Skip to contentSkip to conversation

2.6.17-rc4-mm1

PATCH In review Andrew Morton 27 participants 101 messages May 15, 2006 – May 24, 2006 1 message not in the archive

Subsystems (16)
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-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

- klibc (Kernel libc), as git-klibc.patch (Peter Anvin)

- Header cleanups for compatibility-with-userspace feasibility and for the addition of the `make hdrinstall' target, as git-hdrcleanup.patch and git-hdrinstall.patch (David Woodhouse)

- The cachefs patches are back (local-disk-based caching of network filesystem files) (David Howells)

- Added the ecryptfs filesystem

- per-task delay accounting patches: more advanced process accounting, plus an extensible interface for extended accounting.

- Many memory-management updates

- Plenty of other things

So if problems are found, please put extra effort into Cc:ing the correct mailing list and/or developer on any reports, thanks.

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-rc3-mm1:

origin.patch
git-acpi.patch
git-agpgart.patch
git-alsa.patch
git-block.patch
git-cfq.patch
git-cifs.patch
git-dvb.patch
git-gfs2.patch
git-ia64.patch
git-ieee1394.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-nfs.patch
git-ocfs2.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-viro-bird-m32r.patch
git-viro-bird-m68k.patch
git-viro-bird-frv.patch
git-viro-bird-upf.patch
git-viro-bird-volatile.patch
git-klibc-build-hacks.patch

git trees

-s390-make-qeth-buildable.patch -md-avoid-oops-when-attempting-to-fix-read-errors-on-raid10.patch -md-fixed-refcounting-locking-when-attempting-read-error-correction-in-raid10.patch -md-change-enotsupp-to-eopnotsupp.patch -md-improve-detection-of-lack-of-barrier-support-in-raid1.patch -md-fix-rdev-nr_pending-count-when-retrying-barrier-requests.patch -x86_64-add-compat_sys_vmsplice-and-use-it-in.patch -i386-x86-64-fix-acpi-disabled-lapic-handling.patch -i386-fix-overflow-in-e820_all_mapped.patch -i386-remove-apic=-warning.patch
-silence-initcall-warnings.patch
-uml-fix-iomem-list-traversal.patch -uml-skas0-support-for-2g-2g-hosts.patch -uml-remove-null-checks-and-add-some-codingstyle.patch -uml-clean-up-after-madvise_remove.patch
-uml-update-defconfig.patch
-uml-error-handling-fixes.patch -page-migration-fix-fallback-behavior-for-dirty-pages.patch -altix-correct-ioc3-port-order.patch -sparsemem-interaction-with-memory-add-bug-fixes.patch -spufs-fix-for-config_numa.patch -powerpc-allow-devices-to-register-with-numa-topology.patch -powerpc-cell-add-numa-id-to-struct-spu.patch -s390-fix-ipd-handling.patch
-s390-bug-in-setup_rt_frame.patch -rtc-rtc-dev-tweak-for-64-bit-kernel.patch -genrtc-fix-read-on-64-bit-platforms.patch -amd-alchemy-uart-claim-memory-range.patch -make-pc-speaker-driver-work-on-x86-64.patch -timer-tsc-check-suspend-notifier-change.patch -acpi-memory-leakages-in-drivers-acpi-thermalc.patch -audit-deal-with-deadlocks-in-audit_free.patch -audit-sockaddr-patch.patch -audit-move-call-of-audit_free-into-do_exit.patch -audit-drop-gfp_mask-in-audit_log_exit.patch -audit-drop-task-argument-of-audit_syscall_entryexit.patch -audit-no-need-to-wank-with-task_lock-and-pinning-task-down-in-audit_syscall_exit.patch -audit-support-for-context-based-audit-filtering.patch -audit-support-for-context-based-audit-filtering-2.patch -audit-audit-inode-patch.patch
-audit-change-lspp-ipc-auditing.patch -audit-reworked-patch-for-labels-on-user-space-messages.patch -audit-more-user-space-subject-labels.patch -audit-rework-of-ipc-auditing.patch
-audit-audit-filter-performance.patch -gregkh-driver-class-device-add-attribute_group-creation.patch -gregkh-driver-netdev-create-attribute_groups-with-class_device_add.patch -netdev-hotplug-napi-race-cleanup.patch -dvb-core-ule-fixes-and-rfc4326-additions-kernel-2616.patch -pwc-dec23-oops-fix.patch
-scx200_acb-fix-for-cs5535-eratta.patch
-sem2mutex-drivers-ieee1394.patch -drivers-ieee1394-ohci1394c-function-calls-without-effect.patch -forcedeth-fix-multi-irq-issues.patch -via-rhine-zero-pad-short-packets-on-rhine-i-ethernet-cards.patch -mv643xx_eth-provide-sysfs-class-device-symlink.patch -powerpc-pseries-avoid-crash-in-pci-code-if-mem-system-not-up.patch -git-rbtree-uml-fix.patch
-serial-locking-cleanup.patch
-ieee80211_wxc-remove-dead-code.patch -x86_64-mm-avoid-irq0-ioapic-pin-collision.patch -x86_64-mm-fix-die_lock-nesting.patch -x86_64-mm-add-nmi_exit-to-die_nmi.patch -x86_64-mm-compat-printk-fix.patch
-x86_64-mm-new-northbridge-fix.patch
-uml-fix-patch-mismerge.patch -uml-search-from-uml_net-in-a-more-reasonable-path.patch -uml-use-kbuild-tracking-for-all-files-and-fix-compilation-output.patch -uml-fix-compilation-and-execution-with-hardened-gcc.patch -uml-cleanup-unprofile-expression-and-build-infrastructure.patch -uml-export-symbols-added-by-gcc-hardened.patch -reduce-nr-of-ptr-derefs-in-fs-jffs2-summaryc.patch -remove-fs-jffs2-histoh.patch -cleanup-default-value-of-mtd_pcmcia_anonymous.patch -rcu-introduce-rcu_needs_cpu-interface.patch -rcu-introduce-rcu_needs_cpu-interface-fix.patch -s390-exploit-rcu_needs_cpu-interface.patch -handle-config_lbd-and-config_lsf-in-one-place.patch

Merged into mainline or a subsystem tree

+selinux-check-for-failed-kmalloc-in-security_sid_to_context.patch +fs-openc-unexport-sys_openat.patch +autofs4-nfy_none-wait-race-fix.patch +autofs4-nfy_none-wait-race-fix-tidy.patch +fix-capi-reload-by-unregistering-the-correct-major.patch +tpm-update-module-dependencies.patch +pcmcia-oopses-fixes.patch +via-quirk-fixup-additional-pci-ids.patch +smbfs-chroot-issue-cve-2006-1864.patch +rcu-introduce-rcu_needs_cpu-interface.patch +rcu-introduce-rcu_needs_cpu-interface-fix.patch +s390-exploit-rcu_needs_cpu-interface.patch +setup_per_zone_pages_min-overflow-fix.patch +s390-lcs-incorrect-test.patch +initramfs-fix-cpio-hardlink-check.patch +s390-add-vmsplice-system-call.patch +symbol_put_addr-locks-kernel.patch +contact-info-update.patch +smbfs-fix-slab-corruption-in-samba-error-path.patch +add-slab_is_available-routine-for-boot-code.patch +led-improve-kconfig-information.patch +backlight-lcd-class-fix-sysfs-_store-error-handling.patch +led-add-maintainer-entry-for-the-led-subsystem.patch +led-fix-sysfs-store-function-error-handling.patch +v9fs-twalk-memory-leak.patch +v9fs-signal-handling-fixes.patch +fix-can_share_swap_page-when-config_swap.patch +add-core-solo-and-core-duo-support-to-oprofile.patch +tpm-fix-constant.patch
+final-rio-polish.patch
+final-rio-polish-fix.patch +tpm_register_hardware-gcc-41-warning-fix.patch +fs-compatc-fix-if-a-=-b-typo.patch +root-mount-failure-emit-filesystems-attempted.patch +revert-vfs-propagate-mnt_flags-into-do_loopback-vfsmount.patch +smbus-unhiding-kills-thermal-management.patch +fix-hotplug-kconfig-help.patch +zone-init-check-and-report-unaligned-zone-boundaries.patch +x86-align-highmem-zone-boundaries-with-numa.patch +zone-allow-unaligned-zone-boundaries.patch +gigaset-endian-fix.patch +fix-typos-in-documentation-memory-barrierstxt.patch +ide_cs-add-ibm-microdrive-to-known-ids.patch +nfs-fix-error-handling-on-access_ok-in-compat_sys_nfsservctl.patch +devices-txt-remove-pktcdvd-entry.patch +jffs2-warning-fixes.patch
+dl2k-build-fix.patch

2.6.17 queue

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

Updates and fixes against acpi-dock-driver.patch

+asus_acpi-w3000-support.patch

Asus ACPI driver update

+uninorth-agp-warning-fixes.patch
+alpha-agp-warning-fix.patch

AGP fixes

+blk_start_queue-must-be-called-with-irq-disabled-add-warning.patch

block layer debug check

+cifs-do-not-overwrite-aops-elements.patch

CIFS was being bad.

