xen.patch
+x86_64-nmi-kprobes.patch
+x86_64-apic-main-timer.patch
+x86_64-apic-main-timer-default.patch
+x86_64-timer-resume.patch

x86_64 tree update

+hrtimers-fixup-itimer-conversion.patch +hrtimers-fix-possible-use-of-null-pointer-in.patch +hrtimers-fix-oldvalue-return-in-setitimer.patch +hrtimers-fix-posix-timer-requeue-race.patch +hrtimers-cleanups-and-simplifications.patch +hrtimers-add-back-lost-credit-lines.patch +hrtimers-set-correct-initial-expiry-time-for-relative.patch +hrtimers-set-correct-initial-expiry-time-for-relative-fix.patch

hrtimers fixes

+optimize-off-node-performance-of-zone-reclaim.patch +zone_reclaim-reclaim-on-memory-only-node-support.patch +gfp_zonetypes-add-commentry-on-how-to-calculate.patch +gfp_zonetypes-calculate-from-gfp_zonemask.patch +mm-improve-function-of-sc-may_writepage.patch

mm fixes and updates.

+produce-useful-info-for-kzalloc-with-debug_slab.patch

Make kzalloc() play properly with slab debugging

+dump_stack-in-oom.patch

Do a stack dump in oom-killings.

+selinux-fix-and-cleanup-mprotect-checks.patch +selinux-change-file_alloc_security-to-use-gfp_kernel.patch

SELinux updates

+i386-multi-column-stack-backtraces-update.patch

Make the x86 stack dumps default to two columns.

+i386-print-kernel-version-in-register.patch

Print build number in oopses.

+arm26-fix-find_first_zero_bit-related-warnings.patch +arm26-fix-warnings-about-nr_irqs-being-not-defined.patch +arm26-remove-irq_exit-from-hardirqh.patch +arm26-select-system-type-via-choice.patch +arm26-fixup-get_signal_to_deliver-call.patch +arm26-fixup-asm-statement-in-kernel-fiqc.patch +arm26-drop-local-task_running-copy.patch +arm26-drop-first-arg-of-prepare_arch_switch-finish_arch_switch.patch +arm26-add-__kernel_old_dev_t-for-nfsd.patch +arm26-select-blk_dev_fd-only-on-a5k.patch

arm25 fixes

+efi-dev-mem-simplify-efi_mem_attribute_range.patch +ia64-ioremap-check-efi-for-valid-memory-attributes.patch +ia64-ioremap-check-efi-for-valid-memory-attributes-fix.patch +dmi-only-ioremap-stuff-we-actually-need.patch +efi-keep-physical-table-addresses-in-efi-structure.patch +acpi-clean-up-memory-attribute-checking-for-map-read-write.patch

ia64/DMI work.

+uml-typo-fixup.patch +uml-comments-about-libc-conflict-guards.patch +uml-fix-hugest-stack-users.patch +uml-fix-apples-bananas-typo.patch +uml-tt-syscall_debug-fix-buglet-introduced-in-cleanup.patch +uml-skas0-hold-own-ldt-fixups-for-x86-64.patch +uml-some-harmless-sparse-warning-fixes.patch +uml-avoid-config_nr_cpus-undeclared-bogus-error-messages.patch

UML updates

+s390-build-dasd_cmd-into-dasd_mod.patch +s390-dasd-remove-dynamic-ioctl-registration.patch +s390-remove-cvs-generated-information.patch +s390-overflow-in-sched_clock.patch +s390-monotonic_clock-interface.patch +s390-hangcheck-timer-support.patch +s390-ccw_device_probe_console-return-value.patch +s390-dasd-open-counter.patch +s390-dasd-wait-for-clear-i-o-interrupt.patch

s390 updates

+work-around-ppc64-bootup-bug-by-making-mutex-debugging-save-restore-irqs.patch

Don't accidentally enable interrupts in mutex debugging code: powerpc falls over.

+kernel-kernel-cpuc-to-mutexes.patch

Bring back this mutex conversion.

+sbc-epx-does-not-check-claim-i-o-ports-it-uses-2nd-edition-fix.patch

Fix watchdog driver

+parport-fix-printk-format-warning.patch
+dvb-fix-printk-format-warning.patch

Warning fixes

+fix-cpucontrol-cache_chain_mutex-lock-inversion-bug.patch

Fix deadlock.

+make-bug-messages-more-consistent.patch
+make-bug-messages-more-consistent-update.patch

Make the bug messages say "BUG:"

+add-trylock_kernel.patch
+add-trylock_kernel-fix.patch

More workarounds for the enabling of might_sleep() debugging in early boot.

+turn-on-might_sleep-in-early-bootup-code-too.patch

Enable might_sleep() debugging in early boot.

+dont-allow-users-to-set-config_broken=y.patch

CONFIG_BROKEN really means it.

+kill-_inline_.patch

Remove _INLINE_

+pause_on_oops-command-line-option.patch

I wrote a patch! If your oopses are scrolling off the screen, add `pause_on_oops=100000' to the kernel boot command line.

+pnpbios-missing-small_tag_enddep-tag.patch

pnpbios fix.

+fix-i2o_scsi-oops-on-abort.patch

i2o driver fix

+build_lock_ops-cleanup-preempt_disable-usage.patch

spinlock speedup

+tpm_infineon-fix-printk-format-warning.patch
+tpm_bios-needs-more-securityfs_-functions.patch
+tpm_bios-securityfs-error-checking-fix.patch
+tpm_bios-indexing-fix.patch

tpm driver fixes

+someone-broke-reiserfs-v3-mount-options-this-fixes-it.patch

Fix reiser3 mount option handling.

+parport_serial-printk-warning-fix.patch
+quota_v2-printk-warning-fixes.patch
+sxc-printk-warning-fixes.patch

Warning fixes.

+autofs4-lookup-white-space-cleanup.patch +autofs4-use-libfs-routines-for-readdir.patch +autofs4-cant-mount-due-to-mount-point-dir-not-empty.patch +autofs4-expire-code-readability-cleanup.patch +autofs4-simplify-expire-tree-traversal.patch +autofs4-fix-false-negative-return-from-expire.patch +autofs4-expire-mounts-that-hold-no-extra-references-only.patch +autofs4-expire-mounts-that-hold-no-extra-references-only-fix.patch +autofs4-remove-update_atime-unused-function.patch +autofs4-add-a-show-mount-options-for-proc-filesystem.patch +autofs4-white-space-cleanup-for-waitqc.patch +autofs4-rename-simple_empty_nolock-function.patch +autofs4-change-may_umount-functions-to-boolean.patch +autofs4-increase-module-version.patch

Autofs4 updates.

-reiserfs-fix-is_reusable-bitmap-check-to-not-traverse-the-bitmap-info-array.patch -reiserfs-clean-up-bitmap-block-buffer-head-references.patch -reiserfs-move-bitmap-loading-to-bitmapc.patch -reiserfs-on-demand-bitmap-loading.patch -reiserfs-on-demand-bitmap-loading-fix.patch -reiserfs-on-demand-bitmap-loading-warning-fix.patch

Dropped - baaaad.

+ext3-get-blocks-multiple-block-allocation-cleanup.patch

Tidy ext3-get-blocks-multiple-block-allocation.patch

-powerpc-fastpaths-for-mutex-subsystem.patch

Dropped.

+x86-blacklist-tsc-from-systems-where-it-is-known-to-be-bad.patch

Work around dodgy TSCs

+kernel-kprobesc-fix-a-warning-ifndef-arch_supports_kretprobes.patch

Warning fix.

+dlm-recovery-remove-true-false-defines.patch
+dlm-device-interface-missing-variable.patch
+dlm-device-interface-check-allocation.patch
+dlm-device-interface-fix-unlock-race.patch
+dlm-device-interface-use-kzalloc.patch
+dlm-sem2mutex.patch

DLM updates.

+drivers-ide-ide-ioc-make-__ide_end_request-static.patch

IDE cleanup

+epoll_pwait.patch

epoll feature addition (controversial).

+documentation-ioctl-messtxt-update.patch

ioctl() documentation update.

All 761 patches:

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.16-rc1/2.6.16-rc1-mm2/patch-list

#2 lore raw
Quoted 10 lines · Pekka Enberg

Hi Andrew,

On 1/20/06, Andrew Morton <[email protected]> wrote:
+produce-useful-info-for-kzalloc-with-debug_slab.patch

Quoted 2 lines

Make kzalloc() play properly with slab debugging

Hmm. This still leaves kstrdup() broken which is why I would prefer the following patch to be applied:

kstrdup() doesn't get used much.

Quoted 1 line

http://marc.theaimsgroup.com/?l=linux-kernel&m=113767657400334&w=2

That adds more complexity, IMO. A bit ifdeffy too. __do_kmalloc() should be __always_inline, methinks?

#3 lore raw
Quoted 2 lines · Pekka Enberg
Quoted 2 lines

Hmm. This still leaves kstrdup() broken which is why I would prefer the following patch to be applied:

Quoted 6 lines · Andrew Morton

kstrdup() doesn't get used much.

Quoted 1 line

http://marc.theaimsgroup.com/?l=linux-kernel&m=113767657400334&w=2

That adds more complexity, IMO. A bit ifdeffy too. __do_kmalloc() should be __always_inline, methinks?

Yes it does. This patch does make the caller tracing more explicit, though and less likely to break. And yes, __do_kmalloc() should always be inlined. Here's an updated patch.

Pekka

[PATCH] slab: fix kzalloc and kstrdup caller report for CONFIG_DEBUG_SLAB

From: Pekka Enberg <[email protected]>

This patch fixes kzalloc() and kstrdup() caller report for CONFIG_DEBUG_SLAB. We must pass the caller to __cache_alloc() instead of directly doing __builtin_return_address(0) there; otherwise kzalloc() and kstrdup() are reported as the allocation site instead of the real one.

Thanks to Valdis Kletnieks for reporting the problem and Steven Rostedt for the original idea.

---

 include/linux/slab.h |    7 +++++++
 mm/slab.c            |   29 ++++++++++++++++++++++++-----
 2 files changed, 31 insertions(+), 5 deletions(-)
include/linux/slab.h +7 -0
Index: 2.6-mm/include/linux/slab.h
===================================================================
@@ -76,7 +76,14 @@ struct cache_sizes {
 	kmem_cache_t	*cs_dmacachep;
 };
 extern struct cache_sizes malloc_sizes[];
+
+#ifndef CONFIG_DEBUG_SLAB
 extern void *__kmalloc(size_t, gfp_t);
+#else
+extern void *__kmalloc_track_caller(size_t, gfp_t, void*);
+#define __kmalloc(size, flags) \
+    __kmalloc_track_caller(size, flags, __builtin_return_address(0))
+#endif
 
 static inline void *kmalloc(size_t size, gfp_t flags)
 {
mm/slab.c +24 -5
Index: 2.6-mm/mm/slab.c
===================================================================
@@ -2699,7 +2699,8 @@ static inline void *____cache_alloc(stru
 	return objp;
 }
 
-static inline void *__cache_alloc(struct kmem_cache *cachep, gfp_t flags)
+static __always_inline void *
+__cache_alloc(struct kmem_cache *cachep, gfp_t flags, void *caller)
 {
 	unsigned long save_flags;
 	void *objp;
@@ -2710,7 +2711,7 @@ static inline void *__cache_alloc(struct
 	objp = ____cache_alloc(cachep, flags);
 	local_irq_restore(save_flags);
 	objp = cache_alloc_debugcheck_after(cachep, flags, objp,
-					    __builtin_return_address(0));
+					    caller);
 	prefetchw(objp);
 	return objp;
 }
@@ -2939,7 +2940,7 @@ static inline void __cache_free(struct k
  */
 void *kmem_cache_alloc(struct kmem_cache *cachep, gfp_t flags)
 {
-	return __cache_alloc(cachep, flags);
+	return __cache_alloc(cachep, flags, __builtin_return_address(0));
 }
 EXPORT_SYMBOL(kmem_cache_alloc);
 
@@ -3053,7 +3054,8 @@ EXPORT_SYMBOL(kmalloc_node);
  * platforms.  For example, on i386, it means that the memory must come
  * from the first 16MB.
  */
-void *__kmalloc(size_t size, gfp_t flags)
+static __always_inline void *__do_kmalloc(size_t size, gfp_t flags,
+					  void *caller)
 {
 	struct kmem_cache *cachep;
 
@@ -3065,10 +3067,27 @@ void *__kmalloc(size_t size, gfp_t flags
 	cachep = __find_general_cachep(size, flags);
 	if (unlikely(cachep == NULL))
 		return NULL;
-	return __cache_alloc(cachep, flags);
+	return __cache_alloc(cachep, flags, caller);
+}
+
+#ifndef CONFIG_DEBUG_SLAB
+
+void *__kmalloc(size_t size, gfp_t flags)
+{
+	return __do_kmalloc(size, flags, NULL);
 }
 EXPORT_SYMBOL(__kmalloc);
 
+#else
+
+void *__kmalloc_track_caller(size_t size, gfp_t flags, void *caller)
+{
+	return __do_kmalloc(size, flags, caller);
+}
+EXPORT_SYMBOL(__kmalloc_track_caller);
+
+#endif
+
 #ifdef CONFIG_SMP
 /**
  * __alloc_percpu - allocate one copy of the object for every present
#4 lore raw
In reply to Andrew Morton
Quoted 11 lines · Andrew Morton
Quoted 2 lines · Pekka Enberg

Hmm. This still leaves kstrdup() broken which is why I would prefer the following patch to be applied:

kstrdup() doesn't get used much.

Quoted 1 line

http://marc.theaimsgroup.com/?l=linux-kernel&m=113767657400334&w=2

That adds more complexity, IMO. A bit ifdeffy too. __do_kmalloc() should be __always_inline, methinks?

For what it's worth, I spent some time trying to get kstrdup() fixed as well, but gave up because:

a) The leak I was chasing was using kzalloc() ;)

b) I got stuck in a .h dependency loop - the prototype for kzalloc is in slab.h, and by that point in slab.h, we've seen a prototype for __kmalloc(). Unfortunately, kstrdup() comes out of string.h, and there's usages where string.h has been #included but we've not seen __kmalloc(). Unfortunately, just #include'ing what's needed didn't work, because of the way the code uses #ifndef _LINUX_FOO_H_ - when I started looking at having to #undef stuff to get a prototype of __kmalloc(), I gave up. ;)

#5 lore raw

PATCH 1/1 zone_reclaim cpus_empty needs a real variable

In reply to Andrew Morton

zone_reclaim cpus_empty needs a real variable

On some architectures cpus_empty() attempts to take the address of the mask. Consequently we must store the result of the node_to_cpumask() before applying it.

---
 vmscan.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
mm/vmscan.c +5 -2
diff -upN reference/mm/vmscan.c current/mm/vmscan.c
@@ -1836,18 +1836,21 @@ int zone_reclaim(struct zone *zone, gfp_
 	struct task_struct *p = current;
 	struct reclaim_state reclaim_state;
 	struct scan_control sc;
+	cpumask_t mask;
 
 	if (time_before(jiffies,
 		zone->last_unsuccessful_zone_reclaim + ZONE_RECLAIM_INTERVAL))
 			return 0;
 
 	if (!(gfp_mask & __GFP_WAIT) ||
-		(!cpus_empty(node_to_cpumask(zone->zone_pgdat->node_id)) &&
-			 zone->zone_pgdat->node_id != numa_node_id()) ||
 		zone->all_unreclaimable ||
 		atomic_read(&zone->reclaim_in_progress) > 0)
 			return 0;
 
+	mask = node_to_cpumask(zone->zone_pgdat->node_id);
+	if (!cpus_empty(mask) && zone->zone_pgdat->node_id != numa_node_id())
+		return 0;
+
 	sc.may_writepage = 0;
 	sc.may_swap = 0;
 	sc.nr_scanned = 0;
#6 lore raw

Re: [PATCH] zone_reclaim cpus_empty needs a real variable

Quoted 1 line · Andy Whitcroft

zone_reclaim cpus_empty needs a real variable

Maybe we also need to add a new variable for

zone->zone_pgdat->node_id

Its used multiple times

Quoted 42 lines

On some architectures cpus_empty() attempts to take the address of the mask. Consequently we must store the result of the node_to_cpumask() before applying it.

---
 vmscan.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff -upN reference/mm/vmscan.c current/mm/vmscan.c

mm/vmscan.c +5 -3
@@ -1836,18 +1836,21 @@ int zone_reclaim(struct zone *zone, gfp_
 	struct task_struct *p = current;
 	struct reclaim_state reclaim_state;
 	struct scan_control sc;
+	cpumask_t mask;
 
 	if (time_before(jiffies,
 		zone->last_unsuccessful_zone_reclaim + ZONE_RECLAIM_INTERVAL))
 			return 0;
 
 	if (!(gfp_mask & __GFP_WAIT) ||
-		(!cpus_empty(node_to_cpumask(zone->zone_pgdat->node_id)) &&
-			 zone->zone_pgdat->node_id != numa_node_id()) ||
 		zone->all_unreclaimable ||
 		atomic_read(&zone->reclaim_in_progress) > 0)
 			return 0;
 
+	mask = node_to_cpumask(zone->zone_pgdat->node_id);
+	if (!cpus_empty(mask) && zone->zone_pgdat->node_id != numa_node_id())
+		return 0;
+
 	sc.may_writepage = 0;
 	sc.may_swap = 0;
 	sc.nr_scanned = 0;
-

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#7 lore raw

Re: [PATCH] zone_reclaim cpus_empty needs a real variable

In reply to Andy Whitcroft
Quoted 3 lines · Andy Whitcroft

On some architectures cpus_empty() attempts to take the address
of the mask.

That's pretty awkward. Can they not do that?

#8 lore raw
In reply to Andrew Morton
Quoted 2 lines · Andrew Morton

- drivers/i2c/busses/scx200_acb.c doesn't compile on architectures which don't have asm/msr.h.

It shouldn't be offered on other arches, it's an arch specific driver. Make it behave like the other SCx200 bits in that dir, and hide away from non-x86.

drivers/i2c/busses/Kconfig renamed from drivers/i2c/busses/Kconfig~ +1 -1
@@ -389,7 +389,7 @@ config SCx200_I2C_SDA
 
 config SCx200_ACB
 	tristate "NatSemi SCx200 ACCESS.bus"
-	depends on I2C && PCI
+	depends on SCx200 && I2C && PCI
 	help
 	  Enable the use of the ACCESS.bus controllers of a SCx200 processor.
 
#9 lore raw

Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Andrew Morton
Quoted 4 lines · Andrew Morton

Changes since 2.6.16-rc1-mm1:

linus.patch

Hello,

Commit 4f2d7680cb1ac5c5a70f3ba2447d5aa5c0a1643a (Linus' 2.6 git tree) breaks my iptables (1.3.4):

# iptables -L
execve("/sbin/iptables", ["iptables", "-L"], [/* 24 vars */]) = 0
brk(0)                                  = 0x8056000
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7fdb000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=71332, ...}) = 0
old_mmap(NULL, 71332, PROT_READ, MAP_PRIVATE, 3, 0) = 0xb7fc9000
close(3)                                = 0
open("/lib/libdl.so.2", O_RDONLY)       = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\364\273"..., 512) = 512
fstat64(3, {st_mode=S_IFREG|0755, st_size=16164, ...}) = 0
old_mmap(0x4103b000, 12408, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x4103b000
old_mmap(0x4103d000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x1000) = 0x4103d000
close(3)                                = 0
open("/lib/libc.so.6", O_RDONLY)        = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0|\236\360"..., 512) = 512
fstat64(3, {st_mode=S_IFREG|0755, st_size=1432592, ...}) = 0
old_mmap(0x4fef5000, 1162204, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x4fef5000
old_mmap(0x5000b000, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x116000) = 0x5000b000
old_mmap(0x5000f000, 7132, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x5000f000
close(3)                                = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7fc8000
set_thread_area({entry_number:-1 -> 6, base_addr:0xb7fc86c0, limit:1048575, seg_32bit:1, contents:0, read_exec_only:0, limit_in_pages:1, seg_not_present:0, useable:1}) = 0
mprotect(0x4103d000, 4096, PROT_READ)   = 0
mprotect(0x5000b000, 8192, PROT_READ)   = 0
mprotect(0x4fef1000, 4096, PROT_READ)   = 0
munmap(0xb7fc9000, 71332)               = 0
socket(PF_INET, SOCK_RAW, IPPROTO_RAW)  = 3
getsockopt(3, SOL_IP, 0x40 /* IP_??? */, "filter\0\0\0\0\0\0l\216\4\10\364\317\0PL!\0\0\330\320\0"..., [84]) = 0
brk(0)                                  = 0x8056000
brk(0x8077000)                          = 0x8077000
getsockopt(3, SOL_IP, 0x41 /* IP_??? */, "filter\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., [656]) = 0
write(2, "ERROR: 0 not a valid target)\n", 29ERROR: 0 not a valid target)
) = 29
rt_sigprocmask(SIG_UNBLOCK, [ABRT], NULL, 8) = 0
gettid()                                = 3468
tgkill(3468, 3468, SIGABRT)             = 0
--- SIGABRT (Aborted) @ 0 (0) ---
+++ killed by SIGABRT +++

This is it:
[NETFILTER] x_tables: Make XT_ALIGN align as strictly as necessary.

Or else we break on ppc32 and other 32-bit platforms.

Based upon a patch from Harald Welte.

linux/netfilter/x_tables.h +0 -0
@@ -19,7 +19,7 @@ struct xt_get_revision

/* For standard target */
#define XT_RETURN (-NF_REPEAT - 1)
-#define XT_ALIGN(s) (((s) + (__alignof__(void *)-1)) & ~(__alignof__(void *)-1)) +#define XT_ALIGN(s) (((s) + (__alignof__(u_int64_t)-1)) & ~(__alignof__(u_int64_t)-1)) /* Standard return verdict, or do jump. */
#define XT_STANDARD_TARGET ""

Is there more info needed? Did I miss something?

regards,

Signature · 3 lines
Jiri Slaby         www.fi.muni.cz/~xslaby
\_.-^-._   [email protected]   _.-^-._/
B67499670407CE62ACC8 22A032CC55C339D47A7E
#10 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

Quoted 2 lines · Jiri Slaby

Commit 4f2d7680cb1ac5c5a70f3ba2447d5aa5c0a1643a (Linus' 2.6 git tree) breaks my iptables (1.3.4):

You missed to indicate on which architecture?

Signature · 5 lines
- Harald Welte <[email protected]>                 http://netfilter.org/
============================================================================
  "Fragmentation is like classful addressing -- an interesting early
   architectural error that shows how much experimentation was going
   on while IP was being designed."                    -- Paul Vixie
#11 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

Quoted 6 lines · Harald Welte
Quoted 1 line · Jiri Slaby

Commit 4f2d7680cb1ac5c5a70f3ba2447d5aa5c0a1643a (Linus' 2.6 git tree) bre=

aks my

Quoted 1 line

iptables (1.3.4):

You missed to indicate on which architecture?

Of course I did, sorry:
Linux bellona 2.6.16-rc1-mm2good #89 SMP PREEMPT Fri Jan 20 17:05:23 CET 2006 i686 i686 i386 GNU/Linux not ppc, not 64 bit...

regards,
--
Jiri Slaby www.fi.muni.cz/~xslaby
\_.-^-._ [email protected] _.-^-._/
B67499670407CE62ACC8 22A032CC55C339D47A7E

#12 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Harald Welte
Quoted 6 lines · Harald Welte
Quoted 2 lines · Jiri Slaby

Commit 4f2d7680cb1ac5c5a70f3ba2447d5aa5c0a1643a (Linus' 2.6 git tree) breaks my iptables (1.3.4):

You missed to indicate on which architecture?

On x86 (32bits), i have the same i think:

# iptables  -L
ERROR: 0 not a valid target)
Aborted

I can provide a strace if necessary.

regards,

Benoit

#13 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

Quoted 2 lines · Benoit Boissinot

On x86 (32bits), i have the same i think:

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

Or do I just remember wrong?

Linus

#14 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

Quoted 9 lines · Linus Torvalds
Quoted 2 lines · Benoit Boissinot

On x86 (32bits), i have the same i think:

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

I use gcc-4 (gcc version 4.1.0-beta20060113), but i can try with something more conservative.

regards,

Benoit

#15 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

Quoted 5 lines · Benoit Boissinot
Quoted 4 lines · Linus Torvalds
Quoted 1 line · Benoit Boissinot
Quoted 8 lines
Quoted 5 lines
Quoted 1 line

On x86 (32bits), i have the same i think:

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

I use gcc-4 (gcc version 4.1.0-beta20060113), but i can try with something more conservative.

Same error with gcc-3.4 (gcc version 3.4.4)

regards,

Benoit

Quoted 5 lines

regards,

Benoit

#16 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Linus Torvalds
Quoted 9 lines · Linus Torvalds
Quoted 2 lines · Benoit Boissinot

On x86 (32bits), i have the same i think:

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

Or do I just remember wrong?

$ cat c.c
#include <stdio.h>

int main()

{
	printf("%d\n", __alignof__(unsigned long long));

	return 0;
}
$ gcc --version
gcc (GCC) 4.0.2 20051125 (Red Hat 4.0.2-8)
$ gcc32 --version
gcc32 (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-47.fc4)
$ gcc c.c -occ -Wall && ./cc
8
$ gcc32 c.c -occ -Wall && ./cc
8

If it helps...

regards,
--
Jiri Slaby www.fi.muni.cz/~xslaby
\_.-^-._ [email protected] _.-^-._/
B67499670407CE62ACC8 22A032CC55C339D47A7E

#17 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

Quoted 2 lines · Jiri Slaby
$ gcc --version
gcc (GCC) 4.0.2 20051125 (Red Hat 4.0.2-8)

btw. I use this one.

regards,
--
Jiri Slaby www.fi.muni.cz/~xslaby
\_.-^-._ [email protected] _.-^-._/
B67499670407CE62ACC8 22A032CC55C339D47A7E

#18 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Linus Torvalds
Quoted 11 lines · Linus Torvalds
Quoted 2 lines · Benoit Boissinot

On x86 (32bits), i have the same i think:

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

Or do I just remember wrong?

I don't know what __alignof__ does but afaik, unsigned long long has been 8 bytes at least since gcc 3.3.X. I don't know the size of it in 4.X.X.

#19 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Linus Torvalds
Quoted 7 lines · Linus Torvalds
Quoted 2 lines · Benoit Boissinot

On x86 (32bits), i have the same i think:

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

The problem seems to have been accidentially introduced by DaveM's "simplification" of my original patch.

I've already asked Dave to revert his change and apply my original patch (see attachment), which _should_ fix the problem.

Thanks,

Signature · 5 lines
- Harald Welte <[email protected]>                 http://netfilter.org/
============================================================================
  "Fragmentation is like classful addressing -- an interesting early
   architectural error that shows how much experimentation was going
   on while IP was being designed."                    -- Paul Vixie

[NETFILTER] x_tables: Fix XT_ALIGN() macro on [at least] ppc32

To keep backwards compatibility with old iptables userspace programs, the new XT_ALIGN macro always has to return the same value as IPT_ALIGN, IP6T_ALIGN or ARPT_ALIGN in previous kernels.

However, in those kernels the macro was defined in dependency to the respective layer3 specifi data structures, which we can no longer do with x_tables.

The fix is an ugly kludge, but it has been tested to solve the problem. Yet another reason to move away from the current {ip,ip6,arp,eb}tables like data structures.

---
commit 470faeb379560fe877b685ca69be6a7e4f0e91ed
tree 5732ecd9bcab28469805752514e5c57ba26189a1
parent 44718bbfa186d58477163418d37df173aa2dd079
author Harald Welte <[email protected]> Fri, 20 Jan 2006 01:44:24 +0100 committer Harald Welte <[email protected]> Fri, 20 Jan 2006 01:44:24 +0100

 include/linux/netfilter/x_tables.h |   15 ++++++++++++++-
 1 files changed, 14 insertions(+), 1 deletions(-)
include/linux/netfilter/x_tables.h +14 -1
index 472f048..65f9cd8 100644
@@ -19,7 +19,20 @@ struct xt_get_revision
 /* For standard target */
 #define XT_RETURN (-NF_REPEAT - 1)
 
-#define XT_ALIGN(s) (((s) + (__alignof__(void *)-1)) & ~(__alignof__(void *)-1))
+/* this is a dummy structure to find out the alignment requirement for a struct
+ * containing all the fundamental data types that are used in ipt_entry, ip6t_entry
+ * and arpt_entry.  This sucks, and it is a hack.  It will be my personal pleasure
+ * to remove it -HW */
+struct _xt_align
+{
+	u_int8_t u8;
+	u_int16_t u16;
+	u_int32_t u32;
+	u_int64_t u64;
+};
+
+#define XT_ALIGN(s) (((s) + (__alignof__(struct _xt_align)-1)) 	\
+			& ~(__alignof__(struct _xt_align)-1))
 
 /* Standard return verdict, or do jump. */
 #define XT_STANDARD_TARGET ""
#20 lore raw

Re: Iptables error

From: Harald Welte <[email protected]>
Date: Fri, 20 Jan 2006 20:32:01 +0100

Quoted 5 lines

The problem seems to have been accidentially introduced by DaveM's "simplification" of my original patch.

I've already asked Dave to revert his change and apply my original patch (see attachment), which _should_ fix the problem.

Your struct won't be 8-byte aligned either as far as I can tell on x86_64.

We need to use the aligned_u64 thing if you want that.

#21 lore raw

Re: Iptables error

Quoted 2 lines · David S. Miller

Your struct won't be 8-byte aligned either as far as I can tell on x86_64.

According to my tests, the struct is 8-byte-aligned on x86_64, and that's how I'd like it to be.

Please don't ask me why it happens, I know that the alignment constraint of a u64 on x86_64 is only 4. But at least gcc-3.3.6 and gcc-4.0.3 (debian) result in __alignof__ of that test structure (and a 'u_int64_t alone') to 8 bytes.

When it comes to these things, I can only do trial+error.

Maybe it's because __alignof__ returns the recommended alignment, not the required alignment.

Quoted 1 line

We need to use the aligned_u64 thing if you want that.

That should make sure that we always get what we want, yes.

Signature · 5 lines
- Harald Welte <[email protected]>                 http://netfilter.org/
============================================================================
  "Fragmentation is like classful addressing -- an interesting early
   architectural error that shows how much experimentation was going
   on while IP was being designed."                    -- Paul Vixie
#22 lore raw

Re: Iptables error

In reply to Linus Torvalds

From: Linus Torvalds <[email protected]>
Date: Fri, 20 Jan 2006 11:49:46 -0500 (EST)

Quoted 5 lines

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

Or do I just remember wrong?

I think I remember the gcc folks talking about changing this some time long in the past, aparently they did.

#23 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Linus Torvalds
Quoted 4 lines · Linus Torvalds

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

gcc 2.95 says 4 while gcc 3.2 says 8.

Signature · 4 lines
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
#24 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

Quoted 7 lines · Herbert Xu (on Mon, 23 Jan 2006 13:03:20 +1100) wrote:
Quoted 4 lines · Linus Torvalds

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

gcc 2.95 says 4 while gcc 3.2 says 8.

Has somebody turned on -malign-double in gcc?

#25 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Linus Torvalds
Quoted 9 lines · Linus Torvalds
Quoted 2 lines · Benoit Boissinot

On x86 (32bits), i have the same i think:

Interestingly, __alignof__(unsigned long long) is 8 these days, even though I think historically on x86 it was 4. Is this perhaps different in gcc-3 and gcc-4?

Or do I just remember wrong?

Nope. There are compilers where it is 4 byte aligned. I believe this was actually a C abi change.

I actually had some code break because of it. A 32bit binary generated a structure and a 64bit binary couldn't read it. I hadn't realized they had changed recent versions of gcc.

Eric

#26 lore raw

Re: Iptables error [Was: 2.6.16-rc1-mm2]

In reply to Jiri Slaby
Quoted 21 lines · Jiri Slaby
Quoted 4 lines · Andrew Morton

Changes since 2.6.16-rc1-mm1:

linus.patch

Hello,

Commit 4f2d7680cb1ac5c5a70f3ba2447d5aa5c0a1643a (Linus' 2.6 git tree) breaks my iptables (1.3.4):

This is it:
[NETFILTER] x_tables: Make XT_ALIGN align as strictly as necessary.

Or else we break on ppc32 and other 32-bit platforms.

Based upon a patch from Harald Welte.

linux/netfilter/x_tables.h struct xt_get_revision · line 23

Quoted 6 lines
linux/netfilter/x_tables.h +1 -1
@@ -19,7 +19,7 @@ struct xt_get_revision
 /* For standard target */
 #define XT_RETURN (-NF_REPEAT - 1)
-#define XT_ALIGN(s) (((s) + (__alignof__(void *)-1)) & ~(__alignof__(void *)-1))
+#define XT_ALIGN(s) (((s) + (__alignof__(u_int64_t)-1)) & ~(__alignof__(u_int64_t)-1))
 /* Standard return verdict, or do jump. */
 #define XT_STANDARD_TARGET ""

Confirmed here. Backing out this one-liner makes iptables work for me. i686 on a Pentium-4, gcc 4.1.0 from Fedora -devel tree.

#27 lore raw

Re: Iptables error

From: [email protected]
Date: Fri, 20 Jan 2006 13:13:36 -0500

Quoted 2 lines

Confirmed here. Backing out this one-liner makes iptables work for me. i686 on a Pentium-4, gcc 4.1.0 from Fedora -devel tree.

Ok this is on x86. I think I see how it breaks, but I thought Harald's patch would have the same problem.

I just ran a test program, and indeed __alignof__() gives 8 for "long long" and 4 for a struct containing a "long long" on x86. Yikes...

Linus is likely about to be on his way to the airport so I'll push the fix in New Zealand.

#28 lore raw

2.6.16-rc1-mm2: arch/arm26/kernel/fiq.c still doesn't compile

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.16-rc1-mm1:
...
+arm26-fixup-asm-statement-in-kernel-fiqc.patch
...
arm25 fixes
...

This doesn't seem to be enough to fix the arm27 compilation [1]:

<-- snip -->

...
CC arch/arm26/kernel/fiq.o
/usr/src/ctest/mm/kernel/arch/arm26/kernel/fiq.c:1: note: future releases of GCC will not support -mapcs-26 /usr/src/ctest/mm/kernel/arch/arm26/kernel/fiq.c: In function `set_fiq_regs':
/usr/src/ctest/mm/kernel/arch/arm26/kernel/fiq.c:122: error: fp cannot be used in asm here make[2]: *** [arch/arm26/kernel/fiq.o] Error 1

<-- snip -->

cu
Adrian

[1] http://l4x.org/k/?d=10230

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#29 lore raw

Re: 2.6.16-rc1-mm2: arch/arm26/kernel/fiq.c still doesn't compile

Quoted 9 lines · Adrian Bunk
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.16-rc1-mm1:
...
+arm26-fixup-asm-statement-in-kernel-fiqc.patch
...
arm25 fixes
...

Quoted 7 lines

This doesn't seem to be enough to fix the arm27 compilation [1]:

CC arch/arm26/kernel/fiq.o
/usr/src/ctest/mm/kernel/arch/arm26/kernel/fiq.c:1: note: future releases of GCC will not support -mapcs-26 /usr/src/ctest/mm/kernel/arch/arm26/kernel/fiq.c: In function `set_fiq_regs':
/usr/src/ctest/mm/kernel/arch/arm26/kernel/fiq.c:122: error: fp cannot be used in asm here make[2]: *** [arch/arm26/kernel/fiq.o] Error 1

HOSTCC=gcc-3.4

Downgrading to 3.3.* helped me.

#30 lore raw
In reply to Andrew Morton
Quoted 15 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.16-rc1/2.6.16-rc1-mm2/

- This kernel has a big ACPI update

- reiser3 should be safe(r) to use.

Known problems:

- You'll probably see something like this

Memory: 4017084k/6291456k available (2896k kernel code, 176452k reserved, 1868k data, 208k init)

BUG: sleeping function called from invalid context at kernel/mutex.c:84         in_atomic():1, irqs_disabled():0
Call Trace: <ffffffff8012374e>{__might_sleep+177} <ffffffff803cd7da>{mutex_lock+26}
Quoted 7 lines

<ffffffff8016b533>{kmem_cache_create+161} <ffffffff8063b0fb>{free_all_boo

in early boot. Please ignore.

- drivers/i2c/busses/scx200_acb.c doesn't compile on architectures which don't have asm/msr.h.

Seems good here so far (yet to fully test, -mm1 broke reiser too bad for me to test that very much).

Noted this when shutting down:

Starting killall: [ OK ]
Sending all processes the TERM signal...
Sending all processes the KILL signal...
Saving random seed:
Syncing hardware clock to system time
Turning off swap:
Unmounting pipe file systems:
Unmounting file systems:
Please stand by while rebooting the system...
md: stopping all md devices.
md: md1 switched to read-only mode.

BUG: unable to handle kernel NULL pointer dereference<6>md: md2 switched to read-only mode.
 at virtual address 0000001c
 printing eip:
b02a6951
*pde = 00000000
Oops: 0000 [#1]
SMP
last sysfs file: /devices/pci0000:00/0000:00:1f.3/i2c-0/0-002e/vrm
Modules linked in: iptable_mangle iptable_nat ip_nat ip_conntrack nfnetlink iptable_filter ip_tables nfsd exportfs lockd sunrpc ipv6 ip_gre binfmt_misc serio_raw piix hw_random
CPU:    0
EIP:    0060:[<b02a6951>]    Not tainted VLI
EFLAGS: 00010002   (2.6.16-rc1-mm2 #1)

EIP is at bitmap_daemon_work+0x144/0x391
eax: 0000001c ebx: b17fbc00 ecx: b17fbc00 edx: 00000286
esi: efdb7ec0 edi: 00000206 ebp: efc90e4c esp: efc90e24
ds: 007b es: 007b ss: 0068
Process md2_raid1 (pid: 382, threadinfo=efc90000 task=efca4030)

Stack: <0>00000000 efdb7eec 0003c9f0 b17fbc00 00000000 0003d1ef 00000020 efdc2340
       efdaec00 efc90000 efc90e8c b02a25a7 efc90e5c b0313488 efc90e68 0000001e
       b0456364 00000001 00000000 efc90eb4 b0115d87 b0456368 efc90f2c efdc2340
Call Trace:
 [<b0103bf5>] show_stack_log_lvl+0xc5/0xea
 [<b0103db7>] show_registers+0x19d/0x22b
 [<b0103f70>] die+0x12b/0x23b
 [<b01140d4>] do_page_fault+0x27a/0x5de
 [<b0103737>] error_code+0x4f/0x54
 [<b02a25a7>] md_check_recovery+0x1a/0x44a
 [<b029b645>] raid1d+0x2e/0xf55
 [<b02a32c4>] md_thread+0x44/0x14f
 [<b012ea27>] kthread+0xa5/0xca
 [<b0100d25>] kernel_thread_helper+0x5/0xb
Code: 00 83 45 e0 01 8b 4d e0 39 4e 1c 77 9b 8b 45 e4 85 c0 74 4f 8b 45 dc e8 9e cb 06 00 89 c2 8b 4d e4 8b 41 14 01 c0 01 c0 03 46 4c <8b> 08 f6 c1 04 0f 84 50 01 00 00 83 e1 fb 89 08 8b 45 dc e8 22
 <6>md: md3 switched to read-only mode.

md: md4 switched to read-only mode.
md: md5 switched to read-only mode.
md: md0 still in use.
Synchronizing SCSI cache for disk sdc:
Synchronizing SCSI cache for disk sdb:
Synchronizing SCSI cache for disk sda:
Restarting system.

reuben

#31 lore raw

Hi,

Quoted 1 line

From my perspective 2.6.16-rc1-mm2 still needs work. I did not try 15-mm1 or mm2. Both

mm3 and mm4 had X problems in that the system would lock but the keyboard was still active for Sysrq. The lockups took days to occur on both mm3 and mm4. The reiser3 problem made it impossible to test rc1-mm1, rc2-mm2 locked hard sometime in the first 4 hours of use - this time sysrq was dead too.

The system is a amd64 using x86_64 from the unofficial debian build. The box is stable using 15-rc5-mm3 which has had uptimes of over two weeks.

If anyone has ideas on what to backout let me know. Failing that I will boot with a serial console active and see that it reports.

Ideas,
Ed Tomlinson

#32 lore raw
Quoted 8 lines · Ed Tomlinson

Hi,

Quoted 1 line

From my perspective 2.6.16-rc1-mm2 still needs work. I did not try 15-mm1 or mm2. Both

mm3 and mm4 had X problems in that the system would lock but the keyboard was still active for Sysrq. The lockups took days to occur on both mm3 and mm4. The reiser3 problem made it impossible to test rc1-mm1, rc2-mm2 locked hard sometime in the first 4 hours of use - this time sysrq was dead too.

Quoted 8 lines

The system is a amd64 using x86_64 from the unofficial debian build. The box is stable using 15-rc5-mm3 which has had uptimes of over two weeks.

If anyone has ideas on what to backout let me know. Failing that I will boot with a serial console active and see that it reports.

Ideas,
Ed Tomlinson

Serial console shows that its an I/O error triggering a reiserfs4 kernel panic

[  559.544404] end_request: I/O error, dev sda, sector 19856555
[  559.554791] reiser4 panicked cowardly: reiser4[wget(6000)]: commit_current_atom (fs/reiser4/txnmgr.c:1130)[zam-597]:
[  559.554794] write log failed (-5)
[  559.554795] 
[  559.582807] Kernel panic - not syncing: reiser4[wget(6000)]: commit_current_atom (fs/reiser4/txnmgr.c:1130)[zam-597]:
[  559.582809] write log failed (-5)

Have some new errors started to be passed back thru the scsi / libata stack? I have had no problems with 15-rc5-mm3 though it may just be masking an issue...

some more info on this:

lspci -vvv from 2.6.15-rc5-mm3

0000:00:0a.0 IDE interface: nVidia Corporation CK8S Serial ATA Controller (v2.5) (rev a2) (prog-if 85 [Master SecO PriO])
Subsystem: Micro-Star International Co., Ltd.: Unknown device 0300
Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 0 (750ns min, 250ns max)
Interrupt: pin A routed to IRQ 17
Region 0: I/O ports at 09f0 [size=8]
Region 1: I/O ports at 0bf0 [size=4]
Region 2: I/O ports at 0970 [size=8]
Region 3: I/O ports at 0b70 [size=4]
Region 4: I/O ports at e000 [size=16]
Region 5: I/O ports at e400 [size=128]
Capabilities: [44] Power Management version 2
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-

Extract from the boot of 2.6.16-rc1-mm2

Jan 21 10:50:46 grover kernel: [   26.457745] SCSI subsystem initialized
Jan 21 10:50:46 grover kernel: [   26.495088] libata version 1.20 loaded.
Jan 21 10:50:46 grover kernel: [   26.498691] sata_nv 0000:00:09.0: version 0.8
Jan 21 10:50:46 grover kernel: [   26.499255] ACPI: PCI Interrupt Link [APSI] enabled at IRQ 23
Jan 21 10:50:46 grover kernel: [   26.518542] ACPI: PCI Interrupt 0000:00:09.0[A] -> Link [APSI] -> GSI 23 (level, high) -> IRQ 16
Jan 21 10:50:46 grover kernel: [   26.547990] PCI: Setting latency timer of device 0000:00:09.0 to 64
Jan 21 10:50:46 grover kernel: [   26.548042] ata1: SATA max UDMA/133 cmd 0x9E0 ctl 0xBE2 bmdma 0xC800 irq 16
Jan 21 10:50:46 grover kernel: [   26.570996] ata2: SATA max UDMA/133 cmd 0x960 ctl 0xB62 bmdma 0xC808 irq 16
Jan 21 10:50:46 grover kernel: [   26.669114] usb 2-4.1: new low speed USB device using ohci_hcd and address3
Jan 21 10:50:46 grover kernel: [   26.819069] usb 2-4.1: configuration #1 chosen from 1 choice
Jan 21 10:50:46 grover kernel: [   26.843314] ata1: SATA link up 1.5 Gbps (SStatus 113)
Jan 21 10:50:46 grover kernel: [   27.017383] nv_sata: Primary device added
Jan 21 10:50:46 grover kernel: [   27.030590] nv_sata: Primary device removed
Jan 21 10:50:46 grover kernel: [   27.044365] nv_sata: Secondary device removed
Jan 21 10:50:46 grover kernel: [   27.058744] ata1: dev 0 cfg 49:2f00 82:7c6b 83:7f09 84:4673 85:7c69 86:3e2187:4663 88:407f
Jan 21 10:50:46 grover kernel: [   27.058749] ata1: dev 0 ATA-7, max UDMA/133, 490234752 sectors: LBA48
Jan 21 10:50:46 grover kernel: [   27.101015] usb 2-4.2: new full speed USB device using ohci_hcd and address 4
Jan 21 10:50:46 grover kernel: [   27.125640] ata1: dev 0 configured for UDMA/133
Jan 21 10:50:46 grover kernel: [   27.151867] scsi0 : sata_nv
Jan 21 10:50:46 grover kernel: [   27.239985] usb 2-4.2: not running at top speed; connect to a high speed hub
Jan 21 10:50:46 grover kernel: [   27.276984] usb 2-4.2: configuration #1 chosen from 1 choice
Jan 21 10:50:46 grover kernel: [   27.298013] hub 2-4.2:1.0: USB hub found
Jan 21 10:50:46 grover kernel: [   27.311966] hub 2-4.2:1.0: 4 ports detected
Jan 21 10:50:46 grover kernel: [   27.395432] ata2: SATA link down (SStatus 0)
Jan 21 10:50:46 grover kernel: [   27.409957] scsi1 : sata_nv
Jan 21 10:50:46 grover kernel: [   27.423656]   Vendor: ATA       Model: Maxtor 6L250S0    Rev: BACE
Jan 21 10:50:46 grover kernel: [   27.445172]   Type:   Direct-Access                      ANSI SCSI revision: 05
Jan 21 10:50:46 grover kernel: [   27.469990] ACPI: PCI Interrupt Link [APSJ] enabled at IRQ 22
Jan 21 10:50:46 grover kernel: [   27.488973] ACPI: PCI Interrupt 0000:00:0a.0[A] -> Link [APSJ] -> GSI 22 (level, high) -> IRQ 17
Jan 21 10:50:46 grover kernel: [   27.518040] PCI: Setting latency timer of device 0000:00:0a.0 to 64
Jan 21 10:50:46 grover kernel: [   27.518079] ata3: SATA max UDMA/133 cmd 0x9F0 ctl 0xBF2 bmdma 0xE000 irq 17
Jan 21 10:50:46 grover kernel: [   27.541022] ata4: SATA max UDMA/133 cmd 0x970 ctl 0xB72 bmdma 0xE008 irq 17
Jan 21 10:50:46 grover kernel: [   27.768102] ata3: SATA link down (SStatus 0)
Jan 21 10:50:46 grover kernel: [   27.782177] scsi2 : sata_nv
Jan 21 10:50:46 grover kernel: [   27.793873] usb 2-4.2.1: new full speed USB device using ohci_hcd and address 5
Jan 21 10:50:46 grover kernel: [   27.968838] usb 2-4.2.1: configuration #1 chosen from 1 choice
Jan 21 10:50:46 grover kernel: [   28.007301] ata4: SATA link down (SStatus 0)
Jan 21 10:50:46 grover kernel: [   28.021888] scsi3 : sata_nv
Jan 21 10:50:46 grover kernel: [   28.031774] ACPI: PCI Interrupt Link [APC4] enabled at IRQ 19
Jan 21 10:50:46 grover kernel: [   28.050702] GSI 20 sharing vector 0xD1 and IRQ 20
Jan 21 10:50:46 grover kernel: [   28.066570] ACPI: PCI Interrupt Link [APCJ] enabled at IRQ 21
Jan 21 10:50:46 grover kernel: [   28.085505] ACPI: PCI Interrupt 0000:00:06.0[A] -> <6>ACPI: PCI Interrupt 0
000:02:0c.0[A] -> Link [APC4] -> Link [APCJ] -> GSI 21 (level, high) -> IRQ 18
Jan 21 10:50:46 grover kernel: [   28.132680] GSI 19 (level, low) -> IRQ 20
Jan 21 10:50:46 grover kernel: [   28.145889] PCI: Via IRQ fixup for 0000:02:0c.0, from 10 to 4
Jan 21 10:50:46 grover kernel: [   28.164961] PCI: Setting latency timer of device 0000:00:06.0 to 64
Jan 21 10:50:46 grover kernel: [   28.219223] ohci1394: fw-host0: OHCI-1394 1.0 (PCI): IRQ=[20]  MMIO=[ea000000-ea0007ff]  Max Packet=[2048]  IR/IT contexts=[4/8]
Jan 21 10:50:46 grover kernel: [   28.579166] intel8x0_measure_ac97_clock: measured 56004 usecs
Jan 21 10:50:46 grover kernel: [   28.598090] intel8x0: clocking to 48728
Jan 21 10:50:46 grover kernel: [   29.555763] ieee1394: Host added: ID:BUS[0-00:1023]  GUID[0010dc00006b07c2]
Jan 21 10:50:46 grover kernel: [   32.498302] SCSI device sda: 490234752 512-byte hdwr sectors (251000 MB)
Jan 21 10:50:46 grover kernel: [   32.583207] sda: Write Protect is off
Jan 21 10:50:46 grover kernel: [   32.595993] sda: Mode Sense: 00 3a 00 10
Jan 21 10:50:46 grover kernel: [   32.599979] SCSI device sda: drive cache: write back w/ FUA
Jan 21 10:50:46 grover kernel: [   32.672741] SCSI device sda: 490234752 512-byte hdwr sectors (251000 MB)
Jan 21 10:50:46 grover kernel: [   32.694879] sda: Write Protect is off
Jan 21 10:50:46 grover kernel: [   32.707334] sda: Mode Sense: 00 3a 00 10
Jan 21 10:50:46 grover kernel: [   32.707585] SCSI device sda: drive cache: write back w/ FUA
Jan 21 10:50:46 grover kernel: [   32.726082]  sda: sda1 sda2 sda3 sda4 < sda5 >
Jan 21 10:50:46 grover kernel: [   32.873995] sd 0:0:0:0: Attached scsi disk sda
Jan 21 10:50:46 grover kernel: [   33.541752] eth1394: eth1: IEEE-1394 IPv4 over 1394 Ethernet (fw-host0)
Jan 21 10:50:46 grover kernel: [   34.409131] usbcore: registered new driver hiddev
Jan 21 10:50:46 grover kernel: [   34.569628] input: Microsoft Microsoft IntelliMouse® Optical as /class/input/input1
Jan 21 10:50:46 grover kernel: [   34.594902] input: USB HID v1.00 Mouse [Microsoft Microsoft IntelliMouse® Optical] on usb-0000:00:02.0-4.1
Jan 21 10:50:46 grover kernel: [   34.627294] usbcore: registered new driver usbhid
Jan 21 10:50:46 grover kernel: [   34.642789] drivers/usb/input/hid-core.c: v2.6:USB HID core driver
Jan 21 10:50:46 grover kernel: [   34.816697] Bluetooth: Core ver 2.8
Jan 21 10:50:46 grover kernel: [   34.828192] NET: Registered protocol family 31
Jan 21 10:50:46 grover kernel: [   34.842990] Bluetooth: HCI device and connection manager initialized
Jan 21 10:50:46 grover kernel: [   34.863934] Bluetooth: HCI socket layer initialized
Jan 21 10:50:46 grover kernel: [   34.920716] Bluetooth: HCI USB driver ver 2.9
Jan 21 10:50:46 grover kernel: [   34.941494] usbcore: registered new driver hci_usb
Jan 21 10:50:46 grover kernel: [   36.538804] Adding 979956k swap on /dev/hda2.  Priority:-1 extents:1 across:979956k
Jan 21 10:50:46 grover kernel: [   36.564768] Adding 1020116k swap on /dev/sda2.  Priority:-2 extents:1 across:1020116k
Jan 21 10:50:46 grover kernel: [   38.361425] ieee1394: sbp2: Driver forced to serialize I/O (serialize_io=1)
Jan 21 10:50:46 grover kernel: [   38.384356] ieee1394: sbp2: Try serialize_io=0 for better performance
Jan 21 10:50:46 grover kernel: [   38.552288] Driver 'w83627hf' needs updating - please use bus_type methods
Jan 21 10:50:46 grover kernel: [   38.579591] w83627hf 9191-0290: Reading VID from GPIO5
Jan 21 10:50:46 grover kernel: [   38.703929] powernow-k8: Found 1 AMD Athlon 64 / Opteron processors (version 1.60.0)
Jan 21 10:50:46 grover kernel: [   38.733242] powernow-k8:    0 : fid 0xa (1800 MHz), vid 0x2 (1500 mV)
Jan 21 10:50:46 grover kernel: [   38.754467] powernow-k8:    1 : fid 0x2 (1000 MHz), vid 0x12 (1100 mV)
Jan 21 10:50:46 grover kernel: [   38.776181] cpu_init done, current fid 0xa, vid 0x2
Jan 21 10:50:46 grover kernel: [   38.826859] video1394: Installed video1394 module
Jan 21 10:50:46 grover kernel: [   38.863707] mice: PS/2 mouse device common for all mice
Jan 21 10:50:46 grover kernel: [   48.737454] kjournald starting.  Commit interval 5 seconds
Jan 21 10:50:46 grover kernel: [   48.766804] EXT3 FS on hda1, internal journal
Jan 21 10:50:46 grover kernel: [   48.781215] EXT3-fs: mounted filesystem with ordered data mode.
Jan 21 10:50:46 grover kernel: [   48.844247] ReiserFS: hda5: found reiserfs format "3.6" with standard journal
Jan 21 10:50:46 grover kernel: [   56.740820] ReiserFS: hda5: using ordered data mode
Jan 21 10:50:46 grover kernel: [   56.787279] ReiserFS: hda5: journal params: device hda5, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30
Jan 21 10:50:46 grover kernel: [   56.837125] ReiserFS: hda5: checking transaction log (hda5)
Jan 21 10:50:46 grover kernel: [   56.902508] ReiserFS: hda5: Using r5 hash to sort names
Jan 21 10:50:46 grover kernel: [   57.252649] Loading Reiser4. See www.namesys.com for a description of Reiser4 
grover:/var/log# sdparm -i /dev/sda
    /dev/sda: ATA       Maxtor 6L250S0    BACE
Device identification VPD page:
  Addressed logical unit:
    id_type: vendor specific [0x0],  code_set: ASCII
 00     4c 69 6e 75 78 20 41 54  41 2d 53 43 53 49 20 73    Linux ATA-SCSI s
 10     69 6d 75 6c 61 74 6f 72                             imulator
grover:/var/log# smartctl -i -d ata /dev/sda
smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===

Device Model:     Maxtor 6L250S0
Serial Number:    L50QDF3H
Firmware Version: BACE1G10
User Capacity:    251,000,193,024 bytes
Device is:        Not in smartctl database [for details use: -P showall]
ATA Version is:   7
ATA Standard is:  ATA/ATAPI-7 T13 1532D revision 0
Local Time is:    Sat Jan 21 11:27:21 2006 EST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled
grover:/var/log# smartctl -H -d ata /dev/sda
smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

---

Hope this helps and that I found the correct places to copy the info.

Ed Tomlinson

-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#33 lore raw
Quoted 1 line · Ed Tomlinson
grover:/var/log# smartctl -i -d ata /dev/sda

[snip]

Quoted 10 lines
grover:/var/log# smartctl -H -d ata /dev/sda
smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

---

Hope this helps and that I found the correct places to copy the info.

How about:
smartctl -a -d ata /dev/sda
or, if that produces too much output, then at least the following two:
smartctl -A -d ata /dev/sda
smartctl -l error -d ata /dev/sda

That way we might be able to figure out whether the disk coincidentally started going bad after you updated the kernel. --
-Barry K. Nathan <[email protected]>

#34 lore raw
Quoted 10 lines · Barry K. Nathan
Quoted 1 line · Ed Tomlinson
grover:/var/log# smartctl -i -d ata /dev/sda

[snip]

Quoted 7 lines
grover:/var/log# smartctl -H -d ata /dev/sda
smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

Quoted 6 lines
Quoted 3 lines

---

Hope this helps and that I found the correct places to copy the info.

How about:
smartctl -a -d ata /dev/sdagrover:/poola/home/ed# smartctl -a -d ata /dev/sda

smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===

Device Model:     Maxtor 6L250S0
Serial Number:    L50QDF3H
Firmware Version: BACE1G10
User Capacity:    251,000,193,024 bytes
Device is:        Not in smartctl database [for details use: -P showall]
ATA Version is:   7
ATA Standard is:  ATA/ATAPI-7 T13 1532D revision 0
Local Time is:    Sat Jan 21 16:34:26 2006 EST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

General SMART Values:
Offline data collection status: (0x80) Offline data collection activity
was never started.
Auto Offline Data Collection: Enabled.
Self-test execution status: ( 0) The previous self-test routine completed
without error or no self-test has ever been run.
Total time to complete Offline
data collection: (1922) seconds.
Offline data collection
capabilities: (0x5b) SMART execute Offline immediate.
Auto Offline data collection on/off support. Suspend Offline collection upon new command.
Offline surface scan supported. Self-test supported.
No Conveyance Self-test supported. Selective Self-test supported.
SMART capabilities: (0x0003) Saves SMART data before entering
power-saving mode.
Supports SMART auto save timer.
Error logging capability: (0x01) Error logging supported.
General Purpose Logging supported.
Short self-test routine
recommended polling time: ( 2) minutes.
Extended self-test routine
recommended polling time: ( 99) minutes.

SMART Attributes Data Structure revision number: 16 Vendor Specific SMART Attributes with Thresholds:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  3 Spin_Up_Time            0x0027   252   252   063    Pre-fail  Always       -       571
  4 Start_Stop_Count        0x0032   253   253   000    Old_age   Always       -       2
  5 Reallocated_Sector_Ct   0x0033   253   253   063    Pre-fail  Always       -       0
  6 Read_Channel_Margin     0x0001   253   253   100    Pre-fail  Offline      -       0
  7 Seek_Error_Rate         0x000a   253   252   000    Old_age   Always       -       0
  8 Seek_Time_Performance   0x0027   250   240   187    Pre-fail  Always       -       49844
  9 Power_On_Hours          0x0032   251   251   000    Old_age   Always       -       49644
 10 Spin_Retry_Count        0x002b   252   252   157    Pre-fail  Always       -       0
 11 Calibration_Retry_Count 0x002b   252   252   223    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   253   253   000    Old_age   Always       -       4
192 Power-Off_Retract_Count 0x0032   253   253   000    Old_age   Always       -       0
193 Load_Cycle_Count        0x0032   253   253   000    Old_age   Always       -       0
194 Temperature_Celsius     0x0032   028   253   000    Old_age   Always       -       29
195 Hardware_ECC_Recovered  0x000a   253   252   000    Old_age   Always       -       8656
196 Reallocated_Event_Count 0x0008   253   253   000    Old_age   Offline      -       0
197 Current_Pending_Sector  0x0008   253   253   000    Old_age   Offline      -       0
198 Offline_Uncorrectable   0x0008   253   253   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x0008   199   199   000    Old_age   Offline      -       0
200 Multi_Zone_Error_Rate   0x000a   253   252   000    Old_age   Always       -       0
201 Soft_Read_Error_Rate    0x000a   253   252   000    Old_age   Always       -       1
202 TA_Increase_Count       0x000a   253   252   000    Old_age   Always       -       0
203 Run_Out_Cancel          0x000b   253   252   180    Pre-fail  Always       -       0
204 Shock_Count_Write_Opern 0x000a   253   252   000    Old_age   Always       -       0
205 Shock_Rate_Write_Opern  0x000a   253   252   000    Old_age   Always       -       0
207 Spin_High_Current       0x002a   252   252   000    Old_age   Always       -       0
208 Spin_Buzz               0x002a   252   252   000    Old_age   Always       -       0
209 Offline_Seek_Performnce 0x0024   242   242   000    Old_age   Offline      -       143
210 Unknown_Attribute       0x0032   253   252   000    Old_age   Always       -       0
211 Unknown_Attribute       0x0032   253   252   000    Old_age   Always       -       0
212 Unknown_Attribute       0x0032   253   252   000    Old_age   Always       -       0

SMART Error Log Version: 1
No Errors Logged

SMART Self-test log structure revision number 1
No self-tests have been logged. [To run self-tests, use: smartctl -t]

SMART Selective self-test log data structure revision number 1

 SPAN  MIN_LBA  MAX_LBA  CURRENT_TEST_STATUS
    1        0        0  Not_testing
    2        0        0  Not_testing
    3        0        0  Not_testing
    4        0        0  Not_testing
    5        0        0  Not_testing

Selective self-test flags (0x0):
After scanning selected spans, do NOT read-scan remainder of disk.
If Selective self-test is pending on power-up, resume after 0 minute delay.

Quoted 3 lines

or, if that produces too much output, then at least the following two:
smartctl -A -d ata /dev/sda
smartctl -l error -d ata /dev/sda

grover:/poola/home/ed# smartctl -l error -d ata /dev/sda
smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART Error Log Version: 1
No Errors Logged

Quoted 2 lines

That way we might be able to figure out whether the disk coincidentally started going bad after you updated the kernel.

I suspect the newer kernel (or kernels) since when I revert to 15-rc5-mm3 all is well.

Thanks,
Ed Tomlinson

#35 lore raw

On 1/21/06, Ed Tomlinson <[email protected]> wrote:
[snip]

Quoted 2 lines

smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen Home page is http://smartmontools.sourceforge.net/

[snip]

Quoted 2 lines

Error logging capability: (0x01) Error logging supported.
General Purpose Logging supported.

[snip]

Quoted 3 lines

SMART Error Log Version: 1
No Errors Logged

[snip]

Quoted 1 line

I suspect the newer kernel (or kernels) since when I revert to 15-rc5-mm3 all is well.

That's what it looks like to me, too. Weird...

--
-Barry K. Nathan <[email protected]>

#36 lore raw
In reply to Ed Tomlinson

Summarizing all this. There are two problems here.

1. reserifs4 panics when it gets io errors - I remember this was an issue that needed to be fixed in the R4 code before it moves to mainline...

2. Why does a drive which is fine with 2.6.15-rc5-mm3, return a -5 with 2.6.16-mm3 and above? Smart reports no problems with the drive hardware. What has changed in the libata/scsi stacks?

Thanks,
Ed Tomlinson

Quoted 142 lines · Ed Tomlinson
Quoted 10 lines · Barry K. Nathan
Quoted 1 line · Ed Tomlinson
grover:/var/log# smartctl -i -d ata /dev/sda

[snip]

Quoted 7 lines
grover:/var/log# smartctl -H -d ata /dev/sda
smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

Quoted 6 lines
Quoted 3 lines

---

Hope this helps and that I found the correct places to copy the info.

How about:
smartctl -a -d ata /dev/sdagrover:/poola/home/ed# smartctl -a -d ata /dev/sda

smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===

Device Model:     Maxtor 6L250S0
Serial Number:    L50QDF3H
Firmware Version: BACE1G10
User Capacity:    251,000,193,024 bytes
Device is:        Not in smartctl database [for details use: -P showall]
ATA Version is:   7
ATA Standard is:  ATA/ATAPI-7 T13 1532D revision 0
Local Time is:    Sat Jan 21 16:34:26 2006 EST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

General SMART Values:
Offline data collection status: (0x80) Offline data collection activity
was never started.
Auto Offline Data Collection: Enabled.
Self-test execution status: ( 0) The previous self-test routine completed
without error or no self-test has ever been run.
Total time to complete Offline
data collection: (1922) seconds.
Offline data collection
capabilities: (0x5b) SMART execute Offline immediate.
Auto Offline data collection on/off support. Suspend Offline collection upon new command.
Offline surface scan supported. Self-test supported.
No Conveyance Self-test supported. Selective Self-test supported.
SMART capabilities: (0x0003) Saves SMART data before entering
power-saving mode.
Supports SMART auto save timer.
Error logging capability: (0x01) Error logging supported.
General Purpose Logging supported.
Short self-test routine
recommended polling time: ( 2) minutes.
Extended self-test routine
recommended polling time: ( 99) minutes.

SMART Attributes Data Structure revision number: 16 Vendor Specific SMART Attributes with Thresholds:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  3 Spin_Up_Time            0x0027   252   252   063    Pre-fail  Always       -       571
  4 Start_Stop_Count        0x0032   253   253   000    Old_age   Always       -       2
  5 Reallocated_Sector_Ct   0x0033   253   253   063    Pre-fail  Always       -       0
  6 Read_Channel_Margin     0x0001   253   253   100    Pre-fail  Offline      -       0
  7 Seek_Error_Rate         0x000a   253   252   000    Old_age   Always       -       0
  8 Seek_Time_Performance   0x0027   250   240   187    Pre-fail  Always       -       49844
  9 Power_On_Hours          0x0032   251   251   000    Old_age   Always       -       49644
 10 Spin_Retry_Count        0x002b   252   252   157    Pre-fail  Always       -       0
 11 Calibration_Retry_Count 0x002b   252   252   223    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   253   253   000    Old_age   Always       -       4
192 Power-Off_Retract_Count 0x0032   253   253   000    Old_age   Always       -       0
193 Load_Cycle_Count        0x0032   253   253   000    Old_age   Always       -       0
194 Temperature_Celsius     0x0032   028   253   000    Old_age   Always       -       29
195 Hardware_ECC_Recovered  0x000a   253   252   000    Old_age   Always       -       8656
196 Reallocated_Event_Count 0x0008   253   253   000    Old_age   Offline      -       0
197 Current_Pending_Sector  0x0008   253   253   000    Old_age   Offline      -       0
198 Offline_Uncorrectable   0x0008   253   253   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x0008   199   199   000    Old_age   Offline      -       0
200 Multi_Zone_Error_Rate   0x000a   253   252   000    Old_age   Always       -       0
201 Soft_Read_Error_Rate    0x000a   253   252   000    Old_age   Always       -       1
202 TA_Increase_Count       0x000a   253   252   000    Old_age   Always       -       0
203 Run_Out_Cancel          0x000b   253   252   180    Pre-fail  Always       -       0
204 Shock_Count_Write_Opern 0x000a   253   252   000    Old_age   Always       -       0
205 Shock_Rate_Write_Opern  0x000a   253   252   000    Old_age   Always       -       0
207 Spin_High_Current       0x002a   252   252   000    Old_age   Always       -       0
208 Spin_Buzz               0x002a   252   252   000    Old_age   Always       -       0
209 Offline_Seek_Performnce 0x0024   242   242   000    Old_age   Offline      -       143
210 Unknown_Attribute       0x0032   253   252   000    Old_age   Always       -       0
211 Unknown_Attribute       0x0032   253   252   000    Old_age   Always       -       0
212 Unknown_Attribute       0x0032   253   252   000    Old_age   Always       -       0

SMART Error Log Version: 1
No Errors Logged

SMART Self-test log structure revision number 1
No self-tests have been logged. [To run self-tests, use: smartctl -t]

SMART Selective self-test log data structure revision number 1

 SPAN  MIN_LBA  MAX_LBA  CURRENT_TEST_STATUS
    1        0        0  Not_testing
    2        0        0  Not_testing
    3        0        0  Not_testing
    4        0        0  Not_testing
    5        0        0  Not_testing

Selective self-test flags (0x0):
After scanning selected spans, do NOT read-scan remainder of disk.
If Selective self-test is pending on power-up, resume after 0 minute delay.

Quoted 3 lines

or, if that produces too much output, then at least the following two:
smartctl -A -d ata /dev/sda
smartctl -l error -d ata /dev/sda

grover:/poola/home/ed# smartctl -l error -d ata /dev/sda
smartctl version 5.34 [x86_64-unknown-linux-gnu] Copyright (C) 2002-5 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART Error Log Version: 1
No Errors Logged

Quoted 2 lines

That way we might be able to figure out whether the disk coincidentally started going bad after you updated the kernel.

I suspect the newer kernel (or kernels) since when I revert to 15-rc5-mm3 all is well.

Thanks,
Ed Tomlinson
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#37 lore raw
Quoted 7 lines · Ed Tomlinson

Summarizing all this. There are two problems here.

1. reserifs4 panics when it gets io errors - I remember this was an issue that needed to be fixed in the R4 code before it moves to mainline...

2. Why does a drive which is fine with 2.6.15-rc5-mm3, return a -5 with 2.6.16-mm3
and above? Smart reports no problems with the drive hardware. What has changed in the libata/scsi stacks?

That's a long answer. Could you assist in narrowing down the versions which are affected?

It would also be useful if you could try vanilla kernels, and help us discover whether problems surfaces in 2.6.15, 2.6.15-git[1234], 2.6.16-rc1, etc.

Jeff

A message that is not in the archive (sent off-list or not yet received): [email protected]

#39 lore raw

Re: 2.6.16-rc1-mm2 (mm5 too) panics

Hi,

I need some help figuring this one out. I cannot use git bisect and applying the 2.6.15-1 reiserfs4 patch does not work with newer 2.6.16 levels - looks like the mutex conversion hits it hard. Looking in mm there are about 50 reiser4 patches...

To reproduce this problem I need reiser4 though the issue maybe in libata or scsi. Does anyone have an idea how I can proceed to find what makes newer kernels get io errors which trigger a reiser4
panic? The panic is a zam-597 (fs/reiser4/txnmgr.c) with an rc=-5

Anyone have a git tree tracking linus with reiser4 patches applied staring before 2.6.15 - if so git bisect could be used to find the change causing the problem.

One datapoint. In 2.6.15 + reiser4 2.6.15-1 works fine - the resiser4 filesystem is heavily used with no errors. Smart report no problems with the drive being used. The libata driver used is sata_nv.

Please reply to my email - I am only subscribed to lkml.

Help!
Ed Tomlinson

ret = reiser4_write_logs(nr_submitted);
if (ret < 0)

    reiser4_panic("zam-597", "write log failed (%ld)\n", ret);
Quoted 27 lines · Ed Tomlinson
Quoted 10 lines · you
Quoted 8 lines · Ed Tomlinson

Summarizing all this. There are two problems here.

1. reserifs4 panics when it gets io errors - I remember this was an issue that needed to be fixed in the R4 code before it moves to mainline...

2. Why does a drive which is fine with 2.6.15-rc5-mm3, return a -5 with 2.6.16-mm3 and above? Smart reports no problems with the drive hardware. What has changed in the libata/scsi stacks?

Quoted 6 lines

That's a long answer. Could you assist in narrowing down the versions which are affected?

It would also be useful if you could try vanilla kernels, and help us discover whether problems surfaces in 2.6.15, 2.6.15-git[1234], 2.6.16-rc1, etc.

Jeff,

I'll see what I can do with git bisect. Given that reserifs4 is in the picture this may be
fun... I expect it will be a slow process (kernels take 40min to build here).

Thanks
Ed Tomlinson

#40 lore raw

2.6.16-rc4-mm1 & 2.6.16-rc1-mm2 (mm5 too) panics

Hi,

2.6.16-rc4-mm1 panics with a zam-597 (fs/reiser4/txnmgr.c) with an rc=-5

This has been happening since 2.6.16-rc1-mm2. It makes testing with reiserfs4 impossible here.

I have not been able to figure out how to use bisect etc to find out what is triggering this. I had hoped that reversing "jbd: split checkpoint lists" as per the fix for ocfs2 might help (its in rc4-mm1) but no such luck.

Does anyone have a suggestion on how I can help find what is broken here?

TIA,
Ed Tomlinson

Quoted 53 lines · Ed Tomlinson

I need some help figuring this one out. I cannot use git bisect and applying the 2.6.15-1 reiserfs4 patch does not work with newer 2.6.16 levels - looks like the mutex conversion hits it hard. Looking in mm there are about 50 reiser4 patches...

To reproduce this problem I need reiser4 though the issue maybe in libata or scsi. Does anyone have an idea how I can proceed to find what makes newer kernels get io errors which trigger a reiser4
panic? The panic is a zam-597 (fs/reiser4/txnmgr.c) with an rc=-5

Anyone have a git tree tracking linus with reiser4 patches applied staring before 2.6.15 - if so git bisect could be used to find the change causing the problem.

One datapoint. In 2.6.15 + reiser4 2.6.15-1 works fine - the resiser4 filesystem is heavily used with no errors. Smart report no problems with the drive being used. The libata driver used is sata_nv.

Please reply to my email - I am only subscribed to lkml.

Help!
Ed Tomlinson

ret = reiser4_write_logs(nr_submitted);
if (ret < 0)

    reiser4_panic("zam-597", "write log failed (%ld)\n", ret);
Quoted 27 lines · Ed Tomlinson
Quoted 10 lines · you
Quoted 8 lines · Ed Tomlinson

Summarizing all this. There are two problems here.

1. reserifs4 panics when it gets io errors - I remember this was an issue that needed to be fixed in the R4 code before it moves to mainline...

2. Why does a drive which is fine with 2.6.15-rc5-mm3, return a -5 with 2.6.16-mm3 and above? Smart reports no problems with the drive hardware. What has changed in the libata/scsi stacks?

Quoted 6 lines

That's a long answer. Could you assist in narrowing down the versions which are affected?

It would also be useful if you could try vanilla kernels, and help us discover whether problems surfaces in 2.6.15, 2.6.15-git[1234], 2.6.16-rc1, etc.

Jeff,

I'll see what I can do with git bisect. Given that reserifs4 is in the picture this may be
fun... I expect it will be a slow process (kernels take 40min to build here).

Thanks
Ed Tomlinson

#41 lore raw
In reply to Reuben Farrelly
Quoted 17 lines · On Saturday January 21, [email protected] wrote:

md: stopping all md devices.
md: md1 switched to read-only mode.

BUG: unable to handle kernel NULL pointer dereference<6>md: md2 switched to 

read-only mode.
at virtual address 0000001c
printing eip:
b02a6951
*pde = 00000000

Oops: 0000 [#1]
SMP
last sysfs file: /devices/pci0000:00/0000:00:1f.3/i2c-0/0-002e/vrm
Modules linked in: iptable_mangle iptable_nat ip_nat ip_conntrack nfnetlink 
iptable_filter ip_tables nfsd exportfs lockd sunrpc ipv6 ip_gre binfmt_misc 
serio_raw piix hw_random
CPU:    0
EIP:    0060:[<b02a6951>]    Not tainted VLI
EFLAGS: 00010002   (2.6.16-rc1-mm2 #1)
Quoted 1 line

EIP is at bitmap_daemon_work+0x144/0x391

Hmmmm.... yep. I see the problem. We shouldn't be tearing down the bitmap when switching to read-only. Patch to follow.

Thanks,

NeilBrown

#42 lore raw
In reply to Andrew Morton
Quoted 2 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.16-rc1/2.6.16-rc1-mm2/

I have a problem on powerpc from zone_reclaim-reclaim-on-memory-only-node-support.patch:

mm/vmscan.c: In function `zone_reclaim':
mm/vmscan.c:1845: error: invalid lvalue in unary `&'

The preprocessed output is:

if (!(gfp_mask & (( gfp_t)0x10u)) || (!__cpus_empty(&(node_to_cpumask(zone->zone_pgdat->node_id)), 128) &&
...

I don't understand why this wouldn't die on every architecture, since node_to_cpumask is an inline function.

(Using gcc version 3.4.4 20050721 (Red Hat 3.4.4-2))

thanks,
-serge

#43 lore raw
Quoted 2 lines · Serge E. Hallyn

I don't understand why this wouldn't die on every architecture, since node_to_cpumask is an inline function.

Its an array lookup on ia64.

#44 lore raw
Quoted 6 lines · Christoph Lameter
Quoted 2 lines · Serge E. Hallyn

I don't understand why this wouldn't die on every architecture, since node_to_cpumask is an inline function.

Its an array lookup on ia64.

Oh I see, sorry, I was looking at only partial lxr ouput.

Is the following patch an ok fix?

thanks
-serge

On alpha, powerpc, and i386, node_to_cpumask is an inline function rather than a #define to an array lookup.

--

mm/vmscan.c +3 -1
Index: linux-2.6.15/mm/vmscan.c
===================================================================
@@ -1836,13 +1836,15 @@ int zone_reclaim(struct zone *zone, gfp_
 	struct task_struct *p = current;
 	struct reclaim_state reclaim_state;
 	struct scan_control sc;
+	cpumask_t mask;
 
 	if (time_before(jiffies,
 		zone->last_unsuccessful_zone_reclaim + ZONE_RECLAIM_INTERVAL))
 			return 0;
 
+	mask = node_to_cpumask(zone->zone_pgdat->node_id);
 	if (!(gfp_mask & __GFP_WAIT) ||
-		(!cpus_empty(node_to_cpumask(zone->zone_pgdat->node_id)) &&
+		(!cpus_empty(mask) &&
 			 zone->zone_pgdat->node_id != numa_node_id()) ||
 		zone->all_unreclaimable ||
 		atomic_read(&zone->reclaim_in_progress) > 0)
#45 lore raw
Quoted 1 line · Serge E. Hallyn

Is the following patch an ok fix?

It is already fixed along the same line that you are proposing.