Skip to contentSkip to conversation

2.6.17-rc5-mm3

PATCH New Andrew Morton 16 participants 71 messages Jun 4, 2006 – Jun 7, 2006 1 message not in the archive

Subsystems (6)
lore
Share & tools
lore

Andrew Morton

author Applied
#0 lore raw

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

Boilerplate:

- See the `hot-fixes' directory for any important updates to this patchset.

- To fetch an -mm tree using git, use (for example)

git fetch git://git.kernel.org/pub/scm/linux/kernel/git/smurf/linux-trees.git v2.6.16-rc2-mm1

- -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 is 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 2.6.17-rc5-mm2:

origin.patch
git-acpi.patch
git-agpgart.patch
git-alsa.patch
git-audit-master.patch
git-block.patch
git-cifs.patch
git-cpufreq.patch
git-cpufreq-fixup.patch
git-dvb.patch
git-gfs2.patch
git-ia64.patch
git-infiniband.patch
git-intelfb.patch
git-klibc.patch
git-hdrcleanup.patch
git-hdrinstall.patch
git-libata-all.patch
git-mips.patch
git-mtd.patch
git-netdev-all.patch
git-net.patch
git-nfs.patch
git-powerpc.patch
git-rbtree.patch
git-sas.patch
git-pcmcia.patch
git-scsi-target.patch
git-supertrak.patch
git-watchdog.patch
git-cryptodev.patch

git trees

-drivers-usb-core-devioc-dereference-userspace-pointer.patch -nbd-endian-annotations.patch
-cifs-build-fix.patch
-git-cifs-kconfig-fix.patch -cifs-do-not-overwrite-aops-elements.patch -scx200_acb-use-pci-i-o-resource-when-appropriate-fix.patch -git-mtd-cs553x_nand-build-fix.patch -pmf_register_irq_client-gives-sleep-with-locks-held-warning.patch -64-bit-resources-arch-powerpc-changes-update.patch -fix-pciehp-compile-issue-when-config_acpi-is-not.patch -gregkh-pci-pci-64-bit-resources-drivers-others-changes-amba-fix.patch -i386-export-memory-more-than-4g-through-proc-iomem.patch -pci-pci-64-bit-resources-arch-changes-update.patch -improve-pci-config-space-writeback.patch -reverse-pci-config-space-restore-order.patch -pci-add-pci_assign_resource_fixed-allow-fixed-address.patch -add-a-enable-sysfs-attribute-to-the-pci-devices-to-allow.patch -fix-recovery-path-from-errors-during-pcie_init.patch -move-various-pci-ids-to-header-file.patch -kconfigurable-resources-core-changes.patch -kconfigurable-resources-core-changes-i386-fix.patch -kconfigurable-resources-core-changes-fix.patch -kconfigurable-resources-driver-pci-changes.patch -kconfigurable-resources-driver-others-changes.patch -kconfigurable-resources-arch-dependent-changes-arch-a-i.patch -kconfigurable-resources-arch-dependent-changes-arch-a-i-fix.patch -kconfigurable-resources-arch-dependent-changes-arch-j-p.patch -kconfigurable-resources-arch-dependent-changes-arch-q-z.patch -typesh-sector_t-and-blkcnt_t-arent-for-userspace.patch -allow-msi-to-work-on-kexec-kernel.patch -pci-disable-msi-mode-in-pci_disable_device.patch -pci-dont-move-ioapics-below-pci-bridge.patch -git-scsi-rc-fixes.patch -scsi-properly-count-the-number-of-pages-in-scsi_req_map_sg-fix.patch -revert-gregkh-usb-usb-ohci-avoids-root-hub-timer-polling.patch -gregkh-usb-usb-serial-mos7720-powerpc-wrokaround.patch -usb-add-sierra-wireless-mc5720-id-to-airprimec.patch -usb-negative-index-in-drivers-usb-host-isp116x-hcdc.patch -xfs-sparc32-build-fix.patch
-add-pci_cap_id_vndr.patch

Merged into mainline or a subsystem tree

+alpha-smp-irq-routing-fix.patch +fs-nameic-call-to-file_permission-under-a-spinlock-in-do_lookup_path.patch +fs-nameic-call-to-file_permission-under-a-spinlock-in-do_lookup_path-fix.patch +pmf_register_irq_client-gives-sleep-with-locks-held-warning.patch +implement-get--set-tso-for-forcedeth-driver.patch +fix-hpet-operation-on-32-bit-nvidia-platforms.patch +fix-hpet-operation-on-32-bit-nvidia-platforms-build-fix.patch +fix-hpet-operation-on-64-bit-nvidia-platforms.patch +maintainers-add-entries-for-bnx2-and-tg3.patch +sbp2-fix-check-of-return-value-of.patch +sata_sil24-sii3124-sata-driver-endian-problem.patch +m48t86-ia64-build-fix.patch
+m68k-get_user-build-fix.patch
+uml-add-asm-irqflagsh.patch +uml-fix-wall_to_monotonic-initialization.patch +uml-fix-a-typo-in-do_uml_initcalls.patch +uml-__user-annotation-in-arch_prctl.patch +uml-more-__user-annotations.patch +uml-add-ffreestanding-to-cflags.patch

2.6.17 queue

+kevent-add-new-uevent.patch

Required for acpi-dock-driver.patch

-acpi-dock-driver-v3.patch
-acpi-dock-driver-v4.patch
-acpi-dock-driver-interface-fixups.patch

Folded into acpi-dock-driver.patch

-acpiphp-use-new-dock-driver-fix.patch
-acpiphp-use-new-dock-driver-v2.patch

Folded into acpiphp-use-new-dock-driver.patch

-acpi-atlas-acpi-driver-v2-tidy.patch

Folded into acpi-atlas-acpi-driver.patch

+acpi-atlas-acpi-driver-fix.patch

Fix acpi-atlas-acpi-driver.patch

+ieee1394-video1394-be-quiet.patch +ieee1394-ohci1394c-function-calls-without.patch +ieee1394-sbp2-make-tsb42aa9-workaround-specific.patch +ieee1394-semaphore-to-mutex-conversion.patch +ieee1394-raw1394-fix-whitespace-after-x86_64.patch +ieee1394-ieee1394-ohci1394-cycletoolong.patch +ieee1394-ieee1394-support-for-slow-links-or-slow.patch +ieee1394-ieee1394-save-ram-by-using-a-single.patch +ieee1394-sbp2-remove-manipulation-of-inquiry.patch +ieee1394-sbp2-log-number-of-supported-concurrent.patch +ieee1394-ieee1394-extend-lowlevel-api-for.patch +ieee1394-ohci1394-set-address-range-properties.patch +ieee1394-ohci1394-make-phys_dma-parameter.patch +ieee1394-sbp2-sbp2-remove-ohci1394-specific.patch +ieee1394-sbp2-fix-s800-transfers-if-phys_dma-is.patch +ieee1394-update-feature-removal-of-obsolete.patch +ieee1394-sbp2-provide-helptext-for.patch +ieee1394-sbp2-kconfig-fix.patch +ieee1394-sbp2-use-__attribute__packed-for.patch +ieee1394-sbp2-fix-deregistration-of-status-fifo-address-space.patch +ieee1394-add-preprocessor-constant-for-invalid-csr.patch +eth1394-endian-fixes.patch

ieee1394 updates

-ieee1394_core-switch-to-kthread-api-fix.patch

Folded into ieee1394_core-switch-to-kthread-api.patch

-input-fix-oops-on-mk712-load.patch

Dropped

-via-pmu-add-input-device-tidy.patch

Folded into via-pmu-add-input-device.patch

-input-powermac-cleanup-of-mac_hid-and-support-for-ctrlclick-and-commandclick-update.patch

Folded into input-powermac-cleanup-of-mac_hid-and-support-for-ctrlclick-and-commandclick.patch

-input-logitech-trackman-trackball-support.patch

Dropped (I think)

-input-new-force-feedback-interface-fix.patch

Folded into input-new-force-feedback-interface.patch

+kconfig-integrate-split-config-into-silentoldconfig-fix.patch

Folded into kconfig-integrate-split-config-into-silentoldconfig.patch

+kbuild-obj-dirs-is-calculated-incorrectly-if-hostprogs-y-is-defined.patch +fix-make-rpm-for-powerpc.patch

kbuild fixes

+revert-sata_sil24-sii3124-sata-driver-endian-problem.patch

Revert earlier patch so that git-libata-all applies OK.

+libata-add-missing-data_xfer-for-pata_pdc2027x-and-pdc_adma-fix.patch

Fix libata-add-missing-data_xfer-for-pata_pdc2027x-and-pdc_adma.patch

+prevent-au1xmmcc-breakage-on-non-au1200-alchemy.patch

mmc driver fix

+myri10ge-alpha-build-fix.patch

Fix git-netdev-all.patch

+forcedeth-config-ring-sizes.patch
+forcedeth-config-flow-control.patch
+forcedeth-config-phy.patch
+forcedeth-config-wol.patch
+forcedeth-config-csum.patch
+forcedeth-config-statistics.patch
+forcedeth-config-diagnostics.patch
+forcedeth-config-module-parameters.patch
+forcedeth-config-version.patch
+forcedeth-new-device-ids.patch
+forcedeth-typecast-cleanup.patch

forcedeth updates

+lock-validator-netlinkc-netlink_table_grab-fix.patch

netlink locking fix

+recent-match-fix-sleeping-function-called-from-invalid-context.patch +recent-match-missing-refcnt-initialization.patch

netfilter fixes

-fix-for-serial-uart-lockup.patch

Dropped.

+gregkh-pci-pci-add-pci_cap_id_vndr.patch +gregkh-pci-pci-fix-pciehp-compile-issue-when-config_acpi-is-not-enabled.patch +gregkh-pci-pci-64-bit-resource-fixup-pci-resource-dbg-code-to-handle-size-change.patch +gregkh-pci-pci-64-bit-resource-fix-amba-build-warning.patch +gregkh-pci-pci-64-bit-resources-fix-pnp-sysfs-interface.patch +gregkh-pci-pci-64-bit-resources-arch-powerpc-changes-update.patch +gregkh-pci-kconfigurable-resources-core-changes.patch
+gregkh-pci-kconfigurable-resources-driver-pci-changes.patch +gregkh-pci-kconfigurable-resources-driver-others-changes.patch +gregkh-pci-kconfigurable-resources-arch-dependent-changes.patch +gregkh-pci-kconfigurable-resources-arch-dependent-changes-arch.patch +gregkh-pci-kconfigurable-resources-arch-dependent-changes-arch-q-z.patch +gregkh-pci-i386-export-memory-more-than-4g-through-proc-iomem.patch +gregkh-pci-pci-error-handling-on-pci-device-resume.patch +gregkh-pci-pci-ignore-pre-set-64-bit-bars-on-32-bit-platforms.patch +gregkh-pci-pciehp-dont-call-pci_enable_dev.patch
+gregkh-pci-pci-improve-pci-config-space-writeback.patch +gregkh-pci-pci-reverse-pci-config-space-restore-order.patch +gregkh-pci-pci-add-pci_assign_resource_fixed-allow-fixed-address-assignments.patch +gregkh-pci-pci-add-a-enable-sysfs-attribute-to-the-pci-devices-to-allow-userspace-to-enable-devices-without-doing-foul-direct-access.patch +gregkh-pci-pci-don-t-enable-device-if-already-enabled.patch +gregkh-pci-pci-acpi-rename-the-functions-to-avoid-multiple-instances.patch +gregkh-pci-acpi_pcihp-fix-programming-_hpp-values.patch +gregkh-pci-acpi_pcihp-remove-improper-error-message-about-oshp.patch +gregkh-pci-acpi_pcihp-add-support-for-_hpx.patch
+gregkh-pci-pciehp-fix-programming-hotplug-parameters.patch
+gregkh-pci-shpc-cleanup-shpc-register-access.patch
+gregkh-pci-shpc-cleanup-shpc-logical-slot-register-access.patch +gregkh-pci-shpc-cleanup-shpc-logical-slot-register-bits-access.patch +gregkh-pci-shpc-fix-shpc-logical-slot-register-bits-access.patch +gregkh-pci-shpc-fix-shpc-contoller-serr-int-register-bits-access.patch +gregkh-pci-shpchp-mask-global-serr-and-intr-at-controller-release-time.patch +gregkh-pci-shpchp-create-shpchpd-at-controller-probe-time.patch +gregkh-pci-pci-i386-x86_84-disable-pci-resource-decode-on-device-disable.patch +gregkh-pci-sgi-hotplug-incorrect-power-status.patch +gregkh-pci-pci-bus-parity-status-broken-hardware-attribute-edac-foundation.patch +gregkh-pci-pci-hotplug-fix-recovery-path-from-errors-during-pcie_init.patch +gregkh-pci-pciehp-replace-pci_find_slot-with-pci_get_slot.patch
+gregkh-pci-pciehp-add-missing-pci_dev_put.patch
+gregkh-pci-pciehp-implement-get_address-callback.patch +gregkh-pci-shpchp-remove-unnecessary-hpc_ctlr_handle-check.patch
+gregkh-pci-shpchp-cleanup-interrupt-handler.patch
+gregkh-pci-shpchp-cleanup-shpc-commands.patch
+gregkh-pci-shpchp-cleanup-interrupt-polling-timer.patch
+gregkh-pci-shpchp-remove-unused-hpc_evelnt_lock.patch
+gregkh-pci-shpchp-cleanup-improper-info-messages.patch
+gregkh-pci-pci-move-various-pci-ids-to-header-file.patch +gregkh-pci-pci-amd-8131-msi-quirk-called-too-late-bus_flags-not-inherited.patch +gregkh-pci-pci-allow-msi-to-work-on-kexec-kernel.patch
+gregkh-pci-pci-disable-msi-mode-in-pci_disable_device.patch +gregkh-pci-pci-hotplug-fake-null-pointer-dereferences-in-ibm-hot-plug-controller-driver.patch +gregkh-pci-pci-cleanup-unused-variable-about-msi-driver.patch +gregkh-pci-pci-don-t-move-ioapics-below-pci-bridge.patch
+gregkh-pci-pci-remove-unneeded-msi-code.patch
+gregkh-pci-pci-clean-up-pci-documentation-to-be-more-specific.patch +gregkh-pci-pci-fix-race-with-pci_walk_bus-and-pci_destroy_dev.patch +gregkh-pci-pci-test-that-drivers-properly-call-pci_set_master.patch

PCI tree updates

+revert-gregkh-pci-pci-test-that-drivers-properly-call-pci_set_master.patch +gregkh-pci-kconfigurable-resources-arch-dependent-changes-arm-fix.patch +gregkh-pci-pci-64-bit-resources-core-changes-mips-fix.patch

Unbreak it.

-bogus-disk-geometry-on-large-disks-warning-fix.patch

Folded into bogus-disk-geometry-on-large-disks.patch

-areca-raid-linux-scsi-driver-update6-for-2617-rc1-mm3.patch -areca-raid-linux-scsi-driver-update6-for-2617-rc1-mm3-externs-go-in-headers.patch

Folded into areca-raid-linux-scsi-driver.patch

+git-scsi-target-fixup.patch

Fix reject in git-scsi-target.patch.

+gregkh-usb-usb-whiteheat-fix-firmware-spurious-errors.patch +gregkh-usb-usb-add-sierra-wireless-mc5720-id-to-airprime.c.patch +gregkh-usb-usb-negative-index-in-drivers-usb-host-isp116x-hcd.c.patch +gregkh-usb-usb-cdc_ether-recognize-olympus-r1000.patch +gregkh-usb-usbcore-port-reset-for-composite-devices.patch +gregkh-usb-usb-hub-use-usb_reset_composite_device.patch +gregkh-usb-usb-storage-use-usb_reset_composite_device.patch +gregkh-usb-usbhid-use-usb_reset_composite_device.patch +gregkh-usb-usbcore-recovery-from-set-configuration-failure.patch +gregkh-usb-usb-drivers-usb-core-devio.c-dereferences-a-userspace-pointer.patch +gregkh-usb-usb-new-devices-for-the-option-driver.patch

USB tree updates

+x86_64-mm-acpi-blacklist-xw9300.patch +x86_64-mm-apic-support-for-extended-apic-interrupt.patch +x86_64-mm-mce_amd-relocate-sysfs-files.patch +x86_64-mm-mce_amd-support-for-family-0x10-processors.patch +x86_64-mm-mce_amd-cleanup.patch +x86_64-mm-miscellaneous-mm-initc-fixes.patch

x86_64 tree updates

+fall-back-to-old-style-call-trace-if-no-unwinding.patch +allow-unwinder-to-build-without-module-support.patch

Fix it.

+mm-slabc-fix-early-init-assumption.patch

slab fix

+tiacx-ia64-fix.patch

wireless driver fix

+selinux-add-hooks-for-key-subsystem.patch

Wire the key management subsystem into selinux.

+powerpc-vdso-updates.patch

powerpc update

+remove-empty-node-at-boot-time.patch

NUMA fixlet.

+jbd-fix-bug-in-journal_commit_transaction-fix.patch

Fix jbd-fix-bug-in-journal_commit_transaction.patch

+ufs-easy-debug.patch
+ufs-little-directory-lookup-optimization.patch
+ufs-i_blocks-wrong-count.patch
+ufs-unlock_super-without-lock.patch
+ufs-zero-metadata.patch
+ufs-printk-warning-fixes.patch

More UFS fixes

-inotify-kernel-api.patch -inotify-kernel-api-fix.patch +inotify-split-kernel-api-from-userspace-support.patch +inotify-add-names-inode-to-event-handler.patch +inotify-add-interfaces-to-kernel-api.patch +inotify-allow-watch-removal-from-event-handler.patch +inotify-update-kernel-documentation.patch

Updated inotify patch series

+lock-validator-introduce-warn_on_oncecond-speedup.patch

Fix lock-validator-introduce-warn_on_oncecond.patch

+add-max6902-rtc-support-update.patch

Fix add-max6902-rtc-support.patch

+nbd-endian-annotations.patch
+epoll-use-unlocked-wqueue-operations.patch

misc updates

+per-task-delay-accounting-taskstats-interface-fix-2.patch

Fix per-task-delay-accounting-taskstats-interface-fix-1.patch

+sched-fix-smt-nice-lock-contention-and-optimization.patch +sched-fix-smt-nice-lock-contention-and-optimization-tidy.patch

CPu scheduler scability improvements.

+namespaces-utsname-sysctl-hack-cleanup-2-fix.patch

Fix namespaces-utsname-sysctl-hack-cleanup-2.patch

+reiser4-hardirq-include-fix.patch

Fix reiser4.patch

+skeletonfb-remove-duplicate-module-init-exit-license-lines.patch +neofb-fix-unblank-logic-interfering-with-lid-toggled-backlight.patch

fbdev updates

+statistics-infrastructure-prerequisite-timestamp-fix.patch

Fix statistics-infrastructure-prerequisite-timestamp.patch

+genirq-msi-fixes-2.patch

Fix genirq-core.patch

+genirq-add-irq-chip-support-fix.patch

Fix genirq-add-irq-chip-support.patch

+genirq-add-chip-eoi-fastack-fasteoi-fix.patch

Fix genirq-add-chip-eoi-fastack-fasteoi.patch

+lock-validator-floppyc-irq-release-fix-fix.patch

Fix lock-validator-floppyc-irq-release-fix.patch

+lock-validator-locking-api-self-tests-self-test-fix.patch

Fix lock-validator-locking-api-self-tests.patch

+lock-validator-beautify-x86_64-stacktraces-fix-2.patch +lock-validator-beautify-x86_64-stacktraces-fix-3.patch +lock-validator-beautify-x86_64-stacktraces-fix-4.patch

Fix lock-validator-beautify-x86_64-stacktraces.patch some more.

+lock-validator-x86_64-irqflags-trace-entrys-fix.patch

Fix lock-validator-irqtrace-cleanup-include-asm-x86_64-irqflagsh.patch

+lock-validator-core-early_boot_irqs_-build-fix.patch +lock-validator-core-fix-compiler-warning.patch

Fix lock-validator-core.patch

+lock-validator-special-locking-serio.patch +lockdep-add-i_mutex-ordering-annotations-to-the-sunrpc.patch +lockdep-add-parent-child-annotations-to-usbfs.patch

lockdep workarounds

+i386-remove-multi-entry-backtraces.patch

More work on x86 backtraces.

All 1492 patches:

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm2/patch-list

#1 lore raw

When I build ACPI processor support as a module, I get this:

MODPOST
WARNING: drivers/acpi/processor.o - Section mismatch: reference to .init.data: from .text between 'acpi_processor_power_init' (at offset 0xfb0) and 'acpi_safe_halt'

(This is also true of -mm2, but I didn't get a chance to report it before -mm3 was released. Before then, I built it into the kernel and not as a module.)

and I still get this:
WARNING: "scsi_tgt_queue_command" [drivers/scsi/libsrp.ko] undefined!

Signature · 1 line
-Barry K. Nathan <[email protected]>
#2 lore raw
Quoted 6 lines · On Sun, 4 Jun 2006 02:38:03 -0700 "Barry K. Nathan" <[email protected]> wrote:

When I build ACPI processor support as a module, I get this:

MODPOST
WARNING: drivers/acpi/processor.o - Section mismatch: reference to .init.data: from .text between 'acpi_processor_power_init' (at offset 0xfb0) and 'acpi_safe_halt'

yup. The code in there is actually correct (assuming acpi_processor_power_init()'s first invokation is at initcall-time).

Maybe we'll do something to kill the warning, once we're down to the last few thousand of them ;)

Quoted 6 lines

(This is also true of -mm2, but I didn't get a chance to report it before -mm3 was released. Before then, I built it into the kernel and not as a module.)

and I still get this:
WARNING: "scsi_tgt_queue_command" [drivers/scsi/libsrp.ko] undefined!

git-scsi-target Kconfig snafu. I passed it over to James the other day. He might have fixed it - I get my git-scsi-misc via git-infiniband (don't ask) and it's a bit laggy.

#3 lore raw

Hi Andrew,

Quoted 10 lines · Andrew Morton
Quoted 6 lines · On Sun, 4 Jun 2006 02:38:03 -0700 "Barry K. Nathan" <[email protected]> wrote:

When I build ACPI processor support as a module, I get this:

MODPOST
WARNING: drivers/acpi/processor.o - Section mismatch: reference to .init.data: from .text between 'acpi_processor_power_init' (at offset 0xfb0) and 'acpi_safe_halt'

Quoted 5 lines

yup. The code in there is actually correct (assuming acpi_processor_power_init()'s first invokation is at initcall-time).

Maybe we'll do something to kill the warning, once we're down to the last few thousand of them ;)

I have got something similar
WARNING: drivers/usb/storage/usb-storage.o - Section mismatch:
reference to .exit.text: from .smp_locks after '' (at offset 0x3c)
WARNING: net/ipv4/netfilter/ip_conntrack.o - Section mismatch:
reference to .init.text: from .smp_locks after '' (at offset 0x8)
WARNING: net/ipv6/ipv6.o - Section mismatch: reference to .init.text:
from .smp_locks after '' (at offset 0x14c)
WARNING: net/ipv6/ipv6.o - Section mismatch: reference to .init.text:
from .smp_locks after '' (at offset 0x17c)

BTW. I still get this bug
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm2/bug_1.jpg
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm2/mm-config

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#4 lore raw
Quoted 3 lines · Michal Piotrowski

BTW. I still get this bug
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm2/bug_1.jpg
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm2/mm-config

could you please apply the following patches ontop of -mm3:

http://redhat.com/~mingo/lockdep-patches/lockdep-combo-2.6.17-rc5-mm3.patch
http://redhat.com/~mingo/lockdep-patches/lockdep-tracer-2.6.17-rc5-mm3.patch

accept all the default 'make oldconfig' options and reboot into the patched kernel. If everything goes well then the system should still boot up fine and you should still get the lockdep warning - but this time there should be a long trace in /proc/latency_trace. Please upload that trace - it gives us the kernel's function trace, leading up to the warning.

Ingo

#5 lore raw
Quoted 4 lines · Ingo Molnar

could you please apply the following patches ontop of -mm3:

http://redhat.com/~mingo/lockdep-patches/lockdep-combo-2.6.17-rc5-mm3.patch
http://redhat.com/~mingo/lockdep-patches/lockdep-tracer-2.6.17-rc5-mm3.patch

Just for grins, I tried building this, and got this error:

CC kernel/irq/handle.o
kernel/irq/handle.c:246:35: error: macro "early_init_irq_lock_type" passed 1 arguments, but takes just 0 kernel/irq/handle.c:247: error: expected '=', ',', ';', 'asm' or '__attribute__' before '{' token make[2]: *** [kernel/irq/handle.o] Error 1

It won't build if you don't have CONFIG_TRACE_IRQFLAGS defined - and that is defined like this:

config TRACE_IRQFLAGS
bool
default y
depends on TRACE_IRQFLAGS_SUPPORT depends on PROVE_SPIN_LOCKING || PROVE_RW_LOCKING

but my config has:
% grep PROVE .config
# CONFIG_PROVE_SPIN_LOCKING is not set
# CONFIG_PROVE_RW_LOCKING is not set
# CONFIG_PROVE_MUTEX_LOCKING is not set
# CONFIG_PROVE_RWSEM_LOCKING is not set

So using the defaults for the PROVE_* won't compile clean. Yes, probably a stupid setting for anybody applying the patches, but.. ;)

(I'm off to go build kernels without the patch, and with the PROVE_* set)..

#6 lore raw

PATCH 1/1 lock validator: early_init_irq_lock_type() build fix

Quoted 1 line · * [email protected] <[email protected]> wrote:

Just for grins, I tried building this, and got this error:

the patch below should fix this.

Ingo

----
Subject: lock validator: early_init_irq_lock_type() build fix
From: Ingo Molnar <[email protected]>

fix build bug reported by Valdis Kletnieks: if the lock validator is disabled in the .config then kernel/irq/handle.c would not build.

---
 kernel/irq/handle.c |    4 ++++
 1 file changed, 4 insertions(+)
kernel/irq/handle.c +4 -0
Index: linux/kernel/irq/handle.c
===================================================================
@@ -238,6 +238,8 @@ out:
 	return 1;
 }
 
+#ifdef CONFIG_TRACE_IRQFLAGS
+
 /*
  * lockdep: we want to handle all irq_desc locks as a single lock-type:
  */
@@ -250,3 +252,5 @@ void early_init_irq_lock_type(void)
 	for (i = 0; i < NR_IRQS; i++)
 		spin_lock_init_key(&irq_desc[i].lock, &irq_desc_lock_type);
 }
+
+#endif

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

#8 lore raw
Quoted 4 lines · Michal Piotrowski

Unfortunately I can't compile this
Here is output from my build log

/usr/src/linux-mm/kernel/sched.c:3040: error: 'p' redeclared as

i've uploaded a fixed version - does that work for you?

Ingo

#9 lore raw
Quoted 9 lines · Ingo Molnar
Quoted 4 lines · Michal Piotrowski

Unfortunately I can't compile this
Here is output from my build log

/usr/src/linux-mm/kernel/sched.c:3040: error: 'p' redeclared as

i've uploaded a fixed version - does that work for you?

Yes, thanks.

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm3/mm-dmesg
Here is latency trace
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm3/mm-latency.bz2
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm3/mm-config

Here is something new

Jun  4 23:59:44 ltg01-fedora kernel: hdd: set_drive_speed_status:
status=0x51 { DriveReady SeekComplete Error }
Jun  4 23:59:44 ltg01-fedora kernel: hdd: set_drive_speed_status:
error=0xb4 { AbortedCommand LastFailedSense=0x0b }
Jun  4 23:59:44 ltg01-fedora kernel: (          hdparm-1821 |#0): new
164424143 us user-latency.
Jun  4 23:59:44 ltg01-fedora kernel: stopped custom tracer.
Jun  4 23:59:44 ltg01-fedora kernel:
Jun  4 23:59:44 ltg01-fedora kernel: ============================
Jun  4 23:59:44 ltg01-fedora kernel: [ BUG: illegal lock usage! ]
Jun  4 23:59:44 ltg01-fedora kernel: ----------------------------
Jun  4 23:59:44 ltg01-fedora kernel: illegal {in-hardirq-W} ->
{hardirq-on-W} usage.
Jun  4 23:59:44 ltg01-fedora kernel: hdparm/1821 [HC0[0]:SC0[0]:HE1:SE1] takes:
Jun  4 23:59:44 ltg01-fedora kernel:  (ide_lock){++..}, at:
[<c0268388>] ide_dump_opcode+0x13/0x9b
Jun  4 23:59:44 ltg01-fedora kernel: {in-hardirq-W} state was registered at:
Jun  4 23:59:44 ltg01-fedora kernel:   [<c013b536>] lockdep_acquire+0x67/0x7f
Jun  4 23:59:44 ltg01-fedora kernel:   [<c0305755>] _spin_lock_irqsave+0x2d/0x3c
Jun  4 23:59:44 ltg01-fedora kernel:   [<c0265fff>] ide_intr+0x18/0x1ab
Jun  4 23:59:44 ltg01-fedora kernel:   [<c015062c>] handle_IRQ_event+0x1d/0x52
Jun  4 23:59:44 ltg01-fedora kernel:   [<c015169c>] handle_edge_irq+0x113/0x15a
Jun  4 23:59:44 ltg01-fedora kernel:   [<c0105857>] do_IRQ+0xa2/0xc7
Jun  4 23:59:44 ltg01-fedora kernel: irq event stamp: 2011
Jun  4 23:59:44 ltg01-fedora kernel: hardirqs last  enabled at (2011):
[<c0305b29>] _spin_unlock_irq+0x24/0x58
Jun  4 23:59:44 ltg01-fedora kernel: hardirqs last disabled at (2010):
[<c03056c9>] _spin_lock_irq+0x11/0x38
Jun  4 23:59:44 ltg01-fedora kernel: softirqs last  enabled at (2008):
[<c012630c>] __do_softirq+0xf0/0xf8
Jun  4 23:59:44 ltg01-fedora kernel: softirqs last disabled at (2001):
[<c0105741>] do_softirq+0x5e/0xd2
Jun  4 23:59:44 ltg01-fedora kernel:
Jun  4 23:59:44 ltg01-fedora kernel: other info that might help us debug this:
Jun  4 23:59:44 ltg01-fedora kernel: no locks held by hdparm/1821.
Jun  4 23:59:44 ltg01-fedora kernel:
Jun  4 23:59:44 ltg01-fedora kernel: stack backtrace:
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0104513>] show_trace+0x1b/0x20
Jun  4 23:59:44 ltg01-fedora kernel:  [<c01045f1>] dump_stack+0x1f/0x24
Jun  4 23:59:44 ltg01-fedora kernel:  [<c013976c>] print_usage_bug+0x1a5/0x1b1
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0139e90>] mark_lock+0x2ca/0x4f7
Jun  4 23:59:44 ltg01-fedora kernel:  [<c013aa96>] __lockdep_acquire+0x47e/0xaa4
Jun  4 23:59:44 ltg01-fedora kernel:  [<c013b536>] lockdep_acquire+0x67/0x7f
Jun  4 23:59:44 ltg01-fedora kernel:  [<c030552d>] _spin_lock+0x24/0x32
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0268388>] ide_dump_opcode+0x13/0x9b
Jun  4 23:59:44 ltg01-fedora kernel:  [<c02688b6>] ide_dump_status+0x4a6/0x4cc
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0267ae6>]
ide_config_drive_speed+0x32a/0x33a
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0262dc5>] piix_tune_chipset+0x2ed/0x2f8
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0262e31>]
piix_config_drive_xfer_rate+0x61/0xb5
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0263a82>] set_using_dma+0x2f/0x60
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0263bee>] ide_write_setting+0x4a/0xc3
Jun  4 23:59:44 ltg01-fedora kernel:  [<c02647ca>] generic_ide_ioctl+0x8a/0x47f
Jun  4 23:59:44 ltg01-fedora kernel:  [<f886003a>]
idecd_ioctl+0xfd/0x133 [ide_cd]
Jun  4 23:59:44 ltg01-fedora kernel:  [<c01f1fff>] blkdev_driver_ioctl+0x4b/0x5f
Jun  4 23:59:44 ltg01-fedora kernel:  [<c01f2783>] blkdev_ioctl+0x770/0x7bd
Jun  4 23:59:44 ltg01-fedora kernel:  [<c017dc0d>] block_ioctl+0x1f/0x21
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0189353>] do_ioctl+0x27/0x6e
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0189604>] vfs_ioctl+0x26a/0x280
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0189667>] sys_ioctl+0x4d/0x7e
Jun  4 23:59:44 ltg01-fedora kernel:  [<c0305ed2>] sysenter_past_esp+0x63/0xa1
Jun  4 23:59:44 ltg01-fedora kernel: ---------------------------
Jun  4 23:59:44 ltg01-fedora kernel: | preempt count: 00000001 ]
Jun  4 23:59:44 ltg01-fedora kernel: | 1-level deep critical section nesting:
Jun  4 23:59:44 ltg01-fedora kernel: ----------------------------------------
Jun  4 23:59:44 ltg01-fedora kernel: .. [<c030551b>] .... _spin_lock+0x12/0x32
Jun  4 23:59:44 ltg01-fedora kernel: .....[<c0268388>] ..   ( <=
ide_dump_opcode+0x13/0x9b)
Jun  4 23:59:44 ltg01-fedora kernel:
Jun  4 23:59:44 ltg01-fedora kernel: ide: failed opcode was: unknown

I get this when I do "hdparm -c 1 -d 1 /dev/hd{c,d}"

Quoted 3 lines

Ingo

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#10 lore raw

PATCH 1/1 fix IDE deadlock in error reporting code

Quoted 28 lines · Michal Piotrowski
Quoted 1 line

i've uploaded a fixed version - does that work for you?

Yes, thanks.

Here is dmesg
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm3/mm-dmesg
Here is latency trace
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm3/mm-latency.bz2
Here is config
http://www.stardust.webpages.pl/files/mm/2.6.17-rc5-mm3/mm-config

Here is something new

Jun  4 23:59:44 ltg01-fedora kernel: hdd: set_drive_speed_status:
status=0x51 { DriveReady SeekComplete Error }
Jun  4 23:59:44 ltg01-fedora kernel: hdd: set_drive_speed_status:
error=0xb4 { AbortedCommand LastFailedSense=0x0b }
Jun  4 23:59:44 ltg01-fedora kernel: (          hdparm-1821 |#0): new
164424143 us user-latency.
Jun  4 23:59:44 ltg01-fedora kernel: stopped custom tracer.
Jun  4 23:59:44 ltg01-fedora kernel:
Jun  4 23:59:44 ltg01-fedora kernel: ============================
Jun  4 23:59:44 ltg01-fedora kernel: [ BUG: illegal lock usage! ]
Jun  4 23:59:44 ltg01-fedora kernel: ----------------------------
Jun  4 23:59:44 ltg01-fedora kernel: illegal {in-hardirq-W} ->
{hardirq-on-W} usage.
Jun  4 23:59:44 ltg01-fedora kernel: hdparm/1821 [HC0[0]:SC0[0]:HE1:SE1] 
takes:
Jun  4 23:59:44 ltg01-fedora kernel:  (ide_lock){++..}, at:
Quoted 1 line

[<c0268388>] ide_dump_opcode+0x13/0x9b

ah. That's a real deadlock scenario. Does the patch below fix it? If yes then i think this is a candidate for 2.6.17 merging too.

Ingo

--------------------------
Subject: [patch, -rc5-mm3] fix IDE deadlock in error reporting code
From: Ingo Molnar <[email protected]>

Michal Piotrowski reported the following validator assert:

hdd: set_drive_speed_status: status=0x51 { DriveReady SeekComplete Error }
hdd: set_drive_speed_status: error=0xb4 { AbortedCommand LastFailedSense=0x0b }

============================
[ BUG: illegal lock usage! ]
----------------------------
illegal {in-hardirq-W} -> {hardirq-on-W} usage. hdparm/1821 [HC0[0]:SC0[0]:HE1:SE1] takes:
(ide_lock){++..}, at: [<c0268388>] ide_dump_opcode+0x13/0x9b

[...]

 stack backtrace:
  [<c0104513>] show_trace+0x1b/0x20
  [<c01045f1>] dump_stack+0x1f/0x24
  [<c013976c>] print_usage_bug+0x1a5/0x1b1
  [<c0139e90>] mark_lock+0x2ca/0x4f7
  [<c013aa96>] __lockdep_acquire+0x47e/0xaa4
  [<c013b536>] lockdep_acquire+0x67/0x7f
  [<c030552d>] _spin_lock+0x24/0x32
  [<c0268388>] ide_dump_opcode+0x13/0x9b
  [<c02688b6>] ide_dump_status+0x4a6/0x4cc
  [<c0267ae6>] ide_config_drive_speed+0x32a/0x33a
  [<c0262dc5>] piix_tune_chipset+0x2ed/0x2f8
  [<c0262e31>] piix_config_drive_xfer_rate+0x61/0xb5
  [<c0263a82>] set_using_dma+0x2f/0x60
  [<c0263bee>] ide_write_setting+0x4a/0xc3
  [<c02647ca>] generic_ide_ioctl+0x8a/0x47f
  [<f886003a>] idecd_ioctl+0xfd/0x133 [ide_cd]
  [<c01f1fff>] blkdev_driver_ioctl+0x4b/0x5f
  [<c01f2783>] blkdev_ioctl+0x770/0x7bd
  [<c017dc0d>] block_ioctl+0x1f/0x21
  [<c0189353>] do_ioctl+0x27/0x6e
  [<c0189604>] vfs_ioctl+0x26a/0x280
  [<c0189667>] sys_ioctl+0x4d/0x7e
  [<c0305ed2>] sysenter_past_esp+0x63/0xa1

in ide_dump_opcode() takes the ide_lock in an irq-unsafe manner, i.e. this function expects to be called with irqs disabled. But ide_dump_ata[pi]_status() doesnt do that - it enables interrupts specifically. That is a no-no - what guarantees that another IDE port couldnt generate an IDE interrupt while we are dumping this error? The fix is to turn the irq-enabling in these functions into irq-disabling.

---
 drivers/ide/ide-lib.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/ide/ide-lib.c +2 -2
Index: linux/drivers/ide/ide-lib.c
===================================================================
@@ -506,7 +506,7 @@ static u8 ide_dump_ata_status(ide_drive_
 	unsigned long flags;
 	u8 err = 0;
 
-	local_irq_set(flags);
+	local_irq_save(flags);
 	printk("%s: %s: status=0x%02x { ", drive->name, msg, stat);
 	if (stat & BUSY_STAT)
 		printk("Busy ");
@@ -588,7 +588,7 @@ static u8 ide_dump_atapi_status(ide_driv
 
 	status.all = stat;
 	error.all = 0;
-	local_irq_set(flags);
+	local_irq_save(flags);
 	printk("%s: %s: status=0x%02x { ", drive->name, msg, stat);
 	if (status.b.bsy)
 		printk("Busy ");
#11 lore raw

Re: [patch, -rc5-mm3] fix IDE deadlock in error reporting code

Quoted 2 lines · Ingo Molnar

ah. That's a real deadlock scenario. Does the patch below fix it? If yes then i think this is a candidate for 2.6.17 merging too.

actually, the replacement patch below is better i think - it moves the ide_lock taking to outside the printing section. That should still be OK as we dont call other functions from within the section, and it should also result in slightly more robust printing, as the whole printing code will be atomic under ide_lock.

Ingo

---------
Subject: [patch, -rc5-mm3] fix IDE deadlock in error reporting code
From: Ingo Molnar <[email protected]>

Michal Piotrowski reported the following validator assert:

hdd: set_drive_speed_status: status=0x51 { DriveReady SeekComplete Error }
hdd: set_drive_speed_status: error=0xb4 { AbortedCommand LastFailedSense=0x0b }

============================
[ BUG: illegal lock usage! ]
----------------------------
illegal {in-hardirq-W} -> {hardirq-on-W} usage. hdparm/1821 [HC0[0]:SC0[0]:HE1:SE1] takes:
(ide_lock){++..}, at: [<c0268388>] ide_dump_opcode+0x13/0x9b

[...]

 stack backtrace:
  [<c0104513>] show_trace+0x1b/0x20
  [<c01045f1>] dump_stack+0x1f/0x24
  [<c013976c>] print_usage_bug+0x1a5/0x1b1
  [<c0139e90>] mark_lock+0x2ca/0x4f7
  [<c013aa96>] __lockdep_acquire+0x47e/0xaa4
  [<c013b536>] lockdep_acquire+0x67/0x7f
  [<c030552d>] _spin_lock+0x24/0x32
  [<c0268388>] ide_dump_opcode+0x13/0x9b
  [<c02688b6>] ide_dump_status+0x4a6/0x4cc
  [<c0267ae6>] ide_config_drive_speed+0x32a/0x33a
  [<c0262dc5>] piix_tune_chipset+0x2ed/0x2f8
  [<c0262e31>] piix_config_drive_xfer_rate+0x61/0xb5
  [<c0263a82>] set_using_dma+0x2f/0x60
  [<c0263bee>] ide_write_setting+0x4a/0xc3
  [<c02647ca>] generic_ide_ioctl+0x8a/0x47f
  [<f886003a>] idecd_ioctl+0xfd/0x133 [ide_cd]
  [<c01f1fff>] blkdev_driver_ioctl+0x4b/0x5f
  [<c01f2783>] blkdev_ioctl+0x770/0x7bd
  [<c017dc0d>] block_ioctl+0x1f/0x21
  [<c0189353>] do_ioctl+0x27/0x6e
  [<c0189604>] vfs_ioctl+0x26a/0x280
  [<c0189667>] sys_ioctl+0x4d/0x7e
  [<c0305ed2>] sysenter_past_esp+0x63/0xa1

in ide_dump_opcode() takes the ide_lock in an irq-unsafe manner, i.e. this function expects to be called with irqs disabled. But ide_dump_ata[pi]_status() doesnt do that - it enables interrupts specifically. That is a no-no - what guarantees that another IDE port couldnt generate an IDE interrupt while we are dumping this error? The fix is to turn the irq-enabling in these functions into irq-disabling.

---
 drivers/ide/ide-lib.c |   14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)
drivers/ide/ide-lib.c +8 -6
Index: linux/drivers/ide/ide-lib.c
===================================================================
@@ -471,11 +471,9 @@ static void ide_dump_opcode(ide_drive_t 
 	u8 opcode = 0;
 	int found = 0;
 
-	spin_lock(&ide_lock);
 	rq = NULL;
 	if (HWGROUP(drive))
 		rq = HWGROUP(drive)->rq;
-	spin_unlock(&ide_lock);
 	if (!rq)
 		return;
 	if (rq->flags & (REQ_DRIVE_CMD | REQ_DRIVE_TASK)) {
@@ -506,7 +504,7 @@ static u8 ide_dump_ata_status(ide_drive_
 	unsigned long flags;
 	u8 err = 0;
 
-	local_irq_set(flags);
+	spin_lock_irqsave(&ide_lock, flags);
 	printk("%s: %s: status=0x%02x { ", drive->name, msg, stat);
 	if (stat & BUSY_STAT)
 		printk("Busy ");
@@ -566,7 +564,8 @@ static u8 ide_dump_ata_status(ide_drive_
 		printk("\n");
 	}
 	ide_dump_opcode(drive);
-	local_irq_restore(flags);
+	spin_unlock_irqrestore(&ide_lock, flags);
+
 	return err;
 }
 
@@ -588,7 +587,9 @@ static u8 ide_dump_atapi_status(ide_driv
 
 	status.all = stat;
 	error.all = 0;
-	local_irq_set(flags);
+
+	spin_lock_irqsave(&ide_lock, flags);
+
 	printk("%s: %s: status=0x%02x { ", drive->name, msg, stat);
 	if (status.b.bsy)
 		printk("Busy ");
@@ -614,7 +615,8 @@ static u8 ide_dump_atapi_status(ide_driv
 		printk("}\n");
 	}
 	ide_dump_opcode(drive);
-	local_irq_restore(flags);
+	spin_unlock_irqrestore(&ide_lock, flags);
+
 	return error.all;
 }
 
#12 lore raw

Re: [patch, -rc5-mm3] fix IDE deadlock in error reporting code

Hi,

Quoted 11 lines · Ingo Molnar
Quoted 2 lines · Ingo Molnar

ah. That's a real deadlock scenario. Does the patch below fix it? If yes then i think this is a candidate for 2.6.17 merging too.

actually, the replacement patch below is better i think - it moves the ide_lock taking to outside the printing section. That should still be OK as we dont call other functions from within the section, and it should also result in slightly more robust printing, as the whole printing code will be atomic under ide_lock.

Probably fixed, thanks.

Quoted 3 lines

Ingo

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#13 lore raw

Re: [patch, -rc5-mm3] fix IDE deadlock in error reporting code

Quoted 9 lines · Michal Piotrowski

Hi,

Quoted 6 lines · Ingo Molnar
Quoted 2 lines · Ingo Molnar

ah. That's a real deadlock scenario. Does the patch below fix it? If yes then i think this is a candidate for 2.6.17 merging too.

Quoted 7 lines
Quoted 5 lines

actually, the replacement patch below is better i think - it moves the ide_lock taking to outside the printing section. That should still be OK as we dont call other functions from within the section, and it should also result in slightly more robust printing, as the whole printing code will be atomic under ide_lock.

Probably fixed, thanks.

could you send a confirmation if/when you have tried this? I dont get the same message (i have another IDE chipset).

Ingo

#14 lore raw

Re: [patch, -rc5-mm3] fix IDE deadlock in error reporting code

Quoted 4 lines · Ingo Molnar
Quoted 1 line · Michal Piotrowski

Hi,

[snip]

Quoted 4 lines
Quoted 1 line

Probably fixed, thanks.

could you send a confirmation if/when you have tried this? I dont get the same message (i have another IDE chipset).

I have tried to reproduce that bug, it seems that it's very hard to reproduce. After 20 reboots with your patch I haven't seen that.

Quoted 3 lines

Ingo

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#15 lore raw
In reply to Andrew Morton
Quoted 2 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

Small compilation fix needed for x86_64 without SMP:

 arch/x86_64/kernel/mce_amd.c |    4 ++++
 1 files changed, 4 insertions(+)
arch/x86_64/kernel/mce_amd.c +4 -0
Index: linux-2.6.17-rc5-mm3/arch/x86_64/kernel/mce_amd.c
===================================================================
@@ -494,7 +494,11 @@ static __cpuinit int threshold_create_ba
 
 	kobject_set_name(&b->kobj, "threshold_bank%i", bank);
 	b->kobj.parent = &per_cpu(device_mce, cpu).kobj;
+#ifdef CONFIG_SMP
 	b->cpus = cpu_core_map[cpu];
+#else
+	b->cpus = CPU_MASK_CPU0;
+#endif
 
 	err = kobject_register(&b->kobj);
 	if (err)
#16 lore raw

Re: [trivial patch, rc5-mm3] fix typo in acpi pm info message

In reply to Andrew Morton

Hi,

There's a typo in an ACPI info message introduced by the following patch:
http://kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/broken-out/acpi-identify-which-device-is-not-power-manageable.patch

Here's the fix:

acpi/bus.c renamed from acpi/bus.c_ +1 -1
@@ -188,7 +188,7 @@ int acpi_bus_set_power(acpi_handle handl
 	/* Make sure this is a valid target state */
 
 	if (!device->flags.power_manageable) {
-		ACPI_INFO((AE_INFO, "Device `%s]is not power manageable",
+		ACPI_INFO((AE_INFO, "Device [%s] is not power manageable",
 				device->kobj.name));
 		return -ENODEV;
 	}
#17 lore raw
In reply to Andrew Morton
Quoted 8 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

Got this on boot. Looks like another locking bug in firewire:

ACPI: PCI Interrupt 0000:03:03.0[A] -> GSI 20 (level, low) -> IRQ 20
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[20] MMIO=[ec024000-ec0247ff] Max Packet=[2048] IR/IT contexts=[4/8] stopped custom tracer.

============================
[ BUG: illegal lock usage! ]
----------------------------
illegal {hardirq-on-W} -> {in-hardirq-R} usage.
idle/0 [HC1[1]:SC1[0]:HE0:SE0] takes:
(hl_irqs_lock){--+.}, at: [<f8835cb9>] highlevel_host_reset+0x11/0x5b [ieee1394]
{hardirq-on-W} state was registered at:

  [<c0133fe4>] lockdep_acquire+0x4d/0x63
  [<c02f3421>] _write_lock+0x2e/0x3b
  [<f88365ab>] hpsb_register_highlevel+0xac/0xea [ieee1394]
  [<f8836d6a>] init_csr+0x28/0x3f [ieee1394]
  [<f880617d>] 0xf880617d
  [<c01398df>] sys_init_module+0x12a/0x1b7b
  [<c02f3b2d>] sysenter_past_esp+0x56/0x8d
irq event stamp: 258193
hardirqs last  enabled at (258192): [<c011fab5>] __do_softirq+0x67/0xf7
hardirqs last disabled at (258193): [<c0102eb7>] common_interrupt+0x1b/0x2c
softirqs last  enabled at (258186): [<c011fb34>] __do_softirq+0xe6/0xf7
softirqs last disabled at (258191): [<c0104cec>] do_softirq+0x5a/0xc9

other info that might help us debug this:
no locks held by idle/0.

stack backtrace:
 [<c01034ba>] show_trace+0x12/0x14
 [<c0103b8d>] dump_stack+0x19/0x1b
 [<c0132025>] print_usage_bug+0x20b/0x215
 [<c01329cc>] mark_lock+0x4fa/0x5b4
 [<c0133399>] __lockdep_acquire+0x310/0xbc0
 [<c0133fe4>] lockdep_acquire+0x4d/0x63
 [<c02f3153>] _read_lock+0x2e/0x3b
 [<f8835cb9>] highlevel_host_reset+0x11/0x5b [ieee1394]
 [<f8833867>] hpsb_selfid_complete+0x286/0x307 [ieee1394]
 [<f884ec30>] ohci_irq_handler+0x6c9/0x995 [ohci1394]
 [<c013d3a2>] handle_IRQ_event+0x2e/0x63
 [<c013e4c3>] handle_fasteoi_irq+0x6b/0xac
 [<c0104dc7>] do_IRQ+0x6c/0xa5
 =======================
 [<c0102ec1>] common_interrupt+0x25/0x2c
 [<c0104cec>] do_softirq+0x5a/0xc9
 =======================
 [<c011fb90>] irq_exit+0x4b/0x4d
 [<c0104dce>] do_IRQ+0x73/0xa5
 [<c0102ec1>] common_interrupt+0x25/0x2c
 [<c010164e>] cpu_idle+0x63/0x80
 [<c0100599>] rest_init+0x33/0x3a
 [<c03d97af>] start_kernel+0x339/0x3aa
 [<c0100210>] 0xc0100210

ieee1394: Host added: ID:BUS[0-00:1023] GUID[00e018000063814f]

--
J.A. Magallon <jamagallon()ono!com> \ Software is like sex:
\ It's better when it's free
Mandriva Linux release 2007.0 (Cooker) for i586
Linux 2.6.16-jam18 (gcc 4.1.1 20060518 (prerelease)) #2 SMP PREEMPT Mon

#18 lore raw
Quoted 13 lines · On Mon, 5 Jun 2006 01:15:31 +0200 "J.A. Magallón" <[email protected]> wrote:
Quoted 8 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

Got this on boot. Looks like another locking bug in firewire:

Quoted 8 lines

ACPI: PCI Interrupt 0000:03:03.0[A] -> GSI 20 (level, low) -> IRQ 20
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[20] MMIO=[ec024000-ec0247ff] Max Packet=[2048] IR/IT contexts=[4/8] stopped custom tracer.

============================
[ BUG: illegal lock usage! ]
----------------------------
illegal {hardirq-on-W} -> {in-hardirq-R} usage.

So we have an rwlock which was acquired for writing under local_irq_enable() but we later acquired it for reading inside an interrupt handler.

Quoted 10 lines

idle/0 [HC1[1]:SC1[0]:HE0:SE0] takes:
(hl_irqs_lock){--+.}, at: [<f8835cb9>] highlevel_host_reset+0x11/0x5b [ieee1394]
{hardirq-on-W} state was registered at:

  [<c0133fe4>] lockdep_acquire+0x4d/0x63
  [<c02f3421>] _write_lock+0x2e/0x3b
  [<f88365ab>] hpsb_register_highlevel+0xac/0xea [ieee1394]
  [<f8836d6a>] init_csr+0x28/0x3f [ieee1394]
  [<f880617d>] 0xf880617d
  [<c01398df>] sys_init_module+0x12a/0x1b7b
  [<c02f3b2d>] sysenter_past_esp+0x56/0x8d

Here's the irqs-on write_lock.

Quoted 9 lines
irq event stamp: 258193
hardirqs last  enabled at (258192): [<c011fab5>] __do_softirq+0x67/0xf7
hardirqs last disabled at (258193): [<c0102eb7>] common_interrupt+0x1b/0x2c
softirqs last  enabled at (258186): [<c011fb34>] __do_softirq+0xe6/0xf7
softirqs last disabled at (258191): [<c0104cec>] do_softirq+0x5a/0xc9

other info that might help us debug this:
no locks held by idle/0.

Quoted 14 lines
stack backtrace:
 [<c01034ba>] show_trace+0x12/0x14
 [<c0103b8d>] dump_stack+0x19/0x1b
 [<c0132025>] print_usage_bug+0x20b/0x215
 [<c01329cc>] mark_lock+0x4fa/0x5b4
 [<c0133399>] __lockdep_acquire+0x310/0xbc0
 [<c0133fe4>] lockdep_acquire+0x4d/0x63
 [<c02f3153>] _read_lock+0x2e/0x3b
 [<f8835cb9>] highlevel_host_reset+0x11/0x5b [ieee1394]
 [<f8833867>] hpsb_selfid_complete+0x286/0x307 [ieee1394]
 [<f884ec30>] ohci_irq_handler+0x6c9/0x995 [ohci1394]
 [<c013d3a2>] handle_IRQ_event+0x2e/0x63
 [<c013e4c3>] handle_fasteoi_irq+0x6b/0xac
 [<c0104dc7>] do_IRQ+0x6c/0xa5

And here's the in-irq read_lock().

Simple fix would be to take hl_irqs_lock in an irq-safe manner everywhere.

#19 lore raw
In reply to J.A. Magallón
Quoted 10 lines · On Mon, 05 Jun 2006 01:15:31 +0200, "J.A. =?UTF-8?B?TWFnYWxsw7Nu?=" said:
Quoted 1 line · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

============================
[ BUG: illegal lock usage! ]
----------------------------
illegal {hardirq-on-W} -> {in-hardirq-R} usage.
idle/0 [HC1[1]:SC1[0]:HE0:SE0] takes:
(hl_irqs_lock){--+.}, at: [<f8835cb9>] highlevel_host_reset+0x11/0x5b [ieee1394]
{hardirq-on-W} state was registered at:

Quoted 7 lines
  [<c0133fe4>] lockdep_acquire+0x4d/0x63
  [<c02f3421>] _write_lock+0x2e/0x3b
  [<f88365ab>] hpsb_register_highlevel+0xac/0xea [ieee1394]
  [<f8836d6a>] init_csr+0x28/0x3f [ieee1394]
  [<f880617d>] 0xf880617d
  [<c01398df>] sys_init_module+0x12a/0x1b7b
  [<c02f3b2d>] sysenter_past_esp+0x56/0x8d

ACK. I saw this same one too, while udevd was trying to get its act together in very early rc.sysinit....

#20 lore raw
In reply to J.A. Magallón
Quoted 17 lines · J.A. Magallón
Quoted 8 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

Got this on boot. Looks like another locking bug in firewire:

ACPI: PCI Interrupt 0000:03:03.0[A] -> GSI 20 (level, low) -> IRQ 20
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[20] MMIO=[ec024000-ec0247ff] Max Packet=[2048] IR/IT contexts=[4/8] stopped custom tracer.

Quoted 6 lines

============================
[ BUG: illegal lock usage! ]
----------------------------
illegal {hardirq-on-W} -> {in-hardirq-R} usage.
idle/0 [HC1[1]:SC1[0]:HE0:SE0] takes:
(hl_irqs_lock){--+.}, at: [<f8835cb9>] highlevel_host_reset+0x11/0x5b [ieee1394]

this one was reported a few days ago and acknowledged by the firewire people as real.. it seems they haven't sent Andrew a fix yet. If they don't do that today I'll send a provisional fix

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

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

I got this with -mm2, is it supposed to be cured in -mm3 ? I still have to try with mm3:

Jun  2 14:34:39 annwn kernel: Console: colour VGA+ 80x60
Jun  2 14:34:39 annwn kernel: ------------------------
Jun  2 14:34:39 annwn kernel: | Locking API testsuite:
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:                                  | spin |wlock |rlock |mutex | wsem | rsem |
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:                      A-A deadlock:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                  A-B-B-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:              A-B-B-C-C-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:              A-B-C-A-B-C deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-B-C-C-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-C-D-B-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-C-D-B-C-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                     double unlock:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                  bad unlock order:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:               recursive read-lock:             |  ok  |             |failed|
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:      hard-irqs-on + irq-safe-A/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      soft-irqs-on + irq-safe-A/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      hard-irqs-on + irq-safe-A/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      soft-irqs-on + irq-safe-A/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:        sirq-safe-A => hirqs-on/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:        sirq-safe-A => hirqs-on/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:          hard-safe-A + irqs-on/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:          soft-safe-A + irqs-on/12:failed|failed|  ok  |

(all tests failed like this...)

Jun  2 14:34:39 annwn kernel: --------------------------------------------------------
Jun  2 14:34:39 annwn kernel: 141 out of 206 testcases failed, as expected. |
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------

Expected ? Uh ?

--
J.A. Magallon <jamagallon()ono!com> \ Software is like sex:
\ It's better when it's free
Mandriva Linux release 2007.0 (Cooker) for i586
Linux 2.6.16-jam18 (gcc 4.1.1 20060518 (prerelease)) #2 SMP PREEMPT Mon

#22 lore raw
Quoted 5 lines · J.A. Magallón
Jun  2 14:34:39 annwn kernel: --------------------------------------------------------
Jun  2 14:34:39 annwn kernel: 141 out of 206 testcases failed, as expected. |
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------

Expected ? Uh ?

grep PROVE .config

Make sure all 4 of them are set to Y; if any of them are N, then test case failures would in fact be expected.

Signature · 1 line
-Barry K. Nathan <[email protected]>
#23 lore raw
In reply to J.A. Magallón
Quoted 41 lines · J.A. Magallón
Quoted 8 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

I got this with -mm2, is it supposed to be cured in -mm3 ? I still have to try with mm3:

Jun  2 14:34:39 annwn kernel: Console: colour VGA+ 80x60
Jun  2 14:34:39 annwn kernel: ------------------------
Jun  2 14:34:39 annwn kernel: | Locking API testsuite:
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:                                  | spin |wlock |rlock |mutex | wsem | rsem |
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:                      A-A deadlock:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                  A-B-B-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:              A-B-B-C-C-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:              A-B-C-A-B-C deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-B-C-C-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-C-D-B-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-C-D-B-C-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                     double unlock:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                  bad unlock order:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:               recursive read-lock:             |  ok  |             |failed|
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:      hard-irqs-on + irq-safe-A/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      soft-irqs-on + irq-safe-A/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      hard-irqs-on + irq-safe-A/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      soft-irqs-on + irq-safe-A/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:        sirq-safe-A => hirqs-on/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:        sirq-safe-A => hirqs-on/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:          hard-safe-A + irqs-on/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:          soft-safe-A + irqs-on/12:failed|failed|  ok  |
Quoted 7 lines

(all tests failed like this...)

Jun  2 14:34:39 annwn kernel: --------------------------------------------------------
Jun  2 14:34:39 annwn kernel: 141 out of 206 testcases failed, as expected. |
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------

Expected ? Uh ?

I got something like that here before turning on all the test options, suggest ' -- ' for non-selected tests. More info, first four files:
<http://bugsplatter.mine.nu/test/linux-2.6/sempro/?M=D>

dmesg, false positives:

<http://bugsplatter.mine.nu/test/linux-2.6/sempro/dmesg-2.6.17-rc5-mm3a.gz>:

------------------------
| Locking API testsuite:
----------------------------------------------------------------------------
| spin |wlock |rlock |mutex | wsem | rsem |
--------------------------------------------------------------------------

                     A-A deadlock:failed|failed|failed|failed|failed|failed|
                 A-B-B-A deadlock:failed|failed|  ok  |failed|failed|failed|
             A-B-B-C-C-A deadlock:failed|failed|  ok  |failed|failed|failed|
             A-B-C-A-B-C deadlock:failed|failed|  ok  |failed|failed|failed|
         A-B-B-C-C-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
         A-B-C-D-B-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
         A-B-C-D-B-C-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
                    double unlock:failed|failed|failed|failed|failed|failed|
                 bad unlock order:failed|failed|failed|failed|failed|failed|
  --------------------------------------------------------------------------
              recursive read-lock:             |  ok  |             |failed|
  --------------------------------------------------------------------------
     hard-irqs-on + irq-safe-A/12:failed|failed|  ok  |
     soft-irqs-on + irq-safe-A/12:failed|failed|  ok  |
     hard-irqs-on + irq-safe-A/21:failed|failed|  ok  |
     soft-irqs-on + irq-safe-A/21:failed|failed|  ok  |

and dmesg, okay:
<http://bugsplatter.mine.nu/test/linux-2.6/sempro/dmesg-2.6.17-rc5-mm3a-2.gz>:
------------------------
| Locking API testsuite:
----------------------------------------------------------------------------
| spin |wlock |rlock |mutex | wsem | rsem |
--------------------------------------------------------------------------

                     A-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
                 A-B-B-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
             A-B-B-C-C-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
             A-B-C-A-B-C deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
         A-B-B-C-C-D-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
         A-B-C-D-B-D-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
         A-B-C-D-B-C-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
                    double unlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
                 bad unlock order:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
  --------------------------------------------------------------------------
              recursive read-lock:             |  ok  |             |  ok  |
  --------------------------------------------------------------------------
                non-nested unlock:  ok  |  ok  |  ok  |  ok  |
  ------------------------------------------------------------
     hard-irqs-on + irq-safe-A/12:  ok  |  ok  |  ok  |
     soft-irqs-on + irq-safe-A/12:  ok  |  ok  |  ok  |
     hard-irqs-on + irq-safe-A/21:  ok  |  ok  |  ok  |
     soft-irqs-on + irq-safe-A/21:  ok  |  ok  |  ok  |

Grant.

#24 lore raw
In reply to J.A. Magallón
Quoted 41 lines · J.A. Magallón
Quoted 8 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

I got this with -mm2, is it supposed to be cured in -mm3 ? I still have to try with mm3:

Jun  2 14:34:39 annwn kernel: Console: colour VGA+ 80x60
Jun  2 14:34:39 annwn kernel: ------------------------
Jun  2 14:34:39 annwn kernel: | Locking API testsuite:
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:                                  | spin |wlock |rlock |mutex | wsem | rsem |
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:                      A-A deadlock:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                  A-B-B-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:              A-B-B-C-C-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:              A-B-C-A-B-C deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-B-C-C-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-C-D-B-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:          A-B-C-D-B-C-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                     double unlock:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:                  bad unlock order:failed|failed|failed|failed|failed|failed|
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:               recursive read-lock:             |  ok  |             |failed|
Jun  2 14:34:39 annwn kernel:   --------------------------------------------------------------------------
Jun  2 14:34:39 annwn kernel:      hard-irqs-on + irq-safe-A/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      soft-irqs-on + irq-safe-A/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      hard-irqs-on + irq-safe-A/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:      soft-irqs-on + irq-safe-A/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:        sirq-safe-A => hirqs-on/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:        sirq-safe-A => hirqs-on/21:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:          hard-safe-A + irqs-on/12:failed|failed|  ok  |
Jun  2 14:34:39 annwn kernel:          soft-safe-A + irqs-on/12:failed|failed|  ok  |
Quoted 7 lines

(all tests failed like this...)

Jun  2 14:34:39 annwn kernel: --------------------------------------------------------
Jun  2 14:34:39 annwn kernel: 141 out of 206 testcases failed, as expected. |
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------

Expected ? Uh ?

I got something like that here before turning on all the test options, suggest ' -- ' for non-selected tests. More info, first four files:
<http://bugsplatter.mine.nu/test/linux-2.6/sempro/?M=D>

dmesg, false positives:

<http://bugsplatter.mine.nu/test/linux-2.6/sempro/dmesg-2.6.17-rc5-mm3a.gz>:

------------------------
| Locking API testsuite:
----------------------------------------------------------------------------
| spin |wlock |rlock |mutex | wsem | rsem |
--------------------------------------------------------------------------

                     A-A deadlock:failed|failed|failed|failed|failed|failed|
                 A-B-B-A deadlock:failed|failed|  ok  |failed|failed|failed|
             A-B-B-C-C-A deadlock:failed|failed|  ok  |failed|failed|failed|
             A-B-C-A-B-C deadlock:failed|failed|  ok  |failed|failed|failed|
         A-B-B-C-C-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
         A-B-C-D-B-D-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
         A-B-C-D-B-C-D-A deadlock:failed|failed|  ok  |failed|failed|failed|
                    double unlock:failed|failed|failed|failed|failed|failed|
                 bad unlock order:failed|failed|failed|failed|failed|failed|
  --------------------------------------------------------------------------
              recursive read-lock:             |  ok  |             |failed|
  --------------------------------------------------------------------------
     hard-irqs-on + irq-safe-A/12:failed|failed|  ok  |
     soft-irqs-on + irq-safe-A/12:failed|failed|  ok  |
     hard-irqs-on + irq-safe-A/21:failed|failed|  ok  |
     soft-irqs-on + irq-safe-A/21:failed|failed|  ok  |

and dmesg, okay:
<http://bugsplatter.mine.nu/test/linux-2.6/sempro/dmesg-2.6.17-rc5-mm3a-2.gz>:
------------------------
| Locking API testsuite:
----------------------------------------------------------------------------
| spin |wlock |rlock |mutex | wsem | rsem |
--------------------------------------------------------------------------

                     A-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
                 A-B-B-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
             A-B-B-C-C-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
             A-B-C-A-B-C deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
         A-B-B-C-C-D-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
         A-B-C-D-B-D-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
         A-B-C-D-B-C-D-A deadlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
                    double unlock:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
                 bad unlock order:  ok  |  ok  |  ok  |  ok  |  ok  |  ok  |
  --------------------------------------------------------------------------
              recursive read-lock:             |  ok  |             |  ok  |
  --------------------------------------------------------------------------
                non-nested unlock:  ok  |  ok  |  ok  |  ok  |
  ------------------------------------------------------------
     hard-irqs-on + irq-safe-A/12:  ok  |  ok  |  ok  |
     soft-irqs-on + irq-safe-A/12:  ok  |  ok  |  ok  |
     hard-irqs-on + irq-safe-A/21:  ok  |  ok  |  ok  |
     soft-irqs-on + irq-safe-A/21:  ok  |  ok  |  ok  |

Grant.

#25 lore raw
In reply to J.A. Magallón
Quoted 10 lines · J.A. Magallón

I got this with -mm2, is it supposed to be cured in -mm3 ? I still have to try with mm3:

(all tests failed like this...)

Jun  2 14:34:39 annwn kernel: --------------------------------------------------------
Jun  2 14:34:39 annwn kernel: 141 out of 206 testcases failed, as expected. |
Jun  2 14:34:39 annwn kernel: ----------------------------------------------------

Expected ? Uh ?

to have lock validation you should have these options enabled:

CONFIG_PROVE_SPIN_LOCKING=y
CONFIG_PROVE_RW_LOCKING=y
CONFIG_PROVE_MUTEX_LOCKING=y
CONFIG_PROVE_RWSEM_LOCKING=y

otherwise the tests are still run, but the deadlocks are not detected. That's why those 141 testcases are 'expected' failures.

and definitely try -mm3 plus the current combo patch:

http://redhat.com/~mingo/lockdep-patches/lockdep-combo-2.6.17-rc5-mm3.patch

Ingo

#26 lore raw
In reply to Andrew Morton

I am seeing more networking-related funniness with 2.6.17-rc5-mm3 on the same machine previously fixed by git-net-llc-fix.patch. The console log is below. I've done no investigation work in case it's a known problem.

kernel /vmlinuz-autobench ro root=/dev/VolGroup00/LogVol00 rhgb quiet console=t
tyS1,19200 autobench_args: root=30726124 ABAT:1149529388 earlyprintk=serial,tty
S1,19200
[Linux-bzImage, setup=0x1e00, size=0x1e0687]
initrd /initrd-autobench.img
[Linux-initrd @ 0x37e60000, 0x18fbd9 bytes]
Bootdata ok (command line is ro root=/dev/VolGroup00/LogVol00 rhgb quiet console=ttyS1,19200 autobench_args: root=30726124 ABAT:1149529388 earlyprintk=serial,ttyS1,19200)
Linux version 2.6.17-rc5-mm2-autokern1 ([email protected]) (/usr/local/autobench/var/tmp/build/scripts/mkcompile_h: line 61: /usr/local/autobench/sources/x86_64-cross/*/bin/x86_64-unknown-linux-gnu-gcc: No such file or directory) #1 SMP Mon Jun 5 12:36:09 CDT 2006 BIOS-provided physical RAM map:
BIOS-e820: 0000000000000000 - 000000000009d400 (usable)
BIOS-e820: 000000000009d400 - 00000000000a0000 (reserved)
BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved)
BIOS-e820: 0000000000100000 - 000000003ffcddc0 (usable)
BIOS-e820: 000000003ffcddc0 - 000000003ffd0000 (ACPI data)
BIOS-e820: 000000003ffd0000 - 0000000040000000 (reserved)
BIOS-e820: 00000000fec00000 - 0000000100000000 (reserved)
kernel direct mapping tables up to 100000000 @ 8000-8000
DMI 2.3 present.
ACPI: PM-Timer IO Port: 0x2208
ACPI: LAPIC (acpi_id[0x00] lapic_id[0x00] enabled)
Processor #0 15:1 APIC version 16
ACPI: LAPIC (acpi_id[0x01] lapic_id[0x01] enabled)
Processor #1 15:1 APIC version 16
ACPI: LAPIC (acpi_id[0x02] lapic_id[0x02] enabled)
Processor #2 15:1 APIC version 16
ACPI: LAPIC (acpi_id[0x03] lapic_id[0x03] enabled)
Processor #3 15:1 APIC version 16
ACPI: LAPIC_NMI (acpi_id[0x00] dfl dfl lint[0x1])
ACPI: LAPIC_NMI (acpi_id[0x01] dfl dfl lint[0x1])
ACPI: LAPIC_NMI (acpi_id[0x02] dfl dfl lint[0x1])
ACPI: LAPIC_NMI (acpi_id[0x03] dfl dfl lint[0x1])
ACPI: IOAPIC (id[0x0e] address[0xfec00000] gsi_base[0])
IOAPIC[0]: apic_id 14, version 17, address 0xfec00000, GSI 0-23
ACPI: IOAPIC (id[0x0d] address[0xfec10000] gsi_base[24])
IOAPIC[1]: apic_id 13, version 17, address 0xfec10000, GSI 24-27
ACPI: IOAPIC (id[0x0c] address[0xfec20000] gsi_base[48])
IOAPIC[2]: apic_id 12, version 17, address 0xfec20000, GSI 48-51
ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 11 global_irq 11 low level)
Setting APIC routing to physical flat
ACPI: HPET id: 0x10228203 base: 0xfecff000
Using ACPI (MADT) for SMP configuration information
Allocating PCI resources starting at 50000000 (gap: 40000000:bec00000)
SMP: Allowing 4 CPUs, 0 hotplug CPUs
Built 1 zonelists
Kernel command line: ro root=/dev/VolGroup00/LogVol00 rhgb quiet console=ttyS1,19200 autobench_args: root=30726124 ABAT:1149529388 earlyprintk=serial,ttyS1,19200
powernow-k8: MP systems not supported by PSB BIOS structure
powernow-k8: MP systems not supported by PSB BIOS structure
powernow-k8: MP systems not supported by PSB BIOS structure
powernow-k8: MP systems not supported by PSB BIOS structure
Red Hat nash version 5.0.32 starting
Reading all physical volumes. This may take a while...
Found volume group "VolGroup00" using metadata type lvm2
2 logical volume(s) in volume group "VolGroup00" now active
INIT: version 2.86 booting
Welcome to Fedora Core
press 'I' to enter interactive startup.
Setting clock (localtime): Mon Jun 5 12:47:49 CDT 2006 [ OK ]
Starting udev: [ OK ]
Setting hostname bl6-13.ltc.austin.ibm.com: [ OK ]
Setting up Logical Volume Management: 2 logical volume(s) in volume group "VolGroup00" now active
[ OK ]
Checking filesystems
Checking all file systems.
[/sbin/fsck.ext3 (1) -- /] fsck.ext3 -a /dev/VolGroup00/LogVol00
/dev/VolGroup00/LogVol00: clean, 285228/7929856 files, 2745851/7929856 blocks
[/sbin/fsck.ext3 (1) -- /boot] fsck.ext3 -a /dev/sda1
/boot: clean, 63/512512 files, 43614/512064 blocks
[ OK ]
Remounting root filesystem in read-write mode: [ OK ]

Mounting local filesystems:  [  OK  ]
Enabling local filesystem quotas:  [  OK  ]
Enabling swap space:  [  OK  ]
INIT: Entering runlevel: 3
Entering non-interactive startup
Starting readahead_early:  Starting background readahead: [  OK  ]
[  OK  ]
FATAL: Error inserting acpi_cpufreq (/lib/modules/2.6.17-rc5-mm2-autokern1/kernel/arch/x86_64/kernel/cpufreq/acpi-cpufreq.ko): No such device
Bringing up loopback interface:  [  OK  ]
Bringing up interface eth1:  [  OK  ]
Starting system logger: [  OK  ]
Starting kernel logger: [  OK  ]
Starting irqbalance: [  OK  ]
Starting portmap: [  OK  ]
Starting NFS statd: [  OK  ]
Starting RPC idmapd: FATAL: Module sunrpc not found.
FATAL: Error running install command for sunrpc
Starting system message bus: [  OK  ]

Starting Bluetooth services:[ OK ]
[ OK ]
Mounting other filesystems: [ OK ]
Starting hidd: [ OK ]
Starting automount: [ OK ]
Starting smartd: [ OK ]
Starting acpi daemon: [ OK ]
Starting hpiod: [ OK ]
Starting hpssd: [ OK ]
Starting cups: [ OK ]
Starting sshd: [ OK ]
Starting sendmail: [ OK ]
Starting sm-client: [ OK ]
Starting console mouse services: [ OK ]
Starting crond: [ OK ]
Starting xfs: [ OK ]
Starting anacron: [ OK ]
Starting atd: [ OK ]
Starting Avahi daemon: [ OK ]
Starting cups-config-daemon: [ OK ]
Starting HAL daemon: [ OK ]
Fedora Core release 5 (Bordeaux)
Kernel 2.6.17-rc5-mm2-autokern1 on an x86_64
bl6-13.ltc.austin.ibm.com login: -- 0:conmux-control -- time-stamp -- Jun/05/06 10:47:46 -- -- 0:conmux-control -- time-stamp -- Jun/05/06 10:51:12 --

BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81268fc4>] icmp_rcv+0x17c/0x184
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160<7>Losing some ticks... checking if CPU frequency changed.
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81268fc4>] icmp_rcv+0x17c/0x184
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
Unable to handle kernel NULL pointer dereference at 0000000000000010 RIP: 
 [<ffffffff8108b063>] prepare_binprm+0xb/0xf4
PGD ccd9067 PUD e0ce067 PMD 0 
Oops: 0000 [1] SMP 
last sysfs file: /block/sda/sda1/size
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
CPU 2 
Modules linked in: ipv6 ppdev hidp rfcomm l2cap bluetooth video sony_acpi button battery asus_acpi ac lp parport_pc parport nvram
Pid: 18763, comm: sh Not tainted 2.6.17-rc5-mm2-autokern1 #1
RIP: 0010:[<ffffffff8108b063>]  [<ffffffff8108b063>] prepare_binprm+0xb/0xf4
RSP: 0018:ffff81000cb3ded8  EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff81003eae6800 RCX: 0000000000000000
RDX: 0000000000000000 RSI: 0000000000003fff RDI: ffff81003eae6800
RBP: ffff810029649c00 R08: ffff81000cb3c000 R09: 000000000000afa7
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
R13: ffff81001a618000 R14: 000000000070a560 R15: 000000000070b5c0
FS:  00002aba68ba6d30(0000) GS:ffff81003ffbe8c0(0000) knlGS:00000000f7f5a6b0
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000010 CR3: 000000000cdab000 CR4: 00000000000006e0
Process sh (pid: 18763, threadinfo ffff81000cb3c000, task ffff8100285437c0)
Stack: ffff81003eae6800 BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
ffffffff8108b63f ffff81000cb3df58 ffff81001a618000 
       000000000070a560 000000000070b5c0 ffff81001a618000 0000000000709620 
       0000000000000000 ffffffff81007f14 
Call Trace:
 [<ffffffff8108b63f>] do_execve+0x11d/0x24b
 [<ffffffff81007f14>] sys_execve+0x34/0x87
 [<ffffffff81009677>] stub_execve+0x67/0xb0
Code: 48 8b 41 10 48 8b 70 20 b8 f3 ff ff ff 0f b7 56 4c f6 c2 49 
RIP  [<ffffffff8108b063>] prepare_binprm+0xb/0xf4 RSP <ffff81000cb3ded8>
CR2: 0000000000000010
 BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff811d94a3>] sd_rw_intr+0x2a2/0x2b1
 [<ffffffff811cbe47>] scsi_device_unbusy+0x5d/0x77
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff811d94a3>] sd_rw_intr+0x2a2/0x2b1
 [<ffffffff811cbe47>] scsi_device_unbusy+0x5d/0x77
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff811d94a3>] sd_rw_intr+0x2a2/0x2b1
 [<ffffffff811cbe47>] scsi_device_unbusy+0x5d/0x77
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81252ce2>] tcp_recvmsg+0x622/0x7fb
 [<ffffffff8122720e>] sock_common_recvmsg+0x2d/0x44
 [<ffffffff81223aaf>] do_sock_read+0xc6/0xd1
 [<ffffffff81223bff>] sock_aio_read+0x4f/0x5e
 [<ffffffff8102bed5>] __wake_up+0x36/0x4d
 [<ffffffff81080bc8>] do_sync_read+0xc9/0x106
 [<ffffffff81045d20>] autoremove_wake_function+0x0/0x2e
 [<ffffffff81170577>] tty_ldisc_deref+0x65/0x77
 [<ffffffff81080ce9>] vfs_read+0xe4/0x172
 [<ffffffff81081037>] sys_read+0x45/0x6e
 [<ffffffff810092be>] system_call+0x7e/0x83
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81252ce2>] tcp_recvmsg+0x622/0x7fb
 [<ffffffff8122720e>] sock_common_recvmsg+0x2d/0x44
 [<ffffffff81223aaf>] do_sock_read+0xc6/0xd1
 [<ffffffff81223bff>] sock_aio_read+0x4f/0x5e
 [<ffffffff8102bed5>] __wake_up+0x36/0x4d
 [<ffffffff81080bc8>] do_sync_read+0xc9/0x106
 [<ffffffff81045d20>] autoremove_wake_function+0x0/0x2e
 [<ffffffff81170577>] tty_ldisc_deref+0x65/0x77
 [<ffffffff81080ce9>] vfs_read+0xe4/0x172
 [<ffffffff81081037>] sys_read+0x45/0x6e
 [<ffffffff810092be>] system_call+0x7e/0x83
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff811d94a3>] sd_rw_intr+0x2a2/0x2b1
 [<ffffffff811cbe47>] scsi_device_unbusy+0x5d/0x77
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff8125923f>] tcp_rcv_established+0xe3/0x71a
 [<ffffffff8126079d>] tcp_v4_do_rcv+0x2b/0x2ff
 [<ffffffff8126106d>] tcp_v4_rcv+0x5fc/0x996
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
BUG: warning at include/net/dst.h:153/dst_release()
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff811d94a3>] sd_rw_intr+0x2a2/0x2b1
 [<ffffffff811cbe47>] scsi_device_unbusy+0x5d/0x77
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
Call Trace:
 [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81252ce2>] tcp_recvmsg+0x622/0x7fb
 [<ffffffff8122720e>] sock_common_recvmsg+0x2d/0x44
 [<ffffffff81223aaf>] do_sock_read+0xc6/0xd1
 [<ffffffff81223bff>] sock_aio_read+0x4f/0x5e
 [<ffffffff8102bed5>] __wake_up+0x36/0x4d
 [<ffffffff81080bc8>] do_sync_read+0xc9/0x106BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
 [<ffffffff81045d20>] autoremove_wake_function+0x0/0x2e
 [<ffffffff81092210>] do_ioctl+0x64/0x6f
 [<ffffffff81080ce9>] vfs_read+0xe4/0x172
 [<ffffffff81081037>] sys_read+0x45/0x6e
 [<ffffffff810092be>] system_call+0x7e/0x83
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81252ce2>] tcp_recvmsg+0x622/0x7fb
 [<ffffffff8122720e>] sock_common_recvmsg+0x2d/0x44
 [<ffffffff81223aaf>] do_sock_read+0xc6/0xd1
 [<ffffffff81223bff>] sock_aio_read+0x4f/0x5e
 [<ffffffff8102bed5>] __wake_up+0x36/0x4d
 [<ffffffff81080bc8>] do_sync_read+0xc9/0x106
 [<ffffffff81045d20>] autoremove_wake_function+0x0/0x2e
 [<ffffffff81092210>] do_ioctl+0x64/0x6f
 [<ffffffff81080ce9>] vfs_read+0xe4/0x172
 [<ffffffff81081037>] sys_read+0x45/0x6e
 [<ffffffff810092be>] system_call+0x7e/0x83
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81252ce2>] tcp_recvmsg+0x622/0x7fb
 [<ffffffff8122720e>] sock_common_recvmsg+0x2d/0x44
 [<ffffffff81223aaf>] do_sock_read+0xc6/0xd1
 [<ffffffff81223bff>] sock_aio_read+0x4f/0x5e
 [<ffffffff8102bed5>] __wake_up+0x36/0x4d
 [<ffffffff81080bc8>] do_sync_read+0xc9/0x106
 [<ffffffff81045d20>] autoremove_wake_function+0x0/0x2e
 [<ffffffff81092210>] do_ioctl+0x64/0x6f
 [<ffffffff81080ce9>] vfs_read+0xe4/0x172BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
 [<ffffffff81081037>] sys_read+0x45/0x6e
 [<ffffffff810092be>] system_call+0x7e/0x83
general protection fault: 0000 [2] SMP 
last sysfs file: /block/sda/sda1/size
CPU 3 
Modules linked in: ipv6 ppdev hidp rfcomm l2cap bluetooth video sony_acpi button battery asus_acpi ac lp parport_pc parport nvram
Pid: 17887, comm: sshd Not tainted 2.6.17-rc5-mm2-autokern1 #1
RIP: 0010:[<ffffffff81228334>]  [<ffffffff81228334>] skb_drop_fraglist+0x17/0x26
RSP: 0018:ffff81000ef8dc48  EFLAGS: 00010206
RAX: 00000000026b2300 RBX: 4000000000000060 RCX: 000000000000b56c
RDX: ffff8100162f6900 RSI: ffffffff81321250 RDI: 4000000000000060
RBP: 0000000000000090 R08: 0000000000000005 R09: 0000000000000000
R10: 0000000000000000 R11: ffff81000ea78800 R12: 00000000ffffff95
R13: 0000000000000000 R14: ffff810034b84ac0 R15: 0000000000003f70
FS:  00002ae22a4babe0(0000) GS:ffff810037e0cdc0(0000) knlGS:00000000f7f0b6b0
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00000000005b9e5c CR3: 000000000f2ef000 CR4: 00000000000006e0
Process sshd (pid: 17887, threadinfo ffff81000ef8c000, task ffff810015501840)
Stack: ffff810034b84ac0 ffffffff812283c7 00000000000001d8 ffff810034b84ac0 
       0000000000000090 ffffffff812281c2 ffff81000ea78800 ffffffff81252ce2 
       0000000000000246 0000000100000000 
Call Trace:
 [<ffffffff812283c7>] skb_release_data+0x84/0x97
 [<ffffffff812281c2>] kfree_skbmem+0x9/0x7fBUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8103b598>] do_timer+0x9b/0x4bd
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
 [<ffffffff81252ce2>] tcp_recvmsg+0x622/0x7fb
 [<ffffffff8122720e>] sock_common_recvmsg+0x2d/0x44
 [<ffffffff81223aaf>] do_sock_read+0xc6/0xd1
 [<ffffffff81223bff>] sock_aio_read+0x4f/0x5e
 [<ffffffff8102bed5>] __wake_up+0x36/0x4d
 [<ffffffff81080bc8>] do_sync_read+0xc9/0x106
 [<ffffffff81045d20>] autoremove_wake_function+0x0/0x2e
 [<ffffffff81092210>] do_ioctl+0x64/0x6f
 [<ffffffff81080ce9>] vfs_read+0xe4/0x172
 [<ffffffff81081037>] sys_read+0x45/0x6e
 [<ffffffff810092be>] system_call+0x7e/0x83
Code: 48 8b 1b e8 b9 ff ff ff 48 85 db 75 f0 5b c3 55 53 48 89 fb 
RIP  [<ffffffff81228334>] skb_drop_fraglist+0x17/0x26 RSP <ffff81000ef8dc48>
 ----------- [cut here ] --------- [please bite here ] ---------
Kernel BUG at mm/slab.c:3430
invalid opcode: 0000 [3] SMP 
last sysfs file: /block/sda/sda1/size
CPU 1 
Modules linked in: ipv6 ppdev hidp rfcomm l2cap bluetooth video sony_acpi button battery asus_acpi ac lp parport_pc parport nvram
Pid: 6, comm: ksoftirqd/1 Not tainted 2.6.17-rc5-mm2-autokern1 #1
RIP: 0010:[<ffffffff8107d0a4>]  [<ffffffff8107d0a4>] kmem_cache_free+0x5f/0x77
RSP: 0018:ffff810037e1ff28  EFLAGS: 00010287
RAX: 0000000000000080 RBX: ffff81000cfc1480 RCX: 000000000000000a
RDX: ffff81000185c980 RSI: ffff81000e3a6c00 RDI: ffff8100026b2340
RBP: ffff8100024e7d40 R08: 0000000000000008 R09: 0000000000000000
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
R10: 0000000000000000 R11: ffff81003ff950d0 R12: 0000000000000008
R13: 0000000000000001 R14: ffffffff812b1104 R15: 0000000000000000
FS:  00002ae22a4babe0(0000) GS:ffff81003ff81340(0000) knlGS:00000000f7f5c6b0
CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 00000000005b9e5c CR3: 0000000001001000 CR4: 00000000000006e0
Process ksoftirqd/1 (pid: 6, threadinfo ffff810037e14000, task ffff81003ff950d0)
Stack: ffff81000cfc1480 ffffffff8104394c ffff8100024e7dc0 0000000000000000 
       ffffffff814cbc90 ffffffff810439ee 0000000000000000 ffffffff81037bc1 
       0000000000000001 ffffffff81476f90 
Call Trace:
 <IRQ> [<ffffffff8104394c>] __rcu_process_callbacks+0x12a/0x1ab
 [<ffffffff810439ee>] rcu_process_callbacks+0x21/0x42
 [<ffffffff81037bc1>] tasklet_action+0x69/0xa8
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff81037d1b>] ksoftirqd+0x0/0xbf
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 <EOI> [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff81037d84>] ksoftirqd+0x69/0xbf
 [<ffffffff81045806>] kthread+0x107/0x133
 [<ffffffff81037d1b>] ksoftirqd+0x0/0xbf
 [<ffffffff8100a146>] child_rip+0x8/0x12
 [<ffffffff81037d1b>] ksoftirqd+0x0/0xbf
 [<ffffffff810456ff>] kthread+0x0/0x133
 [<ffffffff8100a13e>] child_rip+0x0/0x12
Code: 0f 0b 68 41 56 2b 81 c2 66 0d 9c 5b fa 31 d2 e8 99 f9 ff ff 
RIP  [<ffffffff8107d0a4>] kmem_cache_free+0x5f/0x77BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>

Attempt to release alive inet socket ffff81003f8b2780
RSP <ffff810037e1ff28>
<3>BUG: sleeping function called from invalid context at include/linux/rwsem.h:53
in_atomic():1, irqs_disabled():0

Call Trace:
 <IRQ> [<ffffffff810299a0>] __might_sleep+0xc0/0xc2
 [<ffffffff8103f5a1>] blocking_notifier_call_chain+0x1f/0x4e
 [<ffffffff81034c96>] do_exit+0x22/0x8b2
 [<ffffffff8128a3a7>] _spin_unlock_irqrestore+0xb/0xd
 [<ffffffff8100aa61>] do_divide_error+0x0/0xa2
 [<ffffffff8128ad5e>] do_trap+0xe6/0xf3
 [<ffffffff8100ac90>] do_invalid_op+0x9b/0xa5
 [<ffffffff8107d0a4>] kmem_cache_free+0x5f/0x77
 [<ffffffff81009f8d>] error_exit+0x0/0x84
 [<ffffffff8107d0a4>] kmem_cache_free+0x5f/0x77
 [<ffffffff8104394c>] __rcu_process_callbacks+0x12a/0x1ab
 [<ffffffff810439ee>] rcu_process_callbacks+0x21/0x42
 [<ffffffff81037bc1>] tasklet_action+0x69/0xa8
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff81037d1b>] ksoftirqd+0x0/0xbf
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 <EOI> [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff81037d84>] ksoftirqd+0x69/0xbf
 [<ffffffff81045806>] kthread+0x107/0x133
 [<ffffffff81037d1b>] ksoftirqd+0x0/0xbf
 [<ffffffff8100a146>] child_rip+0x8/0x12
 [<ffffffff81037d1b>] ksoftirqd+0x0/0xbf
 [<ffffffff810456ff>] kthread+0x0/0x133
 [<ffffffff8100a13e>] child_rip+0x0/0x12
Kernel panic - not syncing: Aiee, killing interrupt handler!
Signature · 3 lines
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab
#27 lore raw
Quoted 4 lines · On Mon, 5 Jun 2006 18:56:37 +0100 [email protected] (Mel Gorman) wrote:

I am seeing more networking-related funniness with 2.6.17-rc5-mm3 on the same machine previously fixed by git-net-llc-fix.patch. The console log is below. I've done no investigation work in case it's a known problem.

It's not a known problem, afaik.

Quoted 30 lines

...
Starting anacron: [ OK ]
Starting atd: [ OK ]
Starting Avahi daemon: [ OK ]
Starting cups-config-daemon: [ OK ]
Starting HAL daemon: [ OK ]
Fedora Core release 5 (Bordeaux)
Kernel 2.6.17-rc5-mm2-autokern1 on an x86_64
bl6-13.ltc.austin.ibm.com login: -- 0:conmux-control -- time-stamp -- Jun/05/06 10:47:46 -- -- 0:conmux-control -- time-stamp -- Jun/05/06 10:51:12 --

BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81268fc4>] icmp_rcv+0x17c/0x184
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
Quoted 8 lines
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160<7>Losing some ticks... checking if CPU frequency changed.
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb

There are quite a few changes in the net tree. I guess the first thing to investigate would be 2.6.17-rc5+origin.patch+git-net.patch.

#28 lore raw
Quoted 49 lines · Andrew Morton
Quoted 4 lines · On Mon, 5 Jun 2006 18:56:37 +0100 [email protected] (Mel Gorman) wrote:

I am seeing more networking-related funniness with 2.6.17-rc5-mm3 on the same machine previously fixed by git-net-llc-fix.patch. The console log is below. I've done no investigation work in case it's a known problem.

It's not a known problem, afaik.

Quoted 30 lines

...
Starting anacron: [ OK ]
Starting atd: [ OK ]
Starting Avahi daemon: [ OK ]
Starting cups-config-daemon: [ OK ]
Starting HAL daemon: [ OK ]
Fedora Core release 5 (Bordeaux)
Kernel 2.6.17-rc5-mm2-autokern1 on an x86_64
bl6-13.ltc.austin.ibm.com login: -- 0:conmux-control -- time-stamp -- Jun/05/06 10:47:46 -- -- 0:conmux-control -- time-stamp -- Jun/05/06 10:51:12 --

BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81199568>] tg3_poll+0x1a1/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
 <EOI>
Attempt to release alive inet socket ffff81003f8b2780
BUG: warning at include/net/dst.h:153/dst_release()
Call Trace:
 <IRQ> [<ffffffff81228274>] __kfree_skb+0x3c/0xbd
 [<ffffffff81268fc4>] icmp_rcv+0x17c/0x184
 [<ffffffff812484ca>] ip_local_deliver+0xfe/0x1bf
 [<ffffffff812489bf>] ip_rcv+0x434/0x475
 [<ffffffff8122d615>] netif_receive_skb+0x2c6/0x2e5
Quoted 8 lines
 [<ffffffff81199add>] tg3_poll+0x716/0x94f
 [<ffffffff8122d80c>] net_rx_action+0xac/0x160<7>Losing some ticks... checking if CPU frequency changed.
 [<ffffffff81037904>] __do_softirq+0x48/0xb4
 [<ffffffff8100a496>] call_softirq+0x1e/0x28
 [<ffffffff8100b84e>] do_softirq+0x2c/0x7e
 [<ffffffff8100b6c8>] do_IRQ+0x50/0x59
 [<ffffffff81007807>] default_idle+0x0/0x54
 [<ffffffff810097b8>] ret_from_intr+0x0/0xb
Quoted 3 lines

There are quite a few changes in the net tree. I guess the first thing to investigate would be 2.6.17-rc5+origin.patch+git-net.patch.

That survived long enough to build a kernel, but backing out git-net on top of mm like I did for the LLC bug also survived. Not sure what is going on.

Signature · 3 lines
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab
#29 lore raw
In reply to Andrew Morton
Quoted 4 lines · Andrew Morton
Quoted 1 line · On Mon, 5 Jun 2006 18:56:37 +0100 [email protected] (Mel Gorman) wrote:
Quoted 14 lines
Quoted 3 lines

I am seeing more networking-related funniness with 2.6.17-rc5-mm3 on the same machine previously fixed by git-net-llc-fix.patch. The console log is below. I've done no investigation work in case it's a known problem.

It's not a known problem, afaik.

Quoted 8 lines

...
Starting anacron: [ OK ]
Starting atd: [ OK ]
Starting Avahi daemon: [ OK ]
Starting cups-config-daemon: [ OK ]
Starting HAL daemon: [ OK ]
Fedora Core release 5 (Bordeaux)
Kernel 2.6.17-rc5-mm2-autokern1 on an x86_64

Bah, I'm a spanner. The patches I was testing were rebased to the latest -mm, but the kernel version they were then tested on was not changed. This was probably the LLC bug with a different shaped error and the first set of tests are passing with -mm3. Sorry for the noise.

Signature · 3 lines
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab
#30 lore raw
In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

Thought I'd try my bi-annual "poke at -mm". Results were less than spectacular.

http://www.codemonkey.org.uk/junk/DSC00347.JPG
First the sound driver oopsed.

Then, the whole thing locked up after probing the parallel port. I disabled it in the BIOS, and then it locked up probing the floppy drive..
http://www.codemonkey.org.uk/junk/DSC00348.JPG

System is still alive, and responds to keyboard, but makes no forward progress.

(sysrq-B spewed a lockdep trace and then rebooted. I'll try and get that hooked up to a serial console)

On a whim, I enabled the floppy drive in the BIOS, and rebooted. That got me here. http://www.codemonkey.org.uk/junk/DSC00349.JPG
Same dead userspace.

Off to find a serial cable.

Dave

Signature · 1 line
http://www.codemonkey.org.uk
#31 lore raw
Quoted 6 lines · On Mon, 5 Jun 2006 15:48:45 -0400 Dave Jones <[email protected]> wrote:
Quoted 5 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

Quoted 8 lines
Quoted 2 lines

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

Thought I'd try my bi-annual "poke at -mm". Results were less than spectacular.

http://www.codemonkey.org.uk/junk/DSC00347.JPG
First the sound driver oopsed.

That's a bug in sound/pci/cs4281.c.

There's a debug patch in -mm ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/broken-out/debug-shared-irqs.patch which trips up drivers which request an IRQ before their IRQ handler is
ready to accept IRQs (they'll crash in real life if the IRQ is shared).

Quoted 3 lines

Then, the whole thing locked up after probing the parallel port. I disabled it in the BIOS, and then it locked up probing the floppy drive..
http://www.codemonkey.org.uk/junk/DSC00348.JPG

That looks like the same thing?

Quoted 8 lines

System is still alive, and responds to keyboard, but makes no forward progress.

(sysrq-B spewed a lockdep trace and then rebooted. I'll try and get that hooked up to a serial console)

On a whim, I enabled the floppy drive in the BIOS, and rebooted. That got me here. http://www.codemonkey.org.uk/junk/DSC00349.JPG
Same dead userspace.

So does that.

Quoted 1 line

Off to find a serial cable.

Try reverting debug-shared-irqs.patch, or disable the sound driver?

#32 lore raw
Quoted 8 lines · Andrew Morton
Quoted 3 lines

Then, the whole thing locked up after probing the parallel port. I disabled it in the BIOS, and then it locked up probing the floppy drive..
http://www.codemonkey.org.uk/junk/DSC00348.JPG

That looks like the same thing?

Quoted 2 lines

System is still alive, and responds to keyboard, but makes no forward progress.

Quoted 8 lines
Quoted 6 lines

(sysrq-B spewed a lockdep trace and then rebooted. I'll try and get that hooked up to a serial console)

On a whim, I enabled the floppy drive in the BIOS, and rebooted. That got me here. http://www.codemonkey.org.uk/junk/DSC00349.JPG
Same dead userspace.

So does that.

The top half the screen is the same as the first pic yes, but the purpose of those latter two pics was to show that we're locking up (in aparently different places) shortly afterwards.

Quoted 1 line

Try reverting debug-shared-irqs.patch, or disable the sound driver?

Will turn off the sound driver, and see what happens.

Dave

Signature · 1 line
http://www.codemonkey.org.uk
#33 lore raw
Quoted 2 lines · Dave Jones
Quoted 1 line

Try reverting debug-shared-irqs.patch, or disable the sound driver?

Will turn off the sound driver, and see what happens.

Win! It now boots. I blew it up really easy with a socket-fuzzer though.
(http://people.redhat.com/davej/sfuzz.c)

[  874.865028] ======================================
[  874.943738] [ BUG: bad unlock ordering detected! ]
[  875.002919] --------------------------------------
[  875.062134] sfuzz/23915 is trying to release lock (&sctp_port_alloc_lock) at:
[  875.149619]  [<d128ed4e>] sctp_get_port_local+0xd0/0x285 [sctp]
[  875.222636] but the next lock to release is:
[  875.276019]  (&sctp_port_hashtable[i].lock){-...}, at: [<d128ed0e>] sctp_get_port_local+0x90/0x285 [sctp]
[  875.393031]
[  875.393032] other info that might help us debug this:
[  875.476583] 1 locks held by sfuzz/23915:
[  875.526247]  #0:  (&sctp_port_alloc_lock){-...}, at: [<d128ecd9>] sctp_get_port_local+0x5b/0x285 [sctp]
[  875.641621]
[  875.641623] stack backtrace:
[  875.699891]  [<c0104966>] show_trace_log_lvl+0x54/0xfd
[  875.764425]  [<c0104f1a>] show_trace+0xd/0x10
[  875.819622]  [<c010502f>] dump_stack+0x19/0x1b
[  875.875924]  [<c013b4af>] lockdep_release+0x150/0x2d1
[  875.939610]  [<c032341e>] _spin_unlock+0x16/0x20
[  875.998171]  [<d128ed4e>] sctp_get_port_local+0xd0/0x285 [sctp]
[  876.072345]  [<d128efd4>] sctp_do_bind+0x9a/0x158 [sctp]
[  876.139315]  [<d128f0ce>] sctp_autobind+0x3c/0x44 [sctp]
[  876.206310]  [<d129253d>] sctp_inet_listen+0xe9/0x139 [sctp]
[  876.277539]  [<c02c20af>] sys_listen+0x4a/0x65
[  876.334730]  [<c02c308d>] sys_socketcall+0x98/0x186
[  876.397175]  [<c03239cb>] syscall_call+0x7/0xb
Signature · 1 line
http://www.codemonkey.org.uk
#34 lore raw
Quoted 6 lines · On Mon, 5 Jun 2006 16:44:56 -0400 Dave Jones <[email protected]> wrote:
Quoted 2 lines · Dave Jones
Quoted 1 line

Try reverting debug-shared-irqs.patch, or disable the sound driver?

Will turn off the sound driver, and see what happens.

Win! It now boots.

So does Windows 95.

Quoted 2 lines

I blew it up really easy with a socket-fuzzer though.
(http://people.redhat.com/davej/sfuzz.c)

But it kept running OK, yes?

Quoted 17 lines
[  874.865028] ======================================
[  874.943738] [ BUG: bad unlock ordering detected! ]
[  875.002919] --------------------------------------
[  875.062134] sfuzz/23915 is trying to release lock (&sctp_port_alloc_lock) at:
[  875.149619]  [<d128ed4e>] sctp_get_port_local+0xd0/0x285 [sctp]
[  875.222636] but the next lock to release is:
[  875.276019]  (&sctp_port_hashtable[i].lock){-...}, at: [<d128ed0e>] sctp_get_port_local+0x90/0x285 [sctp]
[  875.393031]
[  875.393032] other info that might help us debug this:
[  875.476583] 1 locks held by sfuzz/23915:
[  875.526247]  #0:  (&sctp_port_alloc_lock){-...}, at: [<d128ecd9>] sctp_get_port_local+0x5b/0x285 [sctp]
[  875.641621]
[  875.641623] stack backtrace:
[  875.699891]  [<c0104966>] show_trace_log_lvl+0x54/0xfd
[  875.764425]  [<c0104f1a>] show_trace+0xd/0x10
[  875.819622]  [<c010502f>] dump_stack+0x19/0x1b
[  875.875924]  [<c013b4af>] lockdep_release+0x150/0x2d1
Quoted 8 lines
[  875.939610]  [<c032341e>] _spin_unlock+0x16/0x20
[  875.998171]  [<d128ed4e>] sctp_get_port_local+0xd0/0x285 [sctp]
[  876.072345]  [<d128efd4>] sctp_do_bind+0x9a/0x158 [sctp]
[  876.139315]  [<d128f0ce>] sctp_autobind+0x3c/0x44 [sctp]
[  876.206310]  [<d129253d>] sctp_inet_listen+0xe9/0x139 [sctp]
[  876.277539]  [<c02c20af>] sys_listen+0x4a/0x65
[  876.334730]  [<c02c308d>] sys_socketcall+0x98/0x186
[  876.397175]  [<c03239cb>] syscall_call+0x7/0xb

This is a really really fussy "BUG", IMO. So we undid the locks in an inappropriate order - big deal.

But often these _are_ things which we should tune up, as an efficiency thing, so it is interesting to hear about them. But calling it a "BUG" is a bit alarmist.

Thanks for booting it.

#35 lore raw
Quoted 4 lines · Andrew Morton
Quoted 3 lines
Quoted 2 lines
Quoted 1 line

Try reverting debug-shared-irqs.patch, or disable the sound driver?

Will turn off the sound driver, and see what happens.

Win! It now boots.

So does Windows 95.

Hey, it's my turn to play "optimist" today. :)

Quoted 4 lines
Quoted 2 lines

I blew it up really easy with a socket-fuzzer though.
(http://people.redhat.com/davej/sfuzz.c)

But it kept running OK, yes?

Yep, still ticking along (for now).

Quoted 26 lines
Quoted 17 lines
[  874.865028] ======================================
[  874.943738] [ BUG: bad unlock ordering detected! ]
[  875.002919] --------------------------------------
[  875.062134] sfuzz/23915 is trying to release lock (&sctp_port_alloc_lock) at:
[  875.149619]  [<d128ed4e>] sctp_get_port_local+0xd0/0x285 [sctp]
[  875.222636] but the next lock to release is:
[  875.276019]  (&sctp_port_hashtable[i].lock){-...}, at: [<d128ed0e>] sctp_get_port_local+0x90/0x285 [sctp]
[  875.393031]
[  875.393032] other info that might help us debug this:
[  875.476583] 1 locks held by sfuzz/23915:
[  875.526247]  #0:  (&sctp_port_alloc_lock){-...}, at: [<d128ecd9>] sctp_get_port_local+0x5b/0x285 [sctp]
[  875.641621]
[  875.641623] stack backtrace:
[  875.699891]  [<c0104966>] show_trace_log_lvl+0x54/0xfd
[  875.764425]  [<c0104f1a>] show_trace+0xd/0x10
[  875.819622]  [<c010502f>] dump_stack+0x19/0x1b
[  875.875924]  [<c013b4af>] lockdep_release+0x150/0x2d1
Quoted 8 lines
[  875.939610]  [<c032341e>] _spin_unlock+0x16/0x20
[  875.998171]  [<d128ed4e>] sctp_get_port_local+0xd0/0x285 [sctp]
[  876.072345]  [<d128efd4>] sctp_do_bind+0x9a/0x158 [sctp]
[  876.139315]  [<d128f0ce>] sctp_autobind+0x3c/0x44 [sctp]
[  876.206310]  [<d129253d>] sctp_inet_listen+0xe9/0x139 [sctp]
[  876.277539]  [<c02c20af>] sys_listen+0x4a/0x65
[  876.334730]  [<c02c308d>] sys_socketcall+0x98/0x186
[  876.397175]  [<c03239cb>] syscall_call+0x7/0xb
Quoted 6 lines

This is a really really fussy "BUG", IMO. So we undid the locks in an inappropriate order - big deal.

But often these _are_ things which we should tune up, as an efficiency thing, so it is interesting to hear about them. But calling it a "BUG" is a bit alarmist.

Maybe so, but it's still pretty grotty though.

Dave

Signature · 1 line
http://www.codemonkey.org.uk
#36 lore raw
In reply to Dave Jones
Quoted 8 lines · Dave Jones
Quoted 2 lines · Dave Jones
Quoted 1 line

Try reverting debug-shared-irqs.patch, or disable the sound driver?

Will turn off the sound driver, and see what happens.

Win! It now boots. I blew it up really easy with a socket-fuzzer though.
(http://people.redhat.com/davej/sfuzz.c)

Quoted 8 lines
[  874.865028] ======================================
[  874.943738] [ BUG: bad unlock ordering detected! ]
[  875.002919] --------------------------------------
[  875.062134] sfuzz/23915 is trying to release lock (&sctp_port_alloc_lock) at:
[  875.149619]  [<d128ed4e>] sctp_get_port_local+0xd0/0x285 [sctp]
[  875.222636] but the next lock to release is:
[  875.276019]  (&sctp_port_hashtable[i].lock){-...}, at: [<d128ed0e>] sctp_get_port_local+0x90/0x285 [sctp]
[  875.393031]

this is "interesting" code to follow but it looks like a honest case of deliberate out of order unlock

---
 net/sctp/socket.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/sctp/socket.c +1 -1
Index: linux-2.6.17-rc5-mm3/net/sctp/socket.c
===================================================================
@@ -4597,7 +4597,7 @@ static long sctp_get_port_local(struct s
 			sctp_spin_unlock(&head->lock);
 		} while (--remaining > 0);
 		sctp_port_rover = rover;
-		sctp_spin_unlock(&sctp_port_alloc_lock);
+		spin_unlock_non_nested(&sctp_port_alloc_lock);
 
 		/* Exhausted local port range during search? */
 		ret = 1;
#37 lore raw
In reply to Andrew Morton

At Mon, 5 Jun 2006 13:06:26 -0700,

Quoted 19 lines · Andrew Morton
Quoted 6 lines · On Mon, 5 Jun 2006 15:48:45 -0400 Dave Jones <[email protected]> wrote:
Quoted 5 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

Quoted 8 lines
Quoted 2 lines

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

Thought I'd try my bi-annual "poke at -mm". Results were less than spectacular.

http://www.codemonkey.org.uk/junk/DSC00347.JPG
First the sound driver oopsed.

Quoted 6 lines

That's a bug in sound/pci/cs4281.c.

There's a debug patch in -mm ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/broken-out/debug-shared-irqs.patch which trips up drivers which request an IRQ before their IRQ handler is
ready to accept IRQs (they'll crash in real life if the IRQ is shared).

I guess that the bug in cs4281 is ioremap too lately issued after the registration of irq handler.

Does the patch below fix the problem?

Takashi

[PATCH] Fix possible Oops in cs4281 irq handler

Call ioremap before request_irq for avoiding possible Oops in cs4281 driver.

---

sound/pci/cs4281.c +7 -7
diff -r 84d14cbbd713 sound/pci/cs4281.c
@@ -1379,6 +1379,13 @@ static int __devinit snd_cs4281_create(s
 	chip->ba0_addr = pci_resource_start(pci, 0);
 	chip->ba1_addr = pci_resource_start(pci, 1);
 
+	chip->ba0 = ioremap_nocache(chip->ba0_addr, pci_resource_len(pci, 0));
+	chip->ba1 = ioremap_nocache(chip->ba1_addr, pci_resource_len(pci, 1));
+	if (!chip->ba0 || !chip->ba1) {
+		snd_cs4281_free(chip);
+		return -ENOMEM;
+	}
+	
 	if (request_irq(pci->irq, snd_cs4281_interrupt, SA_INTERRUPT|SA_SHIRQ,
 			"CS4281", chip)) {
 		snd_printk(KERN_ERR "unable to grab IRQ %d\n", pci->irq);
@@ -1387,13 +1394,6 @@ static int __devinit snd_cs4281_create(s
 	}
 	chip->irq = pci->irq;
 
-	chip->ba0 = ioremap_nocache(chip->ba0_addr, pci_resource_len(pci, 0));
-	chip->ba1 = ioremap_nocache(chip->ba1_addr, pci_resource_len(pci, 1));
-	if (!chip->ba0 || !chip->ba1) {
-		snd_cs4281_free(chip);
-		return -ENOMEM;
-	}
-	
 	tmp = snd_cs4281_chip_init(chip);
 	if (tmp) {
 		snd_cs4281_free(chip);
#38 lore raw
In reply to Andrew Morton
Quoted 2 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol alloc_hdlcdev
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_close
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_set_carrier
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol register_hdlc_device
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_open
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_ioctl
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol unregister_hdlc_device

(19:02:21:root@northwood:mm3)# grep SYNCLINK .config
CONFIG_SYNCLINK_CS=m
CONFIG_SYNCLINK_CS_HDLC=y
(19:02:25:root@northwood:mm3)# grep HDLC .config
CONFIG_HDLC=m
# CONFIG_HDLC_RAW is not set
# CONFIG_HDLC_RAW_ETH is not set
# CONFIG_HDLC_CISCO is not set
# CONFIG_HDLC_FR is not set
# CONFIG_HDLC_PPP is not set
CONFIG_HISAX_HDLC=y
CONFIG_SYNCLINK_CS_HDLC=y

Dave

Signature · 1 line
http://www.codemonkey.org.uk
#39 lore raw
Quoted 12 lines · On Mon, 5 Jun 2006 19:02:48 -0400 Dave Jones wrote:
Quoted 2 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol alloc_hdlcdev
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_close
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_set_carrier
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol register_hdlc_device
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_open
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_ioctl
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol unregister_hdlc_device

Quoted 12 lines
(19:02:21:root@northwood:mm3)# grep SYNCLINK .config
CONFIG_SYNCLINK_CS=m
CONFIG_SYNCLINK_CS_HDLC=y
(19:02:25:root@northwood:mm3)# grep HDLC .config
CONFIG_HDLC=m
# CONFIG_HDLC_RAW is not set
# CONFIG_HDLC_RAW_ETH is not set
# CONFIG_HDLC_CISCO is not set
# CONFIG_HDLC_FR is not set
# CONFIG_HDLC_PPP is not set
CONFIG_HISAX_HDLC=y
CONFIG_SYNCLINK_CS_HDLC=y

Those Kconfig + Makefiles are quite ugly to me. I would rather see SYNCLINK depend on HDLC rather than using some tricks to SELECT HDLC. And then it selects HDLC (and HDLC depends on WAN), but (in my case) WAN was not enabled, and doing "SELECT HDLC" did not enable WAN.

Adding SELECT WAN and changing the hdlc (wan) Makefile to use obj-m or obj-y (it was ONLY obj-y for hdlc) fixes^W makes it build with no missing symbols. However, I'll also see about a fix that uses "depends on HDLC" instead of "selects HDLC".

---
From: Randy Dunlap <[email protected]>

Fix many missing hdlc_generic symbols when CONFIG_HDLC=m. When Selecting HDLC, also Select WAN.
Fix Makefile to build for HDLC=y or HDLC=m.

---
 drivers/char/Kconfig     |    3 +++
 drivers/net/wan/Makefile |    8 ++++++--
 2 files changed, 9 insertions(+), 2 deletions(-)
drivers/net/wan/Makefile +6 -2
@@ -9,14 +9,18 @@ cyclomx-y                       := cycx_
 cyclomx-$(CONFIG_CYCLOMX_X25)	+= cycx_x25.o
 cyclomx-objs			:= $(cyclomx-y)  
 
-hdlc-y				:= hdlc_generic.o
+hdlc-$(CONFIG_HDLC)		:= hdlc_generic.o
 hdlc-$(CONFIG_HDLC_RAW)		+= hdlc_raw.o
 hdlc-$(CONFIG_HDLC_RAW_ETH)	+= hdlc_raw_eth.o
 hdlc-$(CONFIG_HDLC_CISCO)	+= hdlc_cisco.o
 hdlc-$(CONFIG_HDLC_FR)		+= hdlc_fr.o
 hdlc-$(CONFIG_HDLC_PPP)		+= hdlc_ppp.o
 hdlc-$(CONFIG_HDLC_X25)		+= hdlc_x25.o
-hdlc-objs			:= $(hdlc-y)
+ifeq ($(CONFIG_HDLC),y)
+  hdlc-objs			:= $(hdlc-y)
+else
+  hdlc-objs			:= $(hdlc-m)
+endif
 
 pc300-y				:= pc300_drv.o
 pc300-$(CONFIG_PC300_MLPPP)	+= pc300_tty.o
drivers/char/Kconfig +3 -0
@@ -197,6 +197,7 @@ config ISI
 config SYNCLINK
 	tristate "SyncLink PCI/ISA support"
 	depends on SERIAL_NONSTANDARD && PCI && ISA_DMA_API
+	select WAN if SYNCLINK_HDLC
 	select HDLC if SYNCLINK_HDLC
 	help
 	  Driver for SyncLink ISA and PCI synchronous serial adapters.
@@ -214,6 +215,7 @@ config SYNCLINK_HDLC
 config SYNCLINKMP
 	tristate "SyncLink Multiport support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINKMP_HDLC
 	select HDLC if SYNCLINKMP_HDLC
 	help
 	  Driver for SyncLink Multiport (2 or 4 ports) PCI synchronous serial adapter.
@@ -231,6 +233,7 @@ config SYNCLINKMP_HDLC
 config SYNCLINK_GT
 	tristate "SyncLink GT/AC support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINK_GT_HDLC
 	select HDLC if SYNCLINK_GT_HDLC
 	help
 	  Support for SyncLink GT and SyncLink AC families of
#40 lore raw
Quoted 9 lines · Randy.Dunlap

Those Kconfig + Makefiles are quite ugly to me. I would rather see SYNCLINK depend on HDLC rather than using some tricks to SELECT HDLC. And then it selects HDLC (and HDLC depends on WAN), but (in my case) WAN was not enabled, and doing "SELECT HDLC" did not enable WAN.

Adding SELECT WAN and changing the hdlc (wan) Makefile to use obj-m or obj-y (it was ONLY obj-y for hdlc) fixes^W makes it build with no missing symbols. However, I'll also see about a fix that uses "depends on HDLC" instead of "selects HDLC".

Generic HDLC support in the synclink drivers is optional. Should the generic HDLC code be enabled even if it is not used?

Some of our customers would scream if we started forcing them to compile and load unused code.

drivers/char/Kconfig config ISI · line 200

Quoted 5 lines

Fix many missing hdlc_generic symbols when CONFIG_HDLC=m. When Selecting HDLC, also Select WAN.
Fix Makefile to build for HDLC=y or HDLC=m.

drivers/char/Kconfig +1 -0
+	select WAN if SYNCLINK_HDLC

If this is the accepted approach, then synclink_cs should be added also. (drivers/char/pcmcia)

What about select WAN if HDLC instead?
Or does kbuild not propogate the reverse dependency? (SYNCLINK_HDLC selects HDLC, HDLC selects WAN)

--
Paul

#41 lore raw
Quoted 11 lines · On Mon, 05 Jun 2006 20:54:51 -0500 Paul Fulghum wrote:
Quoted 9 lines · Randy.Dunlap

Those Kconfig + Makefiles are quite ugly to me. I would rather see SYNCLINK depend on HDLC rather than using some tricks to SELECT HDLC. And then it selects HDLC (and HDLC depends on WAN), but (in my case) WAN was not enabled, and doing "SELECT HDLC" did not enable WAN.

Adding SELECT WAN and changing the hdlc (wan) Makefile to use obj-m or obj-y (it was ONLY obj-y for hdlc) fixes^W makes it build with no missing symbols. However, I'll also see about a fix that uses "depends on HDLC" instead of "selects HDLC".

Quoted 5 lines

Generic HDLC support in the synclink drivers is optional. Should the generic HDLC code be enabled even if it is not used?

Some of our customers would scream if we started forcing them to compile and load unused code.

OK, I'll try to allow for that.

Quoted 8 lines

drivers/char/Kconfig config ISI · line 200

Quoted 5 lines

Fix many missing hdlc_generic symbols when CONFIG_HDLC=m. When Selecting HDLC, also Select WAN.
Fix Makefile to build for HDLC=y or HDLC=m.

drivers/char/Kconfig +1 -0
+	select WAN if SYNCLINK_HDLC

If this is the accepted approach, then synclink_cs should be added also. (drivers/char/pcmcia)

It's not the desired approach AFAIK, but it may be the only reasonable one. I'm still testing alternatives, but you are welcome to take over and fix it. :)

Quoted 3 lines

What about select WAN if HDLC instead?
Or does kbuild not propogate the reverse dependency? (SYNCLINK_HDLC selects HDLC, HDLC selects WAN)

OK.

---
~Randy

#42 lore raw
Quoted 34 lines · On Mon, 5 Jun 2006 19:03:55 -0700 Randy.Dunlap wrote:
Quoted 11 lines · On Mon, 05 Jun 2006 20:54:51 -0500 Paul Fulghum wrote:
Quoted 9 lines · Randy.Dunlap

Those Kconfig + Makefiles are quite ugly to me. I would rather see SYNCLINK depend on HDLC rather than using some tricks to SELECT HDLC. And then it selects HDLC (and HDLC depends on WAN), but (in my case) WAN was not enabled, and doing "SELECT HDLC" did not enable WAN.

Adding SELECT WAN and changing the hdlc (wan) Makefile to use obj-m or obj-y (it was ONLY obj-y for hdlc) fixes^W makes it build with no missing symbols. However, I'll also see about a fix that uses "depends on HDLC" instead of "selects HDLC".

Quoted 5 lines

Generic HDLC support in the synclink drivers is optional. Should the generic HDLC code be enabled even if it is not used?

Some of our customers would scream if we started forcing them to compile and load unused code.

OK, I'll try to allow for that.

Quoted 8 lines

Review

Quoted 5 lines

Fix many missing hdlc_generic symbols when CONFIG_HDLC=m. When Selecting HDLC, also Select WAN.
Fix Makefile to build for HDLC=y or HDLC=m.

(quoted diff) +1 -0
+	select WAN if SYNCLINK_HDLC

If this is the accepted approach, then synclink_cs should be added also. (drivers/char/pcmcia)

It's not the desired approach AFAIK, but it may be the only reasonable one. I'm still testing alternatives, but you are welcome to take over and fix it. :)

Quoted 5 lines
Quoted 3 lines

What about select WAN if HDLC instead?
Or does kbuild not propogate the reverse dependency? (SYNCLINK_HDLC selects HDLC, HDLC selects WAN)

OK.

Hi Paul,
Here's another version of the patch for you to consider.
---

From: Randy Dunlap <[email protected]>

Fix missing symbol references to hdlc_generic functions. Switch SYNCLINK drivers from using SELECT to using DEPENDS for HDLC.

---
 drivers/char/Kconfig        |    9 +++------
 drivers/char/pcmcia/Kconfig |    3 +--
 drivers/net/wan/Makefile    |    8 ++++++--
 3 files changed, 10 insertions(+), 10 deletions(-)
drivers/char/Kconfig +3 -6
@@ -197,7 +197,6 @@ config ISI
 config SYNCLINK
 	tristate "SyncLink PCI/ISA support"
 	depends on SERIAL_NONSTANDARD && PCI && ISA_DMA_API
-	select HDLC if SYNCLINK_HDLC
 	help
 	  Driver for SyncLink ISA and PCI synchronous serial adapters.
 	  These adapters are no longer in production and have
@@ -205,7 +204,7 @@ config SYNCLINK
 
 config SYNCLINK_HDLC
 	bool "Generic HDLC support for SyncLink driver"
-	depends on SYNCLINK
+	depends on SYNCLINK && HDLC
 	help
 	  Enable generic HDLC support for the SyncLink PCI/ISA driver.
 	  Generic HDLC implements multiple higher layer networking
@@ -214,7 +213,6 @@ config SYNCLINK_HDLC
 config SYNCLINKMP
 	tristate "SyncLink Multiport support"
 	depends on SERIAL_NONSTANDARD && PCI
-	select HDLC if SYNCLINKMP_HDLC
 	help
 	  Driver for SyncLink Multiport (2 or 4 ports) PCI synchronous serial adapter.
 	  These adapters are no longer in production and have
@@ -222,7 +220,7 @@ config SYNCLINKMP
 
 config SYNCLINKMP_HDLC
 	bool "Generic HDLC support for SyncLink Multiport"
-	depends on SYNCLINKMP
+	depends on SYNCLINKMP && HDLC
 	help
 	  Enable generic HDLC support for the SyncLink Multiport driver.
 	  Generic HDLC implements multiple higher layer networking
@@ -231,7 +229,6 @@ config SYNCLINKMP_HDLC
 config SYNCLINK_GT
 	tristate "SyncLink GT/AC support"
 	depends on SERIAL_NONSTANDARD && PCI
-	select HDLC if SYNCLINK_GT_HDLC
 	help
 	  Support for SyncLink GT and SyncLink AC families of
 	  synchronous and asynchronous serial adapters
@@ -239,7 +236,7 @@ config SYNCLINK_GT
 
 config SYNCLINK_GT_HDLC
 	bool "Generic HDLC support for SyncLink GT/AC"
-	depends on SYNCLINK_GT
+	depends on SYNCLINK_GT && HDLC
 	help
 	  Enable generic HDLC support for the SyncLink GT/AC driver.
 	  Generic HDLC implements multiple higher layer networking
drivers/char/pcmcia/Kconfig +1 -2
@@ -8,13 +8,12 @@ menu "PCMCIA character devices"
 config SYNCLINK_CS
 	tristate "SyncLink PC Card support"
 	depends on PCMCIA
-	select HDLC if SYNCLINK_CS_HDLC
 	help
 	  Driver for SyncLink PC Card synchronous serial adapter.
 
 config SYNCLINK_CS_HDLC
 	bool "Generic HDLC support for SyncLink Multiport"
-	depends on SYNCLINK_CS
+	depends on SYNCLINK_CS && HDLC
 	help
 	  Enable generic HDLC support for the SyncLink PC Card driver.
 	  Generic HDLC implements multiple higher layer networking
drivers/net/wan/Makefile +6 -2
@@ -9,14 +9,18 @@ cyclomx-y                       := cycx_
 cyclomx-$(CONFIG_CYCLOMX_X25)	+= cycx_x25.o
 cyclomx-objs			:= $(cyclomx-y)  
 
-hdlc-y				:= hdlc_generic.o
+hdlc-$(CONFIG_HDLC)		:= hdlc_generic.o
 hdlc-$(CONFIG_HDLC_RAW)		+= hdlc_raw.o
 hdlc-$(CONFIG_HDLC_RAW_ETH)	+= hdlc_raw_eth.o
 hdlc-$(CONFIG_HDLC_CISCO)	+= hdlc_cisco.o
 hdlc-$(CONFIG_HDLC_FR)		+= hdlc_fr.o
 hdlc-$(CONFIG_HDLC_PPP)		+= hdlc_ppp.o
 hdlc-$(CONFIG_HDLC_X25)		+= hdlc_x25.o
-hdlc-objs			:= $(hdlc-y)
+ifeq ($(CONFIG_HDLC),y)
+  hdlc-objs			:= $(hdlc-y)
+else
+  hdlc-objs			:= $(hdlc-m)
+endif
 
 pc300-y				:= pc300_drv.o
 pc300-$(CONFIG_PC300_MLPPP)	+= pc300_tty.o
#43 lore raw
Quoted 2 lines · Randy.Dunlap

On Mon, 5 Jun 2006 19:03:55 -0700 Randy.Dunlap wrote:
Here's another version of the patch for you to consider.

This looks like the correct implementation of what I was trying (unsuccessfully) to do when the random config errors were first reported.

I'll do testing tomorrow to make sure I understand this completely.

Thanks,
Paul

#44 lore raw
In reply to Randy.Dunlap
Quoted 13 lines · Randy.Dunlap

Here's another version of the patch for you to consider.
---

drivers/char/Kconfig +0 -1
@@ -197,7 +197,6 @@ config ISI
 config SYNCLINK
 	tristate "SyncLink PCI/ISA support"
 	depends on SERIAL_NONSTANDARD && PCI && ISA_DMA_API
-	select HDLC if SYNCLINK_HDLC
 	help
 	  Driver for SyncLink ISA and PCI synchronous serial adapters.
 	  These adapters are no longer in production and have

drivers/char/Kconfig config SYNCLINK · line 210

Quoted 8 lines
drivers/char/Kconfig +1 -1
@@ -205,7 +204,7 @@ config SYNCLINK
 
 config SYNCLINK_HDLC
 	bool "Generic HDLC support for SyncLink driver"
-	depends on SYNCLINK
+	depends on SYNCLINK && HDLC
 	help
 	  Enable generic HDLC support for the SyncLink PCI/ISA driver.
 	  Generic HDLC implements multiple higher layer networking

Now I remember that this was tried before and does not work because SYNCLINK_HDLC is a bool and will force the HDLC module to 'y' even if the synclink driver is a 'm' which results in build errors.

I also tried 'depends on HDLC if SYNCLINK_HDLC' in the config SYNCLINK section, but that causes a cyclic dependency error. I suppose I could do that if I remove 'depends on SYNCLINK' from config SYNCLINK_HDLC. The only down side of that is the way the SYNCLINK_HDLC option would be displayed.

I'll review this again to find the best solution.

Signature · 2 lines
Paul Fulghum
Microgate Systems, Ltd.
#45 lore raw

PATCH 1/1 fix missing hdlc symbols for synclink drivers

In reply to Randy.Dunlap

Randy:

This patch of yours is the only way I've found to fix the random config build errors and maintain the necessary flexibility of configuration for the synclink drivers. I added the synclink_cs config to the end of your patch.

--

From: Randy Dunlap <[email protected]>

Fix many missing hdlc_generic symbols when CONFIG_HDLC=m. When Selecting HDLC, also Select WAN.
Fix Makefile to build for HDLC=y or HDLC=m.

drivers/net/wan/Makefile +6 -2
@@ -9,14 +9,18 @@ cyclomx-y                       := cycx_
 cyclomx-$(CONFIG_CYCLOMX_X25)	+= cycx_x25.o
 cyclomx-objs			:= $(cyclomx-y)  
 
-hdlc-y				:= hdlc_generic.o
+hdlc-$(CONFIG_HDLC)		:= hdlc_generic.o
 hdlc-$(CONFIG_HDLC_RAW)		+= hdlc_raw.o
 hdlc-$(CONFIG_HDLC_RAW_ETH)	+= hdlc_raw_eth.o
 hdlc-$(CONFIG_HDLC_CISCO)	+= hdlc_cisco.o
 hdlc-$(CONFIG_HDLC_FR)		+= hdlc_fr.o
 hdlc-$(CONFIG_HDLC_PPP)		+= hdlc_ppp.o
 hdlc-$(CONFIG_HDLC_X25)		+= hdlc_x25.o
-hdlc-objs			:= $(hdlc-y)
+ifeq ($(CONFIG_HDLC),y)
+  hdlc-objs			:= $(hdlc-y)
+else
+  hdlc-objs			:= $(hdlc-m)
+endif
 
 pc300-y				:= pc300_drv.o
 pc300-$(CONFIG_PC300_MLPPP)	+= pc300_tty.o
drivers/char/Kconfig +3 -0
@@ -197,6 +197,7 @@ config ISI
 config SYNCLINK
 	tristate "SyncLink PCI/ISA support"
 	depends on SERIAL_NONSTANDARD && PCI && ISA_DMA_API
+	select WAN if SYNCLINK_HDLC
 	select HDLC if SYNCLINK_HDLC
 	help
 	  Driver for SyncLink ISA and PCI synchronous serial adapters.
@@ -214,6 +215,7 @@ config SYNCLINK_HDLC
 config SYNCLINKMP
 	tristate "SyncLink Multiport support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINKMP_HDLC
 	select HDLC if SYNCLINKMP_HDLC
 	help
 	  Driver for SyncLink Multiport (2 or 4 ports) PCI synchronous serial adapter.
@@ -231,6 +233,7 @@ config SYNCLINKMP_HDLC
 config SYNCLINK_GT
 	tristate "SyncLink GT/AC support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINK_GT_HDLC
 	select HDLC if SYNCLINK_GT_HDLC
 	help
 	  Support for SyncLink GT and SyncLink AC families of
drivers/char/pcmcia/Kconfig +1 -0
@@ -8,6 +8,7 @@ menu "PCMCIA character devices"
 config SYNCLINK_CS
 	tristate "SyncLink PC Card support"
 	depends on PCMCIA
+	select WAN if SYNCLINK_CS_HDLC
 	select HDLC if SYNCLINK_CS_HDLC
 	help
 	  Driver for SyncLink PC Card synchronous serial adapter.
#46 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 9 lines · Paul Fulghum

+++ b/drivers/net/wan/Makefile 2006-06-06 14:08:53.000000000 -0500

drivers/net/wan/Makefile +1 -1
@@ -9,14 +9,18 @@ cyclomx-y                       := cycx_
 cyclomx-$(CONFIG_CYCLOMX_X25)	+= cycx_x25.o
 cyclomx-objs			:= $(cyclomx-y)  
 
-hdlc-y				:= hdlc_generic.o
+hdlc-$(CONFIG_HDLC)		:= hdlc_generic.o
 hdlc-$(CONFIG_HDLC_RAW)		+= hdlc_raw.o
 hdlc-$(CONFIG_HDLC_RAW_ETH)	+= hdlc_raw_eth.o

drivers/net/wan/Makefile cyclomx-y := cycx_ · line 23

Quoted 10 lines
drivers/net/wan/Makefile +5 -1
 hdlc-$(CONFIG_HDLC_CISCO)	+= hdlc_cisco.o
 hdlc-$(CONFIG_HDLC_FR)		+= hdlc_fr.o
 hdlc-$(CONFIG_HDLC_PPP)		+= hdlc_ppp.o
 hdlc-$(CONFIG_HDLC_X25)		+= hdlc_x25.o
-hdlc-objs			:= $(hdlc-y)
+ifeq ($(CONFIG_HDLC),y)
+  hdlc-objs			:= $(hdlc-y)
+else
+  hdlc-objs			:= $(hdlc-m)
+endif

How could that work? If CONFIG_HDLC=m and CONFIG_HDLC_*=y it would read:

hdlc-m := hdlc_generic.o
hdlc-y := hdlc_{raw,raw_eth,cisco,fr,ppp,x25}.o
hdlc-objs := $(hdlc-m)

CONFIG_HDLC is 3-state and CONFIG_HDLC_* are simple bools, everything makes a single module.

Not sure what missing symbols do you experience (never had synclink adapter) but almost certainly it's when generic HDLC is not selected (or is 'm' while synclink is 'y') and it's not related to the Makefile (i.e., while I don't exactly like the rest of the patch as I think enabling gHDLC should enable hw drivers - like with other drivers - it would probably work).

Signature · 1 line
Krzysztof Halasa
#47 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 3 lines · Krzysztof Halasa

(i.e., while I don't exactly like the rest of the patch as I think enabling gHDLC should enable hw drivers - like with other drivers - it would probably work).

If I took that approach then you could never use the synclink drivers without generic HDLC.

The synclink drivers can be used independent of the generic HDLC or with generic HDLC depending on space requirements and application.

If you can provide a patch that continues allowing that level of flexibility in a way more to your liking, then please post it and I'll have a look.

Signature · 2 lines
Paul Fulghum
Microgate Systems, Ltd.
#48 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 6 lines · Paul Fulghum

If I took that approach then you could never use the synclink drivers without generic HDLC.

The synclink drivers can be used independent of the generic HDLC or with generic HDLC depending on space requirements and application.

Ah, now I understand. Anyway, the Makefile is not related to the problem.

Quoted 3 lines

If you can provide a patch that continues allowing that level of flexibility in a way more to your liking, then please post it and I'll have a look.

I think I would modify the conditionals in drivers/char/synclink.c
- to assume that generic HDLC is not present if CONFIG_SYNCLINK=y and CONFIG_HDLC=m. I would even put a user-visible comment in the *config (like the one in generic HDLC menu if you have no X.25/LAPB).

I.e., I would probably change:

#ifdef CONFIG_HDLC_MODULE
#define CONFIG_HDLC 1
#endif

into something like:

#if defined(CONFIG_HDLC_MODULE) && defined(CONFIG_SYNCLINK_MODULE) #define CONFIG_HDLC 1
#endif

plus that comment in Kconfig.

(Not sure if the other synclink*.c files are relevant and need similar changes).

Signature · 1 line
Krzysztof Halasa
#49 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 26 lines · On Tue, 06 Jun 2006 23:09:42 +0200 Krzysztof Halasa wrote:
Quoted 6 lines · Paul Fulghum

If I took that approach then you could never use the synclink drivers without generic HDLC.

The synclink drivers can be used independent of the generic HDLC or with generic HDLC depending on space requirements and application.

Ah, now I understand. Anyway, the Makefile is not related to the problem.

Quoted 3 lines

If you can provide a patch that continues allowing that level of flexibility in a way more to your liking, then please post it and I'll have a look.

I think I would modify the conditionals in drivers/char/synclink.c
- to assume that generic HDLC is not present if CONFIG_SYNCLINK=y and CONFIG_HDLC=m. I would even put a user-visible comment in the *config (like the one in generic HDLC menu if you have no X.25/LAPB).

I.e., I would probably change:

#ifdef CONFIG_HDLC_MODULE
#define CONFIG_HDLC 1
#endif

Quoted 5 lines

into something like:

#if defined(CONFIG_HDLC_MODULE) && defined(CONFIG_SYNCLINK_MODULE) #define CONFIG_HDLC 1
#endif

So SYNCLINK has different capabilities depending on whether HDLC=m or HDLC=y ??
Makes no sense to me, but I'll leave it to you and Paul.

Quoted 6 lines

plus that comment in Kconfig.

(Not sure if the other synclink*.c files are relevant and need similar changes).

Signature · 1 line
Krzysztof Halasa

---
~Randy

#50 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 2 lines · Randy.Dunlap

So SYNCLINK has different capabilities depending on whether HDLC=m or HDLC=y ??

Only if SYNCLINK=y. Then if HDLC=m then it's unavailable.

When both are 'm' things should work. I'll check.

Signature · 1 line
Krzysztof Halasa
#51 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Randy.Dunlap
Quoted 6 lines · Randy.Dunlap
Quoted 5 lines · On Tue, 06 Jun 2006 23:09:42 +0200 Krzysztof Halasa wrote:

I.e., I would probably change:

#ifdef CONFIG_HDLC_MODULE
#define CONFIG_HDLC 1
#endif

These macros have already been replaced in the synclink drivers.

Quoted 2 lines

So SYNCLINK has different capabilities depending on whether HDLC=m or HDLC=y ??

No, the current situation is generic HDLC support is an optional configuration for the synclink drivers. The build option of generic HDLC must match that of the synclink driver if that option is enabled.

if synclink is 'y' with generic HDLC option enabled, then the generic HDLC module must be 'y'

if synclink is 'm' with generic HDLC option enabled, then generic HDLC can be 'y' or 'm'

if synclink is 'y' or 'm' without the generic HDLC option, then it is not necessary to build/load the generic HDLC module

--

The Kconfig addition of 'select WAN if SYNCLINK_XX_HDLC' enables the above behavior.

I thought the Makefile change was to allow generic HDLC to build as a module. I'll need to look at that closer.

--
Paul

#52 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Krzysztof Halasa
Quoted 31 lines · On Tue, 06 Jun 2006 22:27:40 +0200 Krzysztof Halasa wrote:
Quoted 9 lines · Paul Fulghum

+++ b/drivers/net/wan/Makefile 2006-06-06 14:08:53.000000000 -0500

drivers/net/wan/Makefile +1 -1
@@ -9,14 +9,18 @@ cyclomx-y                       := cycx_
 cyclomx-$(CONFIG_CYCLOMX_X25)	+= cycx_x25.o
 cyclomx-objs			:= $(cyclomx-y)  
 
-hdlc-y				:= hdlc_generic.o
+hdlc-$(CONFIG_HDLC)		:= hdlc_generic.o
 hdlc-$(CONFIG_HDLC_RAW)		+= hdlc_raw.o
 hdlc-$(CONFIG_HDLC_RAW_ETH)	+= hdlc_raw_eth.o

drivers/net/wan/Makefile cyclomx-y := cycx_ · line 23

Quoted 10 lines
drivers/net/wan/Makefile +5 -1
 hdlc-$(CONFIG_HDLC_CISCO)	+= hdlc_cisco.o
 hdlc-$(CONFIG_HDLC_FR)		+= hdlc_fr.o
 hdlc-$(CONFIG_HDLC_PPP)		+= hdlc_ppp.o
 hdlc-$(CONFIG_HDLC_X25)		+= hdlc_x25.o
-hdlc-objs			:= $(hdlc-y)
+ifeq ($(CONFIG_HDLC),y)
+  hdlc-objs			:= $(hdlc-y)
+else
+  hdlc-objs			:= $(hdlc-m)
+endif

How could that work? If CONFIG_HDLC=m and CONFIG_HDLC_*=y it would read:

hdlc-m := hdlc_generic.o
hdlc-y := hdlc_{raw,raw_eth,cisco,fr,ppp,x25}.o
hdlc-objs := $(hdlc-m)

CONFIG_HDLC is 3-state and CONFIG_HDLC_* are simple bools, everything makes a single module.

Quoted 6 lines

Not sure what missing symbols do you experience (never had synclink adapter) but almost certainly it's when generic HDLC is not selected (or is 'm' while synclink is 'y') and it's not related to the Makefile (i.e., while I don't exactly like the rest of the patch as I think enabling gHDLC should enable hw drivers - like with other drivers - it would probably work).

True, I didn't test every possible config. :)
I'll test some other possibilities...

---
~Randy

#53 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Krzysztof Halasa
Quoted 31 lines · On Tue, 06 Jun 2006 22:27:40 +0200 Krzysztof Halasa wrote:
Quoted 9 lines · Paul Fulghum

+++ b/drivers/net/wan/Makefile 2006-06-06 14:08:53.000000000 -0500

drivers/net/wan/Makefile +1 -1
@@ -9,14 +9,18 @@ cyclomx-y                       := cycx_
 cyclomx-$(CONFIG_CYCLOMX_X25)	+= cycx_x25.o
 cyclomx-objs			:= $(cyclomx-y)  
 
-hdlc-y				:= hdlc_generic.o
+hdlc-$(CONFIG_HDLC)		:= hdlc_generic.o
 hdlc-$(CONFIG_HDLC_RAW)		+= hdlc_raw.o
 hdlc-$(CONFIG_HDLC_RAW_ETH)	+= hdlc_raw_eth.o

drivers/net/wan/Makefile cyclomx-y := cycx_ · line 23

Quoted 10 lines
drivers/net/wan/Makefile +5 -1
 hdlc-$(CONFIG_HDLC_CISCO)	+= hdlc_cisco.o
 hdlc-$(CONFIG_HDLC_FR)		+= hdlc_fr.o
 hdlc-$(CONFIG_HDLC_PPP)		+= hdlc_ppp.o
 hdlc-$(CONFIG_HDLC_X25)		+= hdlc_x25.o
-hdlc-objs			:= $(hdlc-y)
+ifeq ($(CONFIG_HDLC),y)
+  hdlc-objs			:= $(hdlc-y)
+else
+  hdlc-objs			:= $(hdlc-m)
+endif

How could that work? If CONFIG_HDLC=m and CONFIG_HDLC_*=y it would read:

hdlc-m := hdlc_generic.o
hdlc-y := hdlc_{raw,raw_eth,cisco,fr,ppp,x25}.o
hdlc-objs := $(hdlc-m)

CONFIG_HDLC is 3-state and CONFIG_HDLC_* are simple bools, everything makes a single module.

Quoted 6 lines

Not sure what missing symbols do you experience (never had synclink adapter) but almost certainly it's when generic HDLC is not selected (or is 'm' while synclink is 'y') and it's not related to the Makefile (i.e., while I don't exactly like the rest of the patch as I think enabling gHDLC should enable hw drivers - like with other drivers - it would probably work).

Hi,
I think that the main problem is that SYNCLINK wants to be able to use some functions from hdlc_generic.c when CONFIG_HDLC=m. How do you handle that?

---
~Randy

#54 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 40 lines · On Tue, 6 Jun 2006 13:48:16 -0700 Randy.Dunlap wrote:
Quoted 31 lines · On Tue, 06 Jun 2006 22:27:40 +0200 Krzysztof Halasa wrote:
Quoted 9 lines · Paul Fulghum

+++ b/drivers/net/wan/Makefile 2006-06-06 14:08:53.000000000 -0500

drivers/net/wan/Makefile +1 -1
@@ -9,14 +9,18 @@ cyclomx-y                       := cycx_
 cyclomx-$(CONFIG_CYCLOMX_X25)	+= cycx_x25.o
 cyclomx-objs			:= $(cyclomx-y)  
 
-hdlc-y				:= hdlc_generic.o
+hdlc-$(CONFIG_HDLC)		:= hdlc_generic.o
 hdlc-$(CONFIG_HDLC_RAW)		+= hdlc_raw.o
 hdlc-$(CONFIG_HDLC_RAW_ETH)	+= hdlc_raw_eth.o

drivers/net/wan/Makefile cyclomx-y := cycx_ · line 23

Quoted 10 lines
drivers/net/wan/Makefile +5 -1
 hdlc-$(CONFIG_HDLC_CISCO)	+= hdlc_cisco.o
 hdlc-$(CONFIG_HDLC_FR)		+= hdlc_fr.o
 hdlc-$(CONFIG_HDLC_PPP)		+= hdlc_ppp.o
 hdlc-$(CONFIG_HDLC_X25)		+= hdlc_x25.o
-hdlc-objs			:= $(hdlc-y)
+ifeq ($(CONFIG_HDLC),y)
+  hdlc-objs			:= $(hdlc-y)
+else
+  hdlc-objs			:= $(hdlc-m)
+endif

How could that work? If CONFIG_HDLC=m and CONFIG_HDLC_*=y it would read:

hdlc-m := hdlc_generic.o
hdlc-y := hdlc_{raw,raw_eth,cisco,fr,ppp,x25}.o
hdlc-objs := $(hdlc-m)

CONFIG_HDLC is 3-state and CONFIG_HDLC_* are simple bools, everything makes a single module.

Quoted 6 lines

Not sure what missing symbols do you experience (never had synclink adapter) but almost certainly it's when generic HDLC is not selected (or is 'm' while synclink is 'y') and it's not related to the Makefile (i.e., while I don't exactly like the rest of the patch as I think enabling gHDLC should enable hw drivers - like with other drivers - it would probably work).

Quoted 4 lines

Hi,
I think that the main problem is that SYNCLINK wants to be able to use some functions from hdlc_generic.c when CONFIG_HDLC=m. How do you handle that?

Oh, using hdlc-y even for CONFIG_HDLC=m. :)

OK, it's still not working. As Dave Jones reported:

Quoted 7 lines

WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol alloc_hdlcdev
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_close
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_set_carrier
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol register_hdlc_device
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_open
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol hdlc_ioctl
WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/synclink_cs.ko needs unknown symbol unregister_hdlc_device

with:

Quoted 11 lines

CONFIG_SYNCLINK_CS=m
CONFIG_SYNCLINK_CS_HDLC=y

(19:02:25:root@northwood:mm3)# grep HDLC .config
CONFIG_HDLC=m
# CONFIG_HDLC_RAW is not set
# CONFIG_HDLC_RAW_ETH is not set
# CONFIG_HDLC_CISCO is not set
# CONFIG_HDLC_FR is not set
# CONFIG_HDLC_PPP is not set
CONFIG_HISAX_HDLC=y
CONFIG_SYNCLINK_CS_HDLC=y

Or enabling all SYNCLINK drivers =m with HDLC support:

WARNING: "register_hdlc_device" [drivers/char/synclinkmp.ko] undefined!
WARNING: "alloc_hdlcdev" [drivers/char/synclinkmp.ko] undefined!
WARNING: "hdlc_ioctl" [drivers/char/synclinkmp.ko] undefined!
WARNING: "hdlc_open" [drivers/char/synclinkmp.ko] undefined!
WARNING: "hdlc_close" [drivers/char/synclinkmp.ko] undefined!
WARNING: "hdlc_set_carrier" [drivers/char/synclinkmp.ko] undefined!
WARNING: "unregister_hdlc_device" [drivers/char/synclinkmp.ko] undefined!
WARNING: "register_hdlc_device" [drivers/char/synclink_gt.ko] undefined!
WARNING: "alloc_hdlcdev" [drivers/char/synclink_gt.ko] undefined!
WARNING: "hdlc_close" [drivers/char/synclink_gt.ko] undefined!
WARNING: "hdlc_open" [drivers/char/synclink_gt.ko] undefined!
WARNING: "hdlc_ioctl" [drivers/char/synclink_gt.ko] undefined!
WARNING: "hdlc_set_carrier" [drivers/char/synclink_gt.ko] undefined!
WARNING: "unregister_hdlc_device" [drivers/char/synclink_gt.ko] undefined!
WARNING: "register_hdlc_device" [drivers/char/synclink.ko] undefined!
WARNING: "alloc_hdlcdev" [drivers/char/synclink.ko] undefined!
WARNING: "hdlc_close" [drivers/char/synclink.ko] undefined!
WARNING: "hdlc_open" [drivers/char/synclink.ko] undefined!
WARNING: "hdlc_ioctl" [drivers/char/synclink.ko] undefined!
WARNING: "hdlc_set_carrier" [drivers/char/synclink.ko] undefined!
WARNING: "unregister_hdlc_device" [drivers/char/synclink.ko] undefined!
WARNING: "hdlc_close" [drivers/char/pcmcia/synclink_cs.ko] undefined!
WARNING: "hdlc_open" [drivers/char/pcmcia/synclink_cs.ko] undefined!
WARNING: "hdlc_ioctl" [drivers/char/pcmcia/synclink_cs.ko] undefined!
WARNING: "hdlc_set_carrier" [drivers/char/pcmcia/synclink_cs.ko] undefined!
WARNING: "unregister_hdlc_device" [drivers/char/pcmcia/synclink_cs.ko] undefined!
WARNING: "register_hdlc_device" [drivers/char/pcmcia/synclink_cs.ko] undefined!
WARNING: "alloc_hdlcdev" [drivers/char/pcmcia/synclink_cs.ko] undefined!

with partial .config:
# CONFIG_WAN is not set
CONFIG_HDLC=m
# CONFIG_HDLC_RAW is not set
# CONFIG_HDLC_RAW_ETH is not set
# CONFIG_HDLC_CISCO is not set
# CONFIG_HDLC_FR is not set
# CONFIG_HDLC_PPP is not set
...
CONFIG_SYNCLINK=m
CONFIG_SYNCLINK_HDLC=y
CONFIG_SYNCLINKMP=m
CONFIG_SYNCLINKMP_HDLC=y
CONFIG_SYNCLINK_GT=m
CONFIG_SYNCLINK_GT_HDLC=y
CONFIG_N_HDLC=m

so the hdlc_generic code isn't available for some reason.

---
~Randy

#55 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 4 lines · Randy.Dunlap

OK, it's still not working. As Dave Jones reported:

Quoted 2 lines

WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/
synclink_cs.ko needs unknown symbol alloc_hdlcdev

Looks like the only patch in mm3 touching it is:
fix-kbuild-dependencies-for-synclink-drivers.

BTW: selecting CONFIG_HDLC without selecting at least one protocol does nothing good (except that the thing actually builds). As I said, I'd rather let the user decide what it's needed.

Hmm, with 2.6.17-rc5-mm3 it builds fine here (~ up-to-date FC5, i386):

$ egrep '^C.*(SYNCLINK|HDLC)' .config
CONFIG_HDLC=m
CONFIG_SYNCLINK=m
CONFIG_SYNCLINK_HDLC=y
CONFIG_SYNCLINKMP=m
CONFIG_SYNCLINKMP_HDLC=y
CONFIG_SYNCLINK_GT=m
CONFIG_SYNCLINK_GT_HDLC=y
CONFIG_SYNCLINK_CS=m
CONFIG_SYNCLINK_CS_HDLC=y

Are you sure it's not a corrupted build tree or something like that? Can I have a look at the complete .config? Or some other bugzilla #?

There is exactly zero magic in generic HDLC, things are just build and exported as any other symbol in the tree.

Signature · 1 line
Krzysztof Halasa
#56 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 3 lines · Krzysztof Halasa

BTW: selecting CONFIG_HDLC without selecting at least one protocol does nothing good (except that the thing actually builds). As I said, I'd rather let the user decide what it's needed.

Agreed, but this is not about sensible kernel configuration. The synclink drivers have been building and working with generic HDLC for some time with no problem.

What we are cleaning up here is errors from building randomly (and not necessarily usable) generated configs.

The only thing I'm unsure of at this point is the
changes to net/wan/Makefile. I'm setting up some
test builds to check this out.

--
Paul

#57 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Krzysztof Halasa
Quoted 27 lines · On Wed, 07 Jun 2006 00:44:09 +0200 Krzysztof Halasa wrote:
Quoted 4 lines · Randy.Dunlap

OK, it's still not working. As Dave Jones reported:

Quoted 2 lines

WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/
synclink_cs.ko needs unknown symbol alloc_hdlcdev

Looks like the only patch in mm3 touching it is:
fix-kbuild-dependencies-for-synclink-drivers.

BTW: selecting CONFIG_HDLC without selecting at least one protocol does nothing good (except that the thing actually builds). As I said, I'd rather let the user decide what it's needed.

Hmm, with 2.6.17-rc5-mm3 it builds fine here (~ up-to-date FC5, i386):

$ egrep '^C.*(SYNCLINK|HDLC)' .config
CONFIG_HDLC=m
CONFIG_SYNCLINK=m
CONFIG_SYNCLINK_HDLC=y
CONFIG_SYNCLINKMP=m
CONFIG_SYNCLINKMP_HDLC=y
CONFIG_SYNCLINK_GT=m
CONFIG_SYNCLINK_GT_HDLC=y
CONFIG_SYNCLINK_CS=m
CONFIG_SYNCLINK_CS_HDLC=y
Quoted 2 lines

Are you sure it's not a corrupted build tree or something like that? Can I have a look at the complete .config? Or some other bugzilla #?

I did a make mrproper and got the same results.
I'm on x86-64 if it matters.
My .config is attached.

Quoted 2 lines

There is exactly zero magic in generic HDLC, things are just build and exported as any other symbol in the tree.

Thanks for looking into this.
---
~Randy

#58 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 7 lines · Randy.Dunlap
Quoted 2 lines · On Wed, 07 Jun 2006 00:44:09 +0200 Krzysztof Halasa wrote:

Are you sure it's not a corrupted build tree or something like that? Can I have a look at the complete .config? Or some other bugzilla #?

I did a make mrproper and got the same results.
I'm on x86-64 if it matters.
My .config is attached.

I just completed some builds including the setup you indicated (though not your exact .config, which I will try next). I'm also building x86-64.

Everything builds correctly using your patch, but without the Makefile portion, so we can drop that.

It builds for

CONFIG_HDLC=m
CONFIG_SYNCLINK=m
CONFIG_SYNCLINK_HDLC=y
(all modules)

and

CONFIG_HDLC=m
CONFIG_SYNCLINK=y
CONFIG_SYNCLINK_HDLC=y
(all built-in, hdlc promoted to 'y')

--
Paul

#59 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Randy.Dunlap
Quoted 2 lines · Randy.Dunlap

I'm on x86-64 if it matters.
My .config is attached.

Ok, reproduced.

The problem is that CONFIG_WAN is not set, the make system doesn't read drivers/net/wan/Makefile at all, and nothing in drivers/net/wan is being built.

Just another argument against random SELECTs.

Signature · 1 line
Krzysztof Halasa
#60 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 10 lines · On Wed, 07 Jun 2006 01:37:11 +0200 Krzysztof Halasa wrote:
Quoted 2 lines · Randy.Dunlap

I'm on x86-64 if it matters.
My .config is attached.

Ok, reproduced.

The problem is that CONFIG_WAN is not set, the make system doesn't read drivers/net/wan/Makefile at all, and nothing in drivers/net/wan is being built.

Aha. I commented about that early on in this thread and sent a patch to SELECT WAN. However:

Quoted 1 line

Just another argument against random SELECTs.

I agree with that and think that SYNCLINK should be using "depends" instead of "select".

Paul, can you repost the current patch, please?

---
~Randy

#61 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Krzysztof Halasa
Quoted 5 lines · Krzysztof Halasa
Quoted 2 lines · Randy.Dunlap

I'm on x86-64 if it matters.
My .config is attached.

Quoted 7 lines

Ok, reproduced.

The problem is that CONFIG_WAN is not set, the make system doesn't read drivers/net/wan/Makefile at all, and nothing in drivers/net/wan is being built.

Just another argument against random SELECTs.

OK, I thought he was building with the latest patch (attached here), which adds the 'select WAN' reverse dependency.
I tested his .config with the patch (minus the Makefile portion) and it builds just fine.

There is nothing random about these select statements. They are chosen specifically to fix the dependencies. You may feel they are ugly, but 'select' is the only tool I know of that fixes these errors without losing flexibility.

drivers/char/Kconfig +3 -0
@@ -197,6 +197,7 @@ config ISI
 config SYNCLINK
 	tristate "SyncLink PCI/ISA support"
 	depends on SERIAL_NONSTANDARD && PCI && ISA_DMA_API
+	select WAN if SYNCLINK_HDLC
 	select HDLC if SYNCLINK_HDLC
 	help
 	  Driver for SyncLink ISA and PCI synchronous serial adapters.
@@ -214,6 +215,7 @@ config SYNCLINK_HDLC
 config SYNCLINKMP
 	tristate "SyncLink Multiport support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINKMP_HDLC
 	select HDLC if SYNCLINKMP_HDLC
 	help
 	  Driver for SyncLink Multiport (2 or 4 ports) PCI synchronous serial adapter.
@@ -231,6 +233,7 @@ config SYNCLINKMP_HDLC
 config SYNCLINK_GT
 	tristate "SyncLink GT/AC support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINK_GT_HDLC
 	select HDLC if SYNCLINK_GT_HDLC
 	help
 	  Support for SyncLink GT and SyncLink AC families of
drivers/char/pcmcia/Kconfig +1 -0
@@ -8,6 +8,7 @@ menu "PCMCIA character devices"
 config SYNCLINK_CS
 	tristate "SyncLink PC Card support"
 	depends on PCMCIA
+	select WAN if SYNCLINK_CS_HDLC
 	select HDLC if SYNCLINK_CS_HDLC
 	help
 	  Driver for SyncLink PC Card synchronous serial adapter.

-
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/

#62 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 12 lines · On Tue, 06 Jun 2006 18:56:51 -0500 Paul Fulghum wrote:
Quoted 11 lines · Krzysztof Halasa
Quoted 2 lines · Randy.Dunlap

I'm on x86-64 if it matters.
My .config is attached.

Ok, reproduced.

The problem is that CONFIG_WAN is not set, the make system doesn't read drivers/net/wan/Makefile at all, and nothing in drivers/net/wan is being built.

Quoted 6 lines
Quoted 1 line

Just another argument against random SELECTs.

OK, I thought he was building with the latest patch (attached here), which adds the 'select WAN' reverse dependency.
I tested his .config with the patch (minus the Makefile portion) and it builds just fine.

No, I wasn't using any patches...

Quoted 4 lines

There is nothing random about these select statements. They are chosen specifically to fix the dependencies. You may feel they are ugly, but 'select' is the only tool I know of that fixes these errors without losing flexibility.

They are random in the sense that HDLC depends on WAN but only HDLC was being selected. In theory I would have expected config (software) to automatically enable higher-level config symbols in this case (select HDLC to cause select WAN), but that doesn't happen, so we got some "random" config which isn't supported (or even valid) ("random" being "invalid" in this case).

Quoted 29 lines
drivers/char/Kconfig +3 -0
@@ -197,6 +197,7 @@ config ISI
 config SYNCLINK
 	tristate "SyncLink PCI/ISA support"
 	depends on SERIAL_NONSTANDARD && PCI && ISA_DMA_API
+	select WAN if SYNCLINK_HDLC
 	select HDLC if SYNCLINK_HDLC
 	help
 	  Driver for SyncLink ISA and PCI synchronous serial adapters.
@@ -214,6 +215,7 @@ config SYNCLINK_HDLC
 config SYNCLINKMP
 	tristate "SyncLink Multiport support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINKMP_HDLC
 	select HDLC if SYNCLINKMP_HDLC
 	help
 	  Driver for SyncLink Multiport (2 or 4 ports) PCI synchronous serial adapter.
@@ -231,6 +233,7 @@ config SYNCLINKMP_HDLC
 config SYNCLINK_GT
 	tristate "SyncLink GT/AC support"
 	depends on SERIAL_NONSTANDARD && PCI
+	select WAN if SYNCLINK_GT_HDLC
 	select HDLC if SYNCLINK_GT_HDLC
 	help
 	  Support for SyncLink GT and SyncLink AC families of

drivers/char/pcmcia/Kconfig menu "PCMCIA character devices" · line 14

Quoted 7 lines
drivers/char/pcmcia/Kconfig +1 -0
@@ -8,6 +8,7 @@ menu "PCMCIA character devices"
 config SYNCLINK_CS
 	tristate "SyncLink PC Card support"
 	depends on PCMCIA
+	select WAN if SYNCLINK_CS_HDLC
 	select HDLC if SYNCLINK_CS_HDLC
 	help
 	  Driver for SyncLink PC Card synchronous serial adapter.

---
~Randy

#63 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 5 lines · Randy.Dunlap

They are random in the sense that HDLC depends on WAN but only HDLC was being selected. In theory I would have expected config (software) to automatically enable higher-level config symbols in this case (select HDLC to cause select WAN), but that doesn't happen

I absolutely agree, that is the way I thought it would work as all the information to build correctly is contained in the Kconfig files.

But, as you say, kbuild does not work that way. Changing kbuild to do that would take more time than I can commit, so I don't have any room to complain about it.

Quoted 3 lines

, so we got some "random" config
which isn't supported (or even valid) ("random" being "invalid" in this case).

Yes, the config is random, but the select statements are specifically chosen to work with the existing kbuild. Bottom line is the existing kbuild does not seem to propagate reverse dependencies, so you have to explicitly add them all with the select facility. Ugly, but not random.

In the end, it is your original patch (select WAN minus the Makefile changes) that makes it work.

BTW: Thanks for spending your time on this (Randy and Krzysztof), the exchange has been educational and useful.

--
Paul

#64 lore raw

Remove patch from 2.6.17-rc5-mm3

Andrew:

The patch fix-kbuild-dependencies-for-synclink-drivers.patch has proven to be wildly unpopular, so please drop it.

At some point in the 2.6.18 series I will look at an alternative fix.

Thanks,
Paul

#65 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Paul Fulghum
Quoted 2 lines · Paul Fulghum

There is nothing random about these select statements. They are chosen specifically to fix the dependencies.

And they may work until someone changes something - like a CONFIG_* name or directory structure.

Quoted 2 lines

You may feel they are ugly, but 'select' is the only tool I know of that fixes these errors without losing flexibility.

I don't know what special flexibility are you thinking about. DEPENDing on a symbol is as flexible and causes no such problems. The user just needs to enable (in this case) WAN and HDLC manually ((s)he has to select HDLC_* manually anyway). It could even make SYNCLINK_*_HDLC unneeded, the added functionality could (optionally) be selected automagically.

SELECTs are maybe good for small things but such inter-directory ones are IMHO asking for trouble.

Signature · 1 line
Krzysztof Halasa
#66 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 6 lines · Krzysztof Halasa

I don't know what special flexibility are you thinking about. DEPENDing on a symbol is as flexible and causes no such problems. The user just needs to enable (in this case) WAN and HDLC manually ((s)he has to select HDLC_* manually anyway). It could even make SYNCLINK_*_HDLC unneeded, the added functionality could (optionally) be selected automagically.

That is close to the way it was working, but it
had build errors with synclink=y and hdlc=m which
is what started all of these patches.

I guess I could add an extra check for that broken config case in all the individual source files using conditional compilation and disable synclink HDLC support altogether if such
a broken selection of options occurs (through
a random config generation). Still messy, but
since there seems to be a consensus that using the 'select' facility of kbuild is forbidden, that may be the more politically palatable way to go.

--
Paul

#67 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Krzysztof Halasa
Quoted 7 lines · On Wed, 07 Jun 2006 00:44:09 +0200 Krzysztof Halasa wrote:
Quoted 4 lines · Randy.Dunlap

OK, it's still not working. As Dave Jones reported:

Quoted 2 lines

WARNING: /lib/modules/2.6.17-rc5-mm3/kernel/drivers/char/pcmcia/
synclink_cs.ko needs unknown symbol alloc_hdlcdev

Quoted 6 lines

Looks like the only patch in mm3 touching it is:
fix-kbuild-dependencies-for-synclink-drivers.

BTW: selecting CONFIG_HDLC without selecting at least one protocol does nothing good (except that the thing actually builds). As I said, I'd rather let the user decide what it's needed.

That's the problem I think. SYNCLINK wants access to code in hdlc_generic when CONFIG_HDLC=m and SYNCLINK*=m but hdlc_generic.o and hdlc.ko are not being built.

Quoted 1 line

Hmm, with 2.6.17-rc5-mm3 it builds fine here (~ up-to-date FC5, i386):

---
~Randy

#68 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

In reply to Randy.Dunlap

Hi,

Quoted 3 lines · Randy.Dunlap

I think that the main problem is that SYNCLINK wants to be able to use some functions from hdlc_generic.c when CONFIG_HDLC=m. How do you handle that?

I don't :-)

If CONFIG_HDLC=m then all hw drivers below can only be modules.

Signature · 1 line
Krzysztof Halasa
#69 lore raw

Re: [PATCH] fix missing hdlc symbols for synclink drivers

Quoted 11 lines · On Tue, 06 Jun 2006 23:11:27 +0200 Krzysztof Halasa wrote:

Hi,

Quoted 3 lines · Randy.Dunlap

I think that the main problem is that SYNCLINK wants to be able to use some functions from hdlc_generic.c when CONFIG_HDLC=m. How do you handle that?

I don't :-)

If CONFIG_HDLC=m then all hw drivers below can only be modules.

OK. We have:

with partial .config:
# CONFIG_WAN is not set
CONFIG_HDLC=m
# CONFIG_HDLC_RAW is not set
# CONFIG_HDLC_RAW_ETH is not set
# CONFIG_HDLC_CISCO is not set
# CONFIG_HDLC_FR is not set
# CONFIG_HDLC_PPP is not set
...
CONFIG_SYNCLINK=m
CONFIG_SYNCLINK_HDLC=y
CONFIG_SYNCLINKMP=m
CONFIG_SYNCLINKMP_HDLC=y
CONFIG_SYNCLINK_GT=m
CONFIG_SYNCLINK_GT_HDLC=y
CONFIG_N_HDLC=m

and still get missing/unresolved symbols...

---
~Randy

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

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc5/2.6.17-rc5-mm3/

- Lots of PCI and USB updates

- The various lock validator, stack backtracing and IRQ management problems are converging, but we're not quite there yet.

One more, could not find it already reported (if yes, sorry for the noise). It is not in lockdep-combo as 20060606.

ide-floppy driver 0.99.newide
stopped custom tracer.

BUG: warning at kernel/lockdep.c:1856/trace_hardirqs_on()
 [<c01034ba>] show_trace+0x12/0x14
 [<c0103b8d>] dump_stack+0x19/0x1b
 [<c0133c56>] trace_hardirqs_on+0x14d/0x152
 [<f88bcaa9>] idefloppy_pc_intr+0x192/0x6ca [ide_floppy]
 [<f89402e2>] ide_intr+0x74/0x1c7 [ide_core]
 [<c013d212>] handle_IRQ_event+0x2e/0x63
 [<c013e451>] handle_edge_irq+0xad/0x132
 [<c0104dc7>] do_IRQ+0x6c/0xa5
 =======================
 [<c0102ec1>] common_interrupt+0x25/0x2c
 [<c0103029>] error_code+0x39/0x40

hdb: No disk in drive
hdb: 244736kB, 239/64/32 CHS, 4096 kBps, 512 sector size, 2941 rpm
hdb: No disk in drive

--
J.A. Magallon <jamagallon()ono!com> \ Software is like sex:
\ It's better when it's free
Mandriva Linux release 2007.0 (Cooker) for i586
Linux 2.6.16-jam19 (gcc 4.1.1 20060518 (prerelease)) #1 SMP PREEMPT Tue

#71 lore raw

PATCH 1/1 lock validator: add local_irq_enable_in_hardirq() to ide-floppy.c

Quoted 2 lines · J.A. Magallón

One more, could not find it already reported (if yes, sorry for the noise). It is not in lockdep-combo as 20060606.

indeed i missed that. The patch below should fix it.

------------------
Subject: lock validator: add local_irq_enable_in_hardirq() to ide-floppy.c
From: Ingo Molnar <[email protected]>

ide-floppy.c enables hardirqs in IRQ context. (build-tested)

---
 drivers/ide/ide-floppy.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/ide/ide-floppy.c +1 -1
Index: linux/drivers/ide/ide-floppy.c
===================================================================
@@ -839,7 +839,7 @@ static ide_startstop_t idefloppy_pc_intr
 			"transferred\n", pc->actually_transferred);
 		clear_bit(PC_DMA_IN_PROGRESS, &pc->flags);
 
-		local_irq_enable();
+		local_irq_enable_in_hardirq();
 
 		if (status.b.check || test_bit(PC_DMA_ERROR, &pc->flags)) {
 			/* Error detected */