+dprintk-adjustments-to-cpufreq-nforce2.patch +dprintk-adjustments-to-cpufreq-speedstep-centrino.patch +cpufreq-dprintk-adjustments.patch

cpufreq driover cleanups and fixes

+gregkh-driver-driver-core-class_device_add-needs-error-checks.patch +gregkh-driver-driver-core-config_debug_pm-covers-drivers-base-power-too.patch +gregkh-driver-platform_bus-learns-about-modalias.patch +gregkh-driver-driver-core-remove-exports.patch +gregkh-driver-kevent-add-new-uevent.patch +gregkh-driver-driver-core-allow-sysdev_class-have-attributes.patch +gregkh-driver-driver-core-fix-platform_device_add-to-use-device_add.patch +gregkh-driver-driver-core-add-sys-hypervisor-when-needed.patch +gregkh-driver-kobject-warn.patch +gregkh-driver-warn-when-statically-allocated-kobjects-are-used.patch

Driver tree updates

+gregkh-driver-platform_bus-learns-about-modalias-warning-fix.patch +gregkh-driver-warn-when-statically-allocated-kobjects-are-used-fix.patch

Fix them.

+drivers-base-firmware_classc-cleanups.patch
+s3c24xx-gpio-based-spi-driver.patch
+s3c24xx-hardware-spi-driver.patch
+s3c24xx-hardware-spi-driver-tidy.patch

Other driver tree things.

+scx200_acb-use-pci-i-o-resource-when-appropriate-fix.patch

Fix scx200_acb-use-pci-i-o-resource-when-appropriate.patch

+w1-warning-fix.patch

Fix warning

-disable-atykb-warning.patch
+remove-silly-messages-from-input-layer.patch

New version

+via-pmu-add-input-device.patch
+via-pmu-add-input-device-tidy.patch
+fix-mem-leak-in-sidewinder-driver.patch

Input things.

+add-dependency-on-kernelrelease-to-the-package-targets.patch

kbuild feature

+kbuild-export-type-enhancement-to-modpostc.patch +kbuild-prevent-building-modules-that-wont-load.patch +kbuild-export-symbol-usage-report-generator.patch

More kbuild work

+git-klibc-alpha-fixes.patch
+git-klibc-ident-fix.patch
+git-klibc-build-hacks.patch

+git-hdrcleanup-fixup.patch
+git-hdrcleanup-vs-git-klibc-on-ia64.patch
+git-hdrcleanup-vs-git-klibc-on-ia64-2.patch

+git-hdrinstall-fixup.patch

Various fixes against the new git trees

+sdhci-truncated-pointer-fix.patch

MMC fix

+nand_base-modular-fix.patch
+git-mtd-non-arm-fix.patch
+git-mtd-isnt-arm-only.patch

MTD fixes

+drivers-char-hw_randomc-remove-asserts.patch

hw_random cleanup

+fix-phy-id-for-lxt971a-lxt972a.patch

net driver fix

+clean-up-initcall-warning-for-netconsole.patch +remove-dead-entry-in-net-wan-kconfig.patch +ppp_async-hang-fix.patch +selinux-add-security-class-for-appletalk-sockets.patch +fix-mem-leak-in-netfilter.patch +neighbourc-pneigh_get_next-skips-published-entry.patch

networking

+nfs-permit-filesystem-to-override-root-dentry-on-mount.patch +nfs-permit-filesystem-to-perform-statfs-with-a-known-root-dentry.patch +ipath-vs-nfs-permit-filesystem-to-override-root-dentry-on-mount.patch +git-gfs2-vs-nfs-permit-filesystem-to-override-root-dentry-on-mount.patch +nfs-abstract-out-namespace-initialisation.patch +nfs-add-dentry-materialisation-op.patch +nfs-split-fs-nfs-inodec-into-inode-superblock-and-namespace-bits.patch +nfs-share-nfs-superblocks-per-protocol-per-server-per-fsid.patch +nfs-share-nfs-superblocks-per-protocol-per-server-per-fsid-fix.patch +nfs-share-nfs-superblocks-per-protocol-per-server-per-fsid-warning-fixes.patch +fs-cache-provide-a-filesystem-specific-syncable-page-bit.patch +fs-cache-add-notification-of-page-becoming-writable-to-vma-ops.patch +fs-cache-avoid-enfile-checking-for-kernel-specific-open-files.patch +fs-cache-generic-filesystem-caching-facility.patch +fs-cache-make-kafs-use-fs-cache.patch +fs-cache-cachefiles-a-cache-that-backs-onto-a-mounted-filesystem.patch +fs-cache-cachefiles-a-cache-that-backs-onto-a-mounted-filesystem-export-__audit_inode_child.patch +fs-cache-cachefiles-a-cache-that-backs-onto-a-mounted-filesystem-warning-fixes.patch +fs-cache-vs-vfs-add-lock-owner-argument-to-flush-operation.patch +fs-cache-release-page-private-in-failed-readahead.patch +fs-cache-release-page-private-in-failed-readahead-uninlining.patch +fs-cache-release-page-private-in-failed-readahead-uninlining-2.patch +nfs-use-local-caching.patch
+nfs-use-local-caching-fix.patch

cachefs

+powerpc-pseries-increment-fail-counter-in-pci-recovery.patch +powerpc-kbuild-warning-fix.patch

powerpc fixes

+fix-pciehp-compile-issue-when-config_acpi-is-not.patch +i386-export-memory-more-than-4g-through-proc-iomem.patch +fix-recovery-path-from-errors-during-pcie_init.patch +move-various-pci-ids-to-header-file.patch

PCI fixes

-pci-add-pci_assign_resource_fixed-allow-fixed-address-fix.patch

Folded into pci-add-pci_assign_resource_fixed-allow-fixed-address.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

Make the newly-64-bit resource.start and resource.end Kconfigurably 32-bit.

+ti-pcixx12-cardbus-controller-support.patch

Cardbus controller device support

+drivers-scsi-use-array_size-macro.patch +lpfc-sparse-null-warnings.patch +mpt_interrupt-should-return-irq_none-when.patch +aic7-cleanup-module_parm_desc-strings.patch +qla2xxx-lock-ordering-fix.patch +qla2xxx-lock-ordering-fix-warning-fix.patch +random-remove-redundant-sa_sample_random-from-ninjascsi.patch +megaraid-gcc-41-warning-fix.patch +buslogic-gcc-41-warning-fixes.patch +add-scsi_add_host-failure-handling-for-nsp32.patch

scsi things

+gregkh-usb-usb-console-fix-disconnection-issues.patch +gregkh-usb-usb-clean-out-an-unnecessary-null-check-from-ub.patch +gregkh-usb-usbatm-remove-pointless-inline.patch
+gregkh-usb-usbatm-remove-no-longer-needed-include.patch +gregkh-usb-usb-phidget-interfacekit-make-inputs-pollable-and-new-device-support.patch +gregkh-usb-usb-shuttle_usbat-fix-handling-of-scatter-gather-buffers.patch +gregkh-usb-usb-shuttle_usbat-hardcode-detection-of-hp-cdrw-devices.patch +gregkh-usb-usb-shuttle_usbat-hardcode-flash-detection-for-now.patch +gregkh-usb-usb-usb-storage-alauda-fix-transport-info-mismerge.patch +gregkh-usb-usb-net2280-add-a-shutdown-routine.patch +gregkh-usb-usb-uhci-store-the-endpoint-type-in-the-qh-structure.patch +gregkh-usb-usb-uhci-fix-obscure-bug-in-enqueue.patch +gregkh-usb-usb-hid-hidbp-input-drivers-fix-various-usb-input-hid-input.c-bugs-that-make-apple-mighty-mouse-work-poorly.patch

USB tree updates

-fix-sco-on-some-bluetooth-adapters.patch
+fix-sco-on-some-bluetooth-adapters-2.patch

New version

+x86_64-mm-add-end-endproc-annotations-to-entrys.patch +x86_64-mm-i386-numa-summit-check.patch +x86_64-mm-fix-unlikely-profiling--vsyscalls-on-x86_64.patch +x86_64-mm-fix-b44-checks.patch
+x86_64-mm-nommu-warning.patch +x86_64-mm-nmi-watchdog-header-cleanup.patch +x86_64-mm-add-performance-counter-reservation-framework-for-up-kernels.patch +x86_64-mm-utilize-performance-counter-reservation-framework-in-oprofile.patch +x86_64-mm-add-smp-support-on-x86_64-to-reservation-framework.patch +x86_64-mm-add-smp-support-on-i386-to-reservation-framework.patch +x86_64-mm-cleanup-nmi-interrupt-path.patch +x86_64-mm-remove-un-set_nmi_callback-and-reserve-release_lapic_nmi-functions.patch +x86_64-mm-add-abilty-to-enable-disable-nmi-watchdog-from-sysfs.patch

x86_64 tree updates

+x86_64-mm-remove-un-set_nmi_callback-and-reserve-release_lapic_nmi-functions-x86-fix.patch +x86_64-mm-remove-un-set_nmi_callback-and-reserve-release_lapic_nmi-functions-x86-fix-fix.patch +x86_64-mm-remove-un-set_nmi_callback-and-reserve-release_lapic_nmi-functions-x86_64-fix.patch +x86_64-mm-serialize-assign_irq_vector-use-of-static-variables-fix.patch +add-abilty-to-enable-disable-nmi-watchdog-from-procfs.patch

Fix them.

-gregkh-devfs-devfs-die-die-die.patch -gregkh-devfs-devfs-remove-documentation.patch -gregkh-devfs-devfs-scrub-partitions.patch -gregkh-devfs-devfs-scrub-init.patch -gregkh-devfs-devfs-remove-serial-subsystem.patch -gregkh-devfs-devfs-remove-ide-subsystem.patch -gregkh-devfs-devfs-remove-sound-subsystem.patch -gregkh-devfs-devfs-remove-devfs-tape.patch -gregkh-devfs-devfs-remove-devfs_mk_dir.patch -gregkh-devfs-devfs-remove-devfs_mk_symlink.patch -gregkh-devfs-devfs-remove-devfs_mk_bdev.patch -gregkh-devfs-devfs-remove-devfs_mk_cdev.patch -gregkh-devfs-devfs-remove-devfs_remove.patch -gregkh-devfs-devfs-remove-devfs_fs_kernel.h.patch -gregkh-devfs-devfs-remove-misc-devfs_name.patch -gregkh-devfs-devfs-remove-genhd-devfs_name.patch -gregkh-devfs-devfs-remove-videodev-devfs_name.patch -gregkh-devfs-devfs-remove-line-devfs_name.patch -gregkh-devfs-devfs-remove-tty-devfs_name.patch -gregkh-devfs-devfs-tty_driver_no_devfs.patch -gregkh-devfs-devfs-minor-cleanups.patch -gregkh-devfs-devfs-feature-removal.patch

Dropped - too much overlap with the klibc work

+pgdat-allocation-for-new-node-add-specify-node-id-tidy-cleanup.patch +fix-compile-error-undefined-reference-for-sparc64.patch

Fix pgdat-allocation-for-new-node-add-specify-node-id.patch some more

+pgdat-allocation-and-update-for-ia64-of-memory-hotplughold-pgdat-address-at-system-running.patch +pgdat-allocation-and-update-for-ia64-of-memory-hotplug-update-pgdat-address-array.patch +pgdat-allocation-and-update-for-ia64-of-memory-hotplugallocate-pgdat-and-per-node-data.patch +page-migration-cleanup-extract-try_to_unmap-from-migration-functions-update-comments-7.patch +page-migration-cleanup-move-fallback-handling-into-special-function-update-comments-9.patch +swapless-pm-add-r-w-migration-entries-fix.patch +swapless-pm-add-r-w-migration-entries-ifdefs.patch +swapless-pm-add-r-w-migration-entries-update-comments.patch +swapless-pm-add-r-w-migration-entries-update-comments-4.patch +swapless-pm-add-r-w-migration-entries-update-comments-6.patch +swapless-page-migration-modify-core-logic-remove-useless-mapping-checks.patch +more-page-migration-use-migration-entries-for-file-pages-fix.patch +more-page-migration-use-migration-entries-for-file-pages-update-comments-5.patch +more-page-migration-use-migration-entries-for-file-pages-update-comments-8.patch +more-page-migration-use-migration-entries-for-file-pages-remove_migration_ptes.patch +more-page-migration-use-migration-entries-for-file-pages-replace-call-to-pageout-with-writepage-2.patch +page-migration-update-documentation.patch +aop_truncated_page-victims-in-read_pages-belong-in-the-lru.patch +introduce-mechanism-for-registering-active-regions-of-memory.patch +have-power-use-add_active_range-and-free_area_init_nodes.patch +have-x86-use-add_active_range-and-free_area_init_nodes.patch +have-x86_64-use-add_active_range-and-free_area_init_nodes.patch +#have-ia64-use-add_active_range-and-free_area_init_nodes.patch +tracking-dirty-pages-in-shared-mappings-v4.patch +tracking-dirty-pages-in-shared-mappings-v4-fix2.patch +tracking-dirty-pages-in-shared-mappings-v4-fix3.patch +throttle-writers-of-shared-mappings.patch +throttle-writers-of-shared-mappings-tidy.patch
+optimize-follow_pages.patch
+slab-verify-pointers-before-free.patch +sparsemem-record-nid-during-memory-present.patch

Memory management updates

+fix-x86-microcode-driver-handling-of-multiple-matching.patch +i386-break-out-of-recursion-in-stackframe-walk.patch +dont-trigger-full-rebuild-via-config_x86_mce.patch

x86 updates

-x86_64-disable-tsc-with-seccomp.patch

Dropped

+remove-duplicate-symbol-exports-on-alpha.patch

Cleanup

+swsusp-documentation-updates.patch
+swsusp-fix-typo-in-cr0-handling.patch

swsusp updates

-s390-hypervisor-file-system.patch
-s390-hypervisor-file-system-fixes.patch

Dropped (what I had was out of date and it got all confusing. Need a resend)

+add-raw-driver-kconfig-entry-for-s390.patch

Enable raw driver on s390

+percpu-counter-data-type-changes-to-suppport-fix-fix-fix.patch

Fix percpu-counter-data-type-changes-to-suppport.patch some more

+make-rcu-api-inaccessible-to-non-gpl-linux-kernel-modules.patch +doc-add-audit-acct-to-docbook.patch
+ip2-fix-sections.patch
+sgi-ioc4-detect-io-card-variant.patch +two-additions-to-linux-documentation-ioctl-numbertxt.patch +when-config_base_samll=1-the-kernel-261611-cascade-in-kernel-timerc-may-enter-the-infinite-loop.patch +codingstyle-add-typedefs-chapter.patch
+fs-bufferc-possible-cleanups.patch
+rtc-rtc-dev-uie-emulation.patch
+drivers-md-raid6algosc-fix-a-null-dereference.patch +adjust-handle_irr_event-return-type.patch
+sparse-fixes-for-synclink_cs.patch
+jbd-split-checkpoint-lists.patch
+jbd-split-checkpoint-lists-tidy.patch
+add-__iowrite64_copy.patch
+add-pci_cap_id_vndr.patch
+mark-address_space_operations-const.patch +mark-address_space_operations-const-fix.patch +mark-address_space_operations-const-fix-2.patch
+more-bug_on-conversion.patch
+make-kernel-ignore-bogus-partitions.patch +drivers-block-loopc-dont-return-garbage-if-loop_set_status-not-called.patch +docs-update-sparsetxt-with-check_endian.patch +drivers-acorn-char-pcf8583-vs-rtc-subsystem.patch +rewritten-backlight-infrastructure-for-portable-apple-computers.patch +ensure-null-deref-cant-possibly-happen-in-is_exported.patch +bluetooth-fix-potential-null-ptr-deref-in-dtl1_cscdtl1_hci_send_frame.patch +bloat-o-meter-gcc-4-fix.patch +random-remove-sa_sample_random-from-floppy-driver.patch +random-make-cciss-use-add_disk_randomness.patch +random-change-cpqarray-to-use-add_disk_randomness.patch +random-remove-bogus-sa_sample_random-from-at91-compact-flash-driver.patch +random-remove-redundant-sa_sample_random-from-touchscreen-drivers.patch +define-__raw_get_cpu_var-and-use-it.patch +allow-for-per-cpu-data-being-in-tdata-and-tbss-sections.patch +allow-for-per-cpu-data-being-in-tdata-and-tbss-sections-fix.patch +allow-for-per-cpu-data-being-in-tdata-and-tbss-sections-tidy.patch +deprecate-smbfs-in-favour-of-cifs.patch +allow-raw_notifier-callouts-to-unregister-themselves.patch +hptiop-highpoint-rocketraid-3xxx-controller-driver.patch +hptiop-highpoint-rocketraid-3xxx-controller-driver-list-locking.patch +hptiop-highpoint-rocketraid-3xxx-controller-driver-list-locking-updates.patch +hptiop-highpoint-rocketraid-3xxx-controller-driver-list-locking-updates-updates-2.patch +fix-kbuild-dependencies-for-synclink-drivers.patch +fs-freevxfs-cleanup-of-spelling-errors.patch
+pnp-card_probe-fix-memory-leak.patch +fix-unlikely-memory-leak-in-dac960-driver.patch +ufs-ufs_trunc_indirect-infinite-cycle.patch
+ufs-right-block-allocation.patch
+ufs-right-block-allocation-fixes.patch
+sunsu-license-fix.patch

Patches ;)

+per-task-delay-accounting-setup.patch
+per-task-delay-accounting-setup-fix-1.patch
+per-task-delay-accounting-setup-fix-2.patch +per-task-delay-accounting-sync-block-i-o-and-swapin-delay-collection.patch +per-task-delay-accounting-sync-block-i-o-and-swapin-delay-collection-fix-1.patch +per-task-delay-accounting-cpu-delay-collection-via-schedstats.patch +per-task-delay-accounting-cpu-delay-collection-via-schedstats-fix-1.patch +per-task-delay-accounting-utilities-for-genetlink-usage.patch +per-task-delay-accounting-taskstats-interface.patch +per-task-delay-accounting-taskstats-interface-fix-1.patch +per-task-delay-accounting-delay-accounting-usage-of-taskstats-interface.patch +per-task-delay-accounting-delay-accounting-usage-of-taskstats-interface-use-portable-cputime-api-in-__delayacct_add_tsk.patch +per-task-delay-accounting-documentation.patch +per-task-delay-accounting-proc-export-of-aggregated-block-i-o-delays.patch +per-task-delay-accounting-proc-export-of-aggregated-block-i-o-delays-warning-fix.patch

per-task delay accounting

+time-use-clocksource-abstraction-for-ntp-adjustments-optimize-out-some-mults-since-gcc-cant-avoid-them.patch +time-i386-clocksource-drivers-fix-spelling-typos.patch +generic-time-add-macro-to-simplify-hide-mask.patch

Updates to the x86 time management patches

-mmc-multi-sector-writes.patch

Dropped

+capi-crash--race-condition.patch

i4l semifix

+kconfig-select-things-at-the-closest-tristate-instead-of-bool.patch

nfsd Kconfig fix

+sched-comment-bitmap-size-accounting.patch
+unnecessary-long-index-i-in-sched.patch

CPU scheduler work

+mm-implement-swap-prefetching-fix.patch
+mm-implement-swap-prefetching-sched-batch.patch

swap-prefetch fixes

+pi-futex-patchset-v4-fix.patch
+document-futex-pi-design.patch
+document-futex-pi-design-fix.patch
+document-futex-pi-design-fix-fix.patch
+futex_requeue-optimization.patch

pi-futex updates

+de_thread-fix-lockless-do_each_thread.patch

task maangment fix

+coredump-kill-ptrace-related-stuff-fix.patch

Fix coredump-kill-ptrace-related-stuff.patch

+ecryptfs-fs-makefile-and-fs-kconfig.patch
+ecryptfs-documentation.patch
+ecryptfs-makefile.patch
+ecryptfs-main-module-functions.patch +ecryptfs-main-module-functions-vs-nfs-permit-filesystem-to-override-root-dentry-on-mount.patch +ecryptfs-header-declarations.patch
+ecryptfs-superblock-operations.patch
+ecryptfs-dentry-operations.patch
+ecryptfs-file-operations.patch
+ecryptfs-inode-operations.patch
+ecryptfs-mmap-operations.patch +mark-address_space_operations-const-vs-ecryptfs-mmap-operations.patch +ecryptfs-keystore.patch
+ecryptfs-crypto-functions.patch
+ecryptfs-debug-functions.patch
+ecryptfs-alpha-build-fix.patch

ecryptfs

+reiser4-add-missing-txn_restart-before-get_nonexclusive_access-calls.patch +reiser4-check-radix-tree-emptiness-properly.patch +reiser4-check-radix-tree-emptiness-properly-2.patch

reiser4 updates

+reiser4-vs-nfs-permit-filesystem-to-override-root-dentry-on-mount.patch +reiser4-vs-nfs-permit-filesystem-to-perform-statfs-with-a-known-root-dentry.patch

Fix reiser4 damage due to other patches

+ide-fix-hpt37x-timing-tables.patch +ide-optimize-hpt37x-timing-tables.patch +ide-fix-hpt3xx-hotswap-support.patch +ide-fix-the-case-of-multiple-hpt3xx-chips-present.patch +ide-hpt3xx-fix-pci-clock-detection.patch +ide-pdc202xx_old-remove-the-obsolete-busproc.patch +piix-fix-82371mx-enablebits.patch +piix-remove-check-for-broken-mw-dma-mode-0.patch +piix-slc90e66-pio-mode-fallback-fix.patch +make-number-of-ide-interfaces-configurable.patch

IDE driver fixes

+nvidiafb-add-support-for-geforce-6100-and-related-chipsets.patch +fbdev-add-1366x768-wxga-mode-to-mode-database.patch +intelfb-use-firmware-edid-for-mode-database.patch +vesafb-fix-return-code-of-vesafb_setcolreg.patch +vesafb-prefer-vga-registers-over-pmi.patch +vt-delay-the-update-of-the-visible-console.patch +atyfb-fix-dead-code.patch
+fbdev-coverity-bug-85.patch +fbdev-coverity-bug-90.patch

fbdev driver updates

+md-bitmap-fix-online-removal-of-file-backed-bitmaps.patch +md-bitmap-remove-bitmap-writeback-daemon.patch +md-bitmap-cleaner-separation-of-page-attribute-handlers-in-md-bitmap.patch +md-bitmap-use-set_bit-etc-for-bitmap-page-attributes.patch +md-bitmap-remove-unnecessary-page-reference-manipulations-from-md-bitmap-code.patch +md-bitmap-remove-dead-code-from-md-bitmap.patch +md-bitmap-tidy-up-i_writecount-handling-in-md-bitmap.patch +md-bitmap-change-md-bitmap-file-handling-to-use-bmap-to-file-blocks.patch

RAID updates

-nmi-lockup-and-altsysrq-p-dumping-calltraces-on-_all_-cpus.patch -x86_64-ipi-calltraces.patch

Dropped due to rejects

-exit-dont-call-sleeping-things-when-oopsing.patch

Dropped due to lack of interest

All 992 patches:

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/patch-list

#1 lore raw

Hi Andrew

It seems latest kernels have a problem in kmem_cache_destroy()

On a dual opteron machine (NUMA), its quite easy do trigger the bug : doing a oprofile session like :

opcontrol --setup --event=CPU_CLK_UNHALTED:100000 --vmlinux=/usr/src/linux-2.6.17-rc4-mm1/vmlinux
...
opcontrol --dump
...
opcontrol --deinit <<<-- Triggers the bug

slab error in kmem_cache_destroy(): cache `dcookie_cache': Can't free all objects

Call Trace: <ffffffff80278aa4>{kmem_cache_destroy+212}
      <ffffffff802bc559>{dcookie_unregister+345} 
<ffffffff803bbe7a>{event_buffer_release+26}
      <ffffffff8027cee2>{__fput+98} <ffffffff80279ecd>{filp_close+93}
      <ffffffff8023003e>{put_files_struct+110} 
<ffffffff8023143a>{do_exit+650}
      <ffffffff802f58f1>{__up_write+33} 
<ffffffff80231bb8>{do_group_exit+200}
      <ffffffff802097ce>{system_call+126}
# grep dcookie_cache /proc/slabinfo
dcookie_cache          0      0     32  101    1 : tunables  120   60    8 : slabdata      0      0      0

This problem is annoying, because in the oprofile case, we must reboot the machine to be able to start a new profile session.

Eric

#2 lore raw
Quoted 4 lines · Eric Dumazet

Hi Andrew

It seems latest kernels have a problem in kmem_cache_destroy()

Mainline, or just -mm?

Quoted 23 lines

On a dual opteron machine (NUMA), its quite easy do trigger the bug :
doing a oprofile session like :

opcontrol --setup --event=CPU_CLK_UNHALTED:100000
--vmlinux=/usr/src/linux-2.6.17-rc4-mm1/vmlinux
...
opcontrol --dump
...
opcontrol --deinit <<<-- Triggers the bug

slab error in kmem_cache_destroy(): cache `dcookie_cache': Can't free all objects

Call Trace: <ffffffff80278aa4>{kmem_cache_destroy+212}
       <ffffffff802bc559>{dcookie_unregister+345} 
<ffffffff803bbe7a>{event_buffer_release+26}
       <ffffffff8027cee2>{__fput+98} <ffffffff80279ecd>{filp_close+93}
       <ffffffff8023003e>{put_files_struct+110} 
<ffffffff8023143a>{do_exit+650}
       <ffffffff802f58f1>{__up_write+33} 
<ffffffff80231bb8>{do_group_exit+200}
       <ffffffff802097ce>{system_call+126}
Quoted 8 lines
# grep dcookie_cache /proc/slabinfo
dcookie_cache          0      0     32  101    1 : tunables  120   60    
8 : slabdata      0      0      0

This problem is annoying, because in the oprofile case, we must reboot the machine to be able to start a new profile session.

sony:/home/akpm# opcontrol --setup --event=CPU_CLK_UNHALTED:100000  --vmlinux=/boot/vmlinux-2.6.17-rc4-mm1
sony:/home/akpm# opcontrol --start
Using 2.6+ OProfile kernel interface.
Reading module info.
Using log file /var/lib/oprofile/oprofiled.log
Daemon started.
Profiler running.
sony:/home/akpm# opcontrol --dump 
sony:/home/akpm# opcontrol --stop
Stopping profiling.
sony:/home/akpm# opcontrol --deinit
Stopping profiling.
Killing daemon.
Unloading oprofile module
sony:/home/akpm# dmesg

<nothing>

Can you provide a more detailed method for reproducing this?

#3 lore raw

Hi Andrew,

Quoted 1 line · Eric Dumazet
Quoted 1 line

It seems latest kernels have a problem in kmem_cache_destroy()

Quoted 1 line · Andrew Morton

Mainline, or just -mm?

Could be in mainline. See the following thread:
http://lkml.org/lkml/2006/4/27/69. Can't reproduce locally so waiting for git bisect results from the original reporter...

Pekka

#4 lore raw
In reply to Andrew Morton
Quoted 10 lines · Andrew Morton
Quoted 4 lines · Eric Dumazet

Hi Andrew

It seems latest kernels have a problem in kmem_cache_destroy()

Mainline, or just -mm?

Mainline and mm it seems, only if NUMA is ON, on 2.6.17 only (2.6.16.XX is OK)

I added some printks and it seems slab_partials is not empty in __node_shrink() for node 0

I am not a mm/slab.c expert but the following patch cures the problem for me :

[PATCH] slab : NUMA may need __cache_shrink() doing two loops.

mm/slab.c +17 -10
@@ -2225,25 +2225,32 @@
 		spin_lock_irq(&l3->list_lock);
 	}
 	ret = !list_empty(&l3->slabs_full) || !list_empty(&l3->slabs_partial);
+	if (ret)
+		printk(KERN_ERR "__node_shrink(name=%s,node=%d) ret=%d\n",
+			cachep->name, node, ret);
 	return ret;
 }
 
 static int __cache_shrink(struct kmem_cache *cachep)
 {
-	int ret = 0, i = 0;
+	int ret, i;
 	struct kmem_list3 *l3;
+	int loopcount = 0;
 
-	drain_cpu_caches(cachep);
+	do {
+		ret = 0;
+		drain_cpu_caches(cachep);
 
-	check_irq_on();
-	for_each_online_node(i) {
-		l3 = cachep->nodelists[i];
-		if (l3) {
-			spin_lock_irq(&l3->list_lock);
-			ret += __node_shrink(cachep, i);
-			spin_unlock_irq(&l3->list_lock);
+		check_irq_on();
+		for_each_online_node(i) {
+			l3 = cachep->nodelists[i];
+			if (l3) {
+				spin_lock_irq(&l3->list_lock);
+				ret += __node_shrink(cachep, i);
+				spin_unlock_irq(&l3->list_lock);
+			}
 		}
-	}
+	} while (ret && ++loopcount < 2);
 	return (ret ? 1 : 0);
 }
 
#5 lore raw
In reply to Andrew Morton
Quoted 3 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

Indeed. This is the first -mm in a fair while that has crapped itself on boot on my box.

NeilB - is this yours?

usbcore: registered new driver libusual
usbcore: registered new driver hiddev
input: Belkin Components Belkin OmniView KVM Switch as /class/input/input0
input: USB HID v1.00 Keyboard [Belkin Components Belkin OmniView KVM Switch] on usb-0000:00:1d.0-1.1
input: Belkin Components Belkin OmniView KVM Switch as /class/input/input1
input: USB HID v1.00 Mouse [Belkin Components Belkin OmniView KVM Switch] on usb-0000:00:1d.0-1.1
usbcore: registered new driver usbhid
drivers/usb/input/hid-core.c: v2.6:USB HID core driver
serio: i8042 AUX port at 0x60,0x64 irq 12
serio: i8042 KBD port at 0x60,0x64 irq 1
mice: PS/2 mouse device common for all mice
md: raid1 personality registered for level 1
md: md driver 0.90.3 MAX_MD_DEVS=256, MD_SB_DISKS=27
md: bitmap version 4.39
EDAC MC: Ver: 2.0.0 May 15 2006
TCP bic registered
NET: Registered protocol family 1
NET: Registered protocol family 17
drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
Freeing unused kernel memory: 208k freed
Red Hat nash version 5.0.39 starting
Mounting proc filesystem
Mounting sysfs filesystem
Creating /dev
Creating initial device nodes
Setting up hotplug.
Creating block device nodes.
Loading ide-disk.ko module

md: Autodetecting RAID arrays.
md: autorun ...
md: considering sdc11 ...
md:  adding sdc11 ...
md: sdc10 has different UUID to sdc11
md: sdc7 has different UUID to sdc11
md: sdc6 has different UUID to sdc11
md: sdc5 has different UUID to sdc11
md: sdc3 has different UUID to sdc11
md: sdc2 has different UUID to sdc11
md:  adding sda11 ...
md: sda10 has different UUID to sdc11
md: sda7 has different UUID to sdc11
md: sda6 has different UUID to sdc11
md: sda5 has different UUID to sdc11
md: sda3 has different UUID to sdc11
md: sda2 has different UUID to sdc11
md: created md5
md: bind<sda11>
md: bind<sdc11>
md: running: <sdc11><sda11>
raid1: raid set md5 active with 0 out of 2 mirrors
md: considering sdc10 ...
md:  adding sdc10 ...
md: sdc7 has different UUID to sdc10
md: sdc6 has different UUID to sdc10
md: sdc5 has different UUID to sdc10
md: sdc3 has different UUID to sdc10
md: sdc2 has different UUID to sdc10
md:  adding sda10 ...
md: sda7 has different UUID to sdc10
md: sda6 has different UUID to sdc10
md: sda5 has different UUID to sdc10
md: sda3 has different UUID to sdc10
md: sda2 has different UUID to sdc10
md: created md6
md: bind<sda10>
md: bind<sdc10>
md: running: <sdc10><sda10>
raid1: raid set md6 active with 0 out of 2 mirrors
md: considering sdc7 ...
md:  adding sdc7 ...
md: sdc6 has different UUID to sdc7
md: sdc5 has different UUID to sdc7
md: sdc3 has different UUID to sdc7
md: sdc2 has different UUID to sdc7
md:  adding sda7 ...
md: sda6 has different UUID to sdc7
md: sda5 has different UUID to sdc7
md: sda3 has different UUID to sdc7
md: sda2 has different UUID to sdc7
md: created md4
md: bind<sda7>
md: bind<sdc7>
md: running: <sdc7><sda7>
raid1: raid set md4 active with 0 out of 2 mirrors
Unable to handle kernel NULL pointer dereference at 0000000000000010 RIP:
<ffffffff8040e828>{bitmap_create+152}
PGD 3ed90067 PUD 3ed91067 PMD 0
Oops: 0000 [1] SMP
last sysfs file: /block/fd0/removable
CPU 0
Modules linked in: ide_disk
Pid: 1, comm: init Not tainted 2.6.17-rc4-mm1 #1
RIP: 0010:[<ffffffff8040e828>] <ffffffff8040e828>{bitmap_create+152}
RSP: 0000:ffff81003f603ad8  EFLAGS: 00010246
RAX: 0000000000000008 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff810037d424c0
RBP: ffff810037d423c0 R08: 0000000000005d7a R09: 0000000000000000
R10: ffff810037d423c0 R11: 0000000000000100 R12: 00000000fffffff4
R13: ffff81003ed65000 R14: 00000000000f3180 R15: ffff81003ed650d0
FS:  000000000066e890(0063) GS:ffffffff8060e000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000010 CR3: 000000003ed89000 CR4: 00000000000006e0
Process init (pid: 1, threadinfo ffff81003f602000, task ffff81003f601530)
Stack: ffff810037d94fc0 ffffffff80406ee6 0000000100000001 000000013edb40c0
       ffff81003edb40c0 ffff81003ed65018 ffff810037d44f40 ffff810037d427c0
       ffff81003ed65000 ffffffff80404a16
Call Trace: <ffffffff80406ee6>{md_register_thread+167}
       <ffffffff80404a16>{run+942} <ffffffff804080c8>{do_md_run+1208}
       <ffffffff804067ae>{bind_rdev_to_array+494} 
<ffffffff8034d826>{kobject_register+51}
       <ffffffff804069c4>{mddev_find+122} <ffffffff8040865d>{autorun_devices+731}
       <ffffffff80224627>{alloc_inode+246} <ffffffff8040b445>{md_ioctl+165}
       <ffffffff802d641d>{sysfs_make_dirent+27} 
<ffffffff80259cac>{kobject_uevent+220}
       <ffffffff802d5cc4>{sysfs_add_file+118} 
<ffffffff8034690c>{blkdev_driver_ioctl+92}
       <ffffffff80346fd8>{blkdev_ioctl+1704} 
<ffffffff802b536f>{check_disk_change+31}
       <ffffffff8040c84b>{md_open+92} <ffffffff802b5df6>{blkdev_open+0}
       <ffffffff802b5df6>{blkdev_open+0} <ffffffff802b5adb>{do_open+651}
       <ffffffff802b583d>{bdget+277} <ffffffff802b5df6>{blkdev_open+0}
       <ffffffff802b5e17>{blkdev_open+33} <ffffffff8021d712>{__dentry_open+258}
       <ffffffff802b51a8>{block_ioctl+27} <ffffffff802429a1>{do_ioctl+33}
       <ffffffff8022fb0c>{vfs_ioctl+636} <ffffffff8024d8b8>{sys_ioctl+91}
       <ffffffff8025ea72>{system_call+126}
Code: 48 8b 43 10 48 8b 40 10 48 8b b8 10 01 00 00 e8 cf cd e4 ff

RIP <ffffffff8040e828>{bitmap_create+152} RSP <ffff81003f603ad8>
CR2: 0000000000000010
<0>Kernel panic - not syncing: Attempted to kill init! <0>Rebooting in 60 seconds..

-rc3-mm1 doesn't have this problem, but rc4-mm1 does. It seems like the right devices are found and added but no md is created for them.

Here's some output with -rc3-mm1 which is ok:

[root@tornado ~]# cat /proc/mdstat
Personalities : [raid1]
md1 : active raid1 sdc3[1] sda3[0]

      4891712 blocks [2/2] [UU]
      bitmap: 2/150 pages [8KB], 16KB chunk

md2 : active raid1 sdc5[1] sda5[0]

      4891648 blocks [2/2] [UU]
      bitmap: 19/150 pages [76KB], 16KB chunk

md3 : active raid1 sdc6[1] sda6[0]

      104320 blocks [2/2] [UU]
      bitmap: 1/13 pages [4KB], 4KB chunk

md4 : active raid1 sdc7[1] sda7[0]

      497856 blocks [2/2] [UU]
      bitmap: 3/61 pages [12KB], 4KB chunk

md6 : active raid1 sdc10[1] sda10[0]

      20008832 blocks [2/2] [UU]

md5 : active raid1 sdc11[1] sda11[0]

      20008832 blocks [2/2] [UU]

md0 : active raid1 sdc2[1] sda2[0]

      24410688 blocks [2/2] [UU]
      bitmap: 16/187 pages [64KB], 64KB chunk

unused devices: <none>
[root@tornado ~]#

Box is an x86_64 with raid-1/SATA.

Thanks,
Reuben

#6 lore raw
Quoted 11 lines · On Tuesday May 16, [email protected] wrote:
Quoted 3 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

Indeed. This is the first -mm in a fair while that has crapped itself on boot on my box.

NeilB - is this yours?

Uhmmm. yes. I have a patch that I was going to send of a little later today. Just give me an hour or two...

NeilBrown

#7 lore raw

PATCH 1/1 x86 NUMA panic compile error

In reply to Andrew Morton

Is this check now needed given we have the zone alignment patches in this -mm also? I think we want to make it at least possible to boot such a kernel on a 'flat' machine.

-apw

=== 8< ===
x86 NUMA panic compile error

Seem we have a syntax error rising from the the new panic added to let people know NUMA didn't work on physically non-numa hardware.

---
 srat.c |    2 +-
 1 files changed, 1 insertion(+), 1 deletion(-)
arch/i386/kernel/srat.c +1 -1
diff -upN reference/arch/i386/kernel/srat.c current/arch/i386/kernel/srat.c
@@ -269,7 +269,7 @@ int __init get_memcfg_from_srat(void)
 	extern int use_cyclone;
 	if (use_cyclone == 0) {
 		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
+		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else.";
 		early_printk(s);
 		panic(s);
 	}
#8 lore raw

Re: [PATCH] x86 NUMA panic compile error

arch/i386/kernel/srat.c int __init get_memcfg_from_srat(void) · line 275

Quoted 7 lines · Andy Whitcroft
arch/i386/kernel/srat.c +1 -1
 	if (use_cyclone == 0) {
 		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
+		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else.";
 		early_printk(s);
 		panic(s);
 	}

i still strongly oppose the original Andi hack... numerous reasons were given not to apply it (it's nice to simulate/trigger rarer features on mainstream hardware too, and this ability to boot NUMA on my flat x86 testbox found at least one other NUMA bug already). Furthermore, the crash i reported was fixed by the NUMA patchset. Andrew, please drop:

x86_64-mm-i386-numa-summit-check.patch

(which has nothing to do with x86_64 anyway)

Ingo

#9 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 16 lines · Ingo Molnar

arch/i386/kernel/srat.c int __init get_memcfg_from_srat(void) · line 275

Quoted 7 lines · Andy Whitcroft
arch/i386/kernel/srat.c +1 -1
 	if (use_cyclone == 0) {
 		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
+		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else.";
 		early_printk(s);
 		panic(s);
 	}

i still strongly oppose the original Andi hack... numerous reasons were given not to apply it (it's nice to simulate/trigger rarer features on mainstream hardware too, and this ability to boot NUMA on my flat x86 testbox found at least one other NUMA bug already). Furthermore, the crash i reported was fixed by the NUMA patchset. Andrew, please drop:

The problem is that it's not regularly used on a wide range of boxes so it will eventually break again. We had this cycle several times already.

It's also missing a lot of the workarounds for broken SRATs that are needed for many of the existing NUMA systems.

If there's consensus i386 NUMA is useful I can drop it, but I predict it will just eventually break again.

Quoted 3 lines

x86_64-mm-i386-numa-summit-check.patch

(which has nothing to do with x86_64 anyway)

I have a lot of i386 or combined i386/x86-64 patches in my tree - just Andrew's merge script doesn't pick that up.

-Andi

#10 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 9 lines · Andi Kleen
Quoted 5 lines

i still strongly oppose the original Andi hack... numerous reasons were given not to apply it (it's nice to simulate/trigger rarer features on mainstream hardware too, and this ability to boot NUMA on my flat x86 testbox found at least one other NUMA bug already). Furthermore, the crash i reported was fixed by the NUMA patchset. Andrew, please drop:

The problem is that it's not regularly used on a wide range of boxes so it will eventually break again. We had this cycle several times already.

so it will find new bugs. Since you wrote that patch the ability to try numa on 'flat' x86 hardware has found two bugs, one of which (the zone alignment thing) could very well affect non-obsolete hardware too.

Andi, triggering bugs as widely as possible is the _whole point_ of QA and diversity of testing! Do you disable the compilation of the floppy driver just because it's obsolete hardware? Will you remove NUMA emulation from x86_64 just because it regularly broke in the past?

Fact is, the more hardware a given feature can be tried on, the better tested that feature becomes. I find it quite extreme that this point has to even be argued...

Ingo

#11 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Ingo Molnar
Quoted 4 lines · Ingo Molnar

* Andy Whitcroft <[email protected]> wrote:

Quoted 13 lines

arch/i386/kernel/srat.c int __init get_memcfg_from_srat(void) · line 275

Quoted 7 lines
arch/i386/kernel/srat.c +1 -1
 	if (use_cyclone == 0) {
 		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
+		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else.";
 		early_printk(s);
 		panic(s);
 	}

i still strongly oppose the original Andi hack... numerous reasons were given not to apply it (it's nice to simulate/trigger rarer features on mainstream hardware too, and this ability to boot NUMA on my flat x86 testbox found at least one other NUMA bug already). Furthermore, the crash i reported was fixed by the NUMA patchset.

I'll be darned. I never knew it was even possible to run x86 numa kernels on non-numa boxen. I'd have tested about 1000000 of Christoph Lameter's patches if someone had told me. Yes, it's useful.

Quoted 3 lines

Andrew, please drop:

x86_64-mm-i386-numa-summit-check.patch

bang.

Quoted 1 line

(which has nothing to do with x86_64 anyway)

True.

I guess the concern here is that we don't want people building these frankenkernels and then sending us bug reports against them.

So it is perhaps reasonable to do this panic, but only if !CONFIG_EMBEDDED? (It really is time to start renaming CONFIG_EMBEDDED to CONFIG_DONT_DO_THIS or something).

#12 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 21 lines · Andrew Morton
Quoted 4 lines · Ingo Molnar

* Andy Whitcroft <[email protected]> wrote:

Quoted 13 lines
Quoted 7 lines
	if (use_cyclone == 0) {
		/* Make sure user sees something */
arch/i386/kernel/srat.c +1 -1
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
+		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else.";
		early_printk(s);
		panic(s);
	}

i still strongly oppose the original Andi hack... numerous reasons were given not to apply it (it's nice to simulate/trigger rarer features on mainstream hardware too, and this ability to boot NUMA on my flat x86 testbox found at least one other NUMA bug already). Furthermore, the crash i reported was fixed by the NUMA patchset.

Quoted 4 lines

I'll be darned. I never knew it was even possible to run x86 numa kernels on non-numa boxen. I'd have tested about 1000000 of Christoph Lameter's patches if someone had told me. Yes, it's useful.

We always assumed it might be reasonable for a distro to want a single installer kernel for all machines. So having a combined numa not numa capable kernel always seemed like a good idea.

Quoted 11 lines
Quoted 3 lines

Andrew, please drop:

x86_64-mm-i386-numa-summit-check.patch

bang.

Quoted 1 line

(which has nothing to do with x86_64 anyway)

Quoted 8 lines

True.

I guess the concern here is that we don't want people building these frankenkernels and then sending us bug reports against them.

So it is perhaps reasonable to do this panic, but only if !CONFIG_EMBEDDED? (It really is time to start renaming CONFIG_EMBEDDED to CONFIG_DONT_DO_THIS or something).

How about CONFIG_EXPERIMENTAL?

-apw

#13 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 6 lines · Andy Whitcroft
Quoted 4 lines

So it is perhaps reasonable to do this panic, but only if !CONFIG_EMBEDDED? (It really is time to start renaming CONFIG_EMBEDDED to CONFIG_DONT_DO_THIS or something).

How about CONFIG_EXPERIMENTAL?

Probably CONFIG_ADVANCED would be closer.

#14 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 9 lines · Andrew Morton
Quoted 6 lines · Andy Whitcroft
Quoted 4 lines

So it is perhaps reasonable to do this panic, but only if !CONFIG_EMBEDDED? (It really is time to start renaming CONFIG_EMBEDDED to CONFIG_DONT_DO_THIS or something).

How about CONFIG_EXPERIMENTAL?

Probably CONFIG_ADVANCED would be closer.

Most people who recompile kernels probably think of themselves as advanced. This doesn't mean they will get these things right.

-Andi

#15 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andrew Morton
Quoted 9 lines · Andrew Morton
Quoted 4 lines · Andy Whitcroft
Quoted 2 lines

So it is perhaps reasonable to do this panic, but only if !CONFIG_EMBEDDED? (It really is time to start renaming CONFIG_EMBEDDED to CONFIG_DONT_DO_THIS or something).

How about CONFIG_EXPERIMENTAL?

Probably CONFIG_ADVANCED would be closer.

It defaults to off already - people have to explicitly enable it.

Plus the original point was to be able to build one kernel that'd work across NUMA and non-NUMA boxes.

M.

#16 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andrew Morton
Quoted 19 lines · Andrew Morton
Quoted 4 lines · Ingo Molnar

* Andy Whitcroft <[email protected]> wrote:

Quoted 13 lines

arch/i386/kernel/srat.c int __init get_memcfg_from_srat(void) · line 275

Quoted 7 lines
arch/i386/kernel/srat.c +1 -1
 	if (use_cyclone == 0) {
 		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
+		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else.";
 		early_printk(s);
 		panic(s);
 	}

i still strongly oppose the original Andi hack... numerous reasons were given not to apply it (it's nice to simulate/trigger rarer features on mainstream hardware too, and this ability to boot NUMA on my flat x86 testbox found at least one other NUMA bug already). Furthermore, the crash i reported was fixed by the NUMA patchset.

Quoted 3 lines

I'll be darned. I never knew it was even possible to run x86 numa kernels on non-numa boxen. I'd have tested about 1000000 of Christoph Lameter's patches if someone had told me. Yes, it's useful.

If you want to use it for that I would suggest to port the numa emulation code at least - two or four nodes tends to find more problems than a single node.

But testing on a 64bit box - even with numa emulation - would be much better because on 32bit ZONE_NORMAL often is node 0 only and you won't get much numaness for kernel objects.

Quoted 6 lines
Quoted 4 lines

Andrew, please drop:

x86_64-mm-i386-numa-summit-check.patch

bang.

Ok.

Quoted 7 lines
Quoted 2 lines

(which has nothing to do with x86_64 anyway)

True.

I guess the concern here is that we don't want people building these frankenkernels and then sending us bug reports against them.

Well it will still increase the bug numbers you care so much about.

Another reason I don't like it is that it's ugly and reimplements parts of ACPI on its own for no reason.

If people use it regularly for debugging maybe it won't bit rot as quickly as it used to be, but there is a big difference between theory and practice so we'll see.

Quoted 3 lines

So it is perhaps reasonable to do this panic, but only if !CONFIG_EMBEDDED? (It really is time to start renaming CONFIG_EMBEDDED to CONFIG_DONT_DO_THIS or something).

Fine.

-Andi

#17 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 8 lines · Andi Kleen
Quoted 6 lines
Quoted 1 line

(which has nothing to do with x86_64 anyway)

True.

I guess the concern here is that we don't want people building these frankenkernels and then sending us bug reports against them.

Well it will still increase the bug numbers you care so much about.

so lets ... hide them? ahem? Unaligned NUMA zones were a bug waiting to hit us on some other platform.

Ingo

#18 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andi Kleen
Quoted 6 lines · Andi Kleen
Quoted 5 lines

I'll be darned. I never knew it was even possible to run x86 numa kernels on non-numa boxen. I'd have tested about 1000000 of Christoph Lameter's patches if someone had told me. Yes, it's useful.

Quoted 7 lines

If you want to use it for that I would suggest to port the numa emulation code at least - two or four nodes tends to find more problems than a single node.

But testing on a 64bit box - even with numa emulation - would be much better because on 32bit ZONE_NORMAL often is node 0 only and you won't get much numaness for kernel objects.

That's an excellent point - most developers who are likely to want to test NUMA have x86_64 boxes and x86_64 has NUMA-emulation-on-SMP. I'd semi-forgotten that it existed.

This rather weakens the reasons for retaining support for NUMA-on-non-summit-x86. Ingo?

Quoted 4 lines
Quoted 2 lines

I guess the concern here is that we don't want people building these frankenkernels and then sending us bug reports against them.

Well it will still increase the bug numbers you care so much about.

Not really. If a bug affects something we don't care about (like this) I'll just ignore it. I care about the number of busted machines out there, not the bug counts...

Quoted 2 lines

Another reason I don't like it is that it's ugly and reimplements parts of ACPI on its own for no reason.

So shouldn't such a patch remove that code rather than panicing?

#19 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 6 lines
Quoted 4 lines

Another reason I don't like it is that it's ugly and reimplements parts of ACPI on its own for no reason.

So shouldn't such a patch remove that code rather than panicing?

I would be for remove, but apparently we have one or two users in IBM that run their x440s (32bit only) with CONFIG_NUMA. No distributions do so though and I would expect x440s to usually run distributions because they are quite expensive machines.

My arguments for remove:
- The code is very hackish - it was written before the proper ACPI infrastructure is in place - and NUMA on 32bit in general needs a lot of hacks because of the limited ZONE_NORMAL.
- NUMA on 32bit is kind of broken by design.
- It isn't used much.
- It breaks often
- It tends to not work on Opterons and hits the users who try it there.

Short of remove I would settle on the panic on non Summit.

Also if you only wanted NUMA emulation for testing it could be also done much much simpler removing near all of i386/*/srat.c. But with the inherent 32bit NUMA limitations I have my doubts on its great usefulness.

-Andi

#20 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 11 lines · Andi Kleen
Quoted 1 line

So shouldn't such a patch remove that code rather than panicing?

I would be for remove, but apparently we have one or two users in IBM that run their x440s (32bit only) with CONFIG_NUMA. No distributions do so though and I would expect x440s to usually run distributions because they are quite expensive machines.

My arguments for remove:
- The code is very hackish - it was written before the proper ACPI infrastructure is in place - and NUMA on 32bit in general needs a lot of hacks because of the limited ZONE_NORMAL.

works fine here now. The whole NUMA code is still quite hackish in general, (including most of arch/x86_64/*/*.c), so i'd not judge based on that.

Quoted 1 line

- NUMA on 32bit is kind of broken by design.

well. 32bit itself is broken by design, if you consider RAM larger than say 1GB.

Quoted 1 line

- It isn't used much.

it's an enabler of a feature-set that i couldnt test on these boxes otherwise. Look at it like the highmem= boot option. Or consider it a primitive form of NUMA emulation.

Quoted 1 line

- It breaks often

Martin says he's daily testing it in his grid.

Quoted 2 lines

- It tends to not work on Opterons and hits the users who try it there.

maybe due to the zone alignment problem?

Ingo

#21 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andi Kleen
Quoted 9 lines · Andi Kleen

My arguments for remove:
- The code is very hackish - it was written before the proper ACPI infrastructure is in place - and NUMA on 32bit in general needs a lot of hacks because of the limited ZONE_NORMAL.
- NUMA on 32bit is kind of broken by design.
- It isn't used much.
- It breaks often
- It tends to not work on Opterons and hits the users who try it there.

I think you are right: there are very few end users running with CONFIG_NUMA on normal x86. But, there is a disproportionately large number of developers who do it. There are quite a few IBM (and maybe more via OSDL) developers who's only access to real NUMA hardware is x86 NUMAQs and Summit machines. When somebody says "foo is broken on NUMA", I go right to an x86 box.

Anyway, I'd like to think that we've contributed enough to the generic NUMA code to have earned our keep and allow our little x86 NUMA "hacks" to remain. x86 is a legacy architecture now anyway, right? ;)

-- Dave

#22 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 10 lines

I think you are right: there are very few end users running with CONFIG_NUMA on normal x86. But, there is a disproportionately large number of developers who do it. There are quite a few IBM (and maybe more via OSDL) developers who's only access to real NUMA hardware is x86 NUMAQs and Summit machines. When somebody says "foo is broken on NUMA", I go right to an x86 box.
Anyway, I'd like to think that we've contributed enough to the generic NUMA code to have earned our keep and allow our little x86 NUMA "hacks" to remain.

Yes that is why i did the "only work on Summit" patch as compromise. With that you can have your hacks, but it won't impact anybody else.

Quoted 1 line

x86 is a legacy architecture now anyway, right? ;)

I wish everybody would agree on that @)

-Andi

#23 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 2 lines · Andi Kleen
Quoted 1 line

x86 is a legacy architecture now anyway, right? ;)

I wish everybody would agree on that @)

as far as i'm concerned x86 is obsolete: my main devel and testboxes are 64-bit throughout, and i develop for 64-bit by default.

Nevertheless for hard-to-debug bugs i prefer if they can be reproduced and debugged on 32-bit too, because x86_64 debugging is still quite a PITA and wastes alot of time: for example it has no support for exact kernel stacktraces. Also, the printout of the backtrace is butt-ugly and as un-ergonomic to the human eye as it gets - who came up with that "two-maybe-one function entries per-line" nonsense? [Whoever did it he never had to look at (and make sense of) hundreds of stacktraces in a row.]

Also, the majority of kernel bugs still get reported on 32-bit and most of the testers are on 32-bit. So x86_64 is nice but it still needs some work, mainly in terms of debuggability.

Ingo

#24 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 6 lines · Ingo Molnar

Nevertheless for hard-to-debug bugs i prefer if they can be reproduced and debugged on 32-bit too, because x86_64 debugging is still quite a PITA and wastes alot of time: for example it has no support for exact kernel stacktraces. Also, the printout of the backtrace is butt-ugly and as un-ergonomic to the human eye as it gets

Yes, I find x86_64 traces significantly harder to follow. And I miss the display of the length of the functions (do_md_run+1208 instead of do_md_run+1208/2043). The latter form makes it easier to work out whereabouts in the function things happened.

That, plus the mix of hex and decimal numbers..

Quoted 4 lines

who came up with that
"two-maybe-one function entries per-line" nonsense? [Whoever did it he never had to look at (and make sense of) hundreds of stacktraces in a row.]

Plus they're wide enough to get usefully wordwrapped when someone mails them to you.

#25 lore raw

Re: [PATCH] x86 NUMA panic compile error

[... feels the love ...]

Quoted 15 lines · Andrew Morton
Quoted 6 lines · Ingo Molnar

Nevertheless for hard-to-debug bugs i prefer if they can be reproduced and debugged on 32-bit too, because x86_64 debugging is still quite a PITA and wastes alot of time: for example it has no support for exact kernel stacktraces. Also, the printout of the backtrace is butt-ugly and as un-ergonomic to the human eye as it gets

Yes, I find x86_64 traces significantly harder to follow. And I miss the display of the length of the functions (do_md_run+1208 instead of do_md_run+1208/2043). The latter form makes it easier to work out whereabouts in the function things happened.

That, plus the mix of hex and decimal numbers..

Quoted 7 lines
Quoted 4 lines

who came up with that
"two-maybe-one function entries per-line" nonsense? [Whoever did it he never had to look at (and make sense of) hundreds of stacktraces in a row.]

Plus they're wide enough to get usefully wordwrapped when someone mails them to you.

Hmm, I didn't realize they were _that_ unpopular. If you got the i386 like space wasting backtraces would you guys all switch your development machines to x86-64 ? @)

-Andi

#26 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 3 lines · Andi Kleen

[... feels the love ...]

heh.

Quoted 21 lines · Andrew Morton
Quoted 20 lines
Quoted 6 lines · Ingo Molnar

Nevertheless for hard-to-debug bugs i prefer if they can be reproduced and debugged on 32-bit too, because x86_64 debugging is still quite a PITA and wastes alot of time: for example it has no support for exact kernel stacktraces. Also, the printout of the backtrace is butt-ugly and as un-ergonomic to the human eye as it gets

Yes, I find x86_64 traces significantly harder to follow. And I miss the display of the length of the functions (do_md_run+1208 instead of do_md_run+1208/2043). The latter form makes it easier to work out whereabouts in the function things happened.

That, plus the mix of hex and decimal numbers..

Quoted 4 lines

who came up with that
"two-maybe-one function entries per-line" nonsense? [Whoever did it he never had to look at (and make sense of) hundreds of stacktraces in a row.]

Quoted 7 lines
Quoted 2 lines

Plus they're wide enough to get usefully wordwrapped when someone mails them to you.

Hmm, I didn't realize they were _that_ unpopular. If you got the i386 like space wasting backtraces would you guys all switch your development machines to x86-64 ? @)

Developers use serial consoles for such things. (I discovered `console=uart,...' yesterday. It works nicely as an earlyprintk on ia64..)

It's reports-from-the-field which are the problem.

A lot of these problems can be address by simple cranking up the VGA screen resolution, but I discovered that I don't know how to do that - I've always used `vga=extended', but that doesn't work on an EFI-booted ia64 box.

Does anyone know what the magic option is to make the vga console use 50 rows?

#27 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 24 lines · Andrew Morton
Quoted 24 lines · Andrew Morton
Quoted 22 lines
Quoted 6 lines · Ingo Molnar

Nevertheless for hard-to-debug bugs i prefer if they can be reproduced and debugged on 32-bit too, because x86_64 debugging is still quite a PITA and wastes alot of time: for example it has no support for exact kernel stacktraces. Also, the printout of the backtrace is butt-ugly and as un-ergonomic to the human eye as it gets

Yes, I find x86_64 traces significantly harder to follow. And I miss the display of the length of the functions (do_md_run+1208 instead of do_md_run+1208/2043). The latter form makes it easier to work out whereabouts in the function things happened.

That, plus the mix of hex and decimal numbers..

Quoted 4 lines

who came up with that
"two-maybe-one function entries per-line" nonsense? [Whoever did it he never had to look at (and make sense of) hundreds of stacktraces in a row.]

Plus they're wide enough to get usefully wordwrapped when someone mails them to you.

Quoted 7 lines
Quoted 4 lines

Hmm, I didn't realize they were _that_ unpopular. If you got the i386 like space wasting backtraces would you guys all switch your development machines to x86-64 ? @)

Developers use serial consoles for such things. (I discovered `console=uart,...' yesterday. It works nicely as an earlyprintk on ia64..)

I can also recommend firescope + firewire cards. It's not early yet, but I hope eventually. But it can work without the target still being alive and also does on most laptops.

Quoted 1 line

It's reports-from-the-field which are the problem.

In my experience the biggest problem in the field is that most of it scrolls away. That is why I tweaked the x86-64 format to be as space efficient as possible. That's also why the "executive summary" was added.

But Ingo has a point that it usually doesn't help anyways because backtraces tend to be so overlong now after the code got through 20 callbacks before it can do something actually useful.

Quoted 6 lines

A lot of these problems can be address by simple cranking up the VGA screen resolution, but I discovered that I don't know how to do that - I've always used `vga=extended', but that doesn't work on an EFI-booted ia64 box.

Does anyone know what the magic option is to make the vga console use 50 rows?

I use vga=0x0f07

It's a butt ugly font, but it's the smallest I could find without using the slow fbcon.

If you can't remember the hex number use vga=ask

-Andi

#28 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Ingo Molnar
Quoted 5 lines

Nevertheless for hard-to-debug bugs i prefer if they can be reproduced and debugged on 32-bit too, because x86_64 debugging is still quite a PITA and wastes alot of time: for example it has no support for exact kernel stacktraces.

Hopefully soon.

I think i386 only gained it very recently, so it can't be _that_ big a problem.

The real issue is too deeply nested code like the callback hell we have in some subsystems. Better would be to eliminate that. 2.4 was much nicer in this regard and there has been quite a lot of unnecessary complications in this area when the kernel went to 2.6.

Quoted 5 lines

Also, the printout of the backtrace is butt-ugly and
as un-ergonomic to the human eye as it gets - who came up with that "two-maybe-one function entries per-line" nonsense? [Whoever did it he never had to look at (and make sense of) hundreds of stacktraces in a row.]

The original goal was to make it fit as much as possible on the screen when you don't have a serial/net/fireconsole. But arguably it's less and less useful because the kernel has gotten so huge that most backtraces are very long and scroll away anyways.

-Andi

#29 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 6 lines · Andi Kleen
Quoted 4 lines

Nevertheless for hard-to-debug bugs i prefer if they can be reproduced and debugged on 32-bit too, because x86_64 debugging is still quite a PITA and wastes alot of time: for example it has no support for exact kernel stacktraces.

Hopefully soon.

i've already implemented it for FRAME_POINTERS (i really needed it to not go insane when looking at lock validator output).

As you suggested a few weeks ago the real solution would be a dwarf parser. Maybe ia64's could be taken? Will post the patch for the FRAME_POINTERS solution soon. Sample output:

     [<ffffffff8020af4c>] __show_trace+0x3a/0x66
     [<ffffffff8020b36b>] show_trace+0x17/0x1a
     [<ffffffff80207e36>] show_regs+0x36/0x3c
     [<ffffffff80207e75>] smp_show_regs+0x39/0x52
     [<ffffffff8021559e>] smp_nmi_callback+0x6a/0x85
     [<ffffffff802163f2>] do_nmi+0x69/0x91
     [<ffffffff80601dca>] nmi+0x7e/0x85
     [<ffffffffffffffff>] 0xffffffffffffffff
     [<ffffffff80601ad0>] _spin_unlock_irqrestore+0x3e/0x47
     [<ffffffff8024424c>] prepare_to_wait+0x63/0x6d
     [<ffffffff8022ff0c>] do_syslog+0xf1/0x3ca
     [<ffffffff802ba3ed>] kmsg_read+0x3a/0x46
     [<ffffffff8027db72>] vfs_read+0xe6/0x191
     [<ffffffff8027e79d>] sys_read+0x44/0x82
     [<ffffffff80209b11>] system_call+0x7d/0x83
     [<ffffffffffffffff>] 0xffffffffffffffff

(and it works fine across irq/exception stacks too.)

Quoted 2 lines

I think i386 only gained it very recently, so it can't be _that_ big a problem.

i certainly used exact backtraces on i386 for many many years. Not sure whether those patches were all upstream though. It's also the combination of effects that makes the difference between i386 and x86_64 so striking.

furthermore, the kernel's debugging infrastructure improved significantly, and we get more and more stackdumps to interpret [instead of hard to debug corruptions, etc.].

Quoted 4 lines

The real issue is too deeply nested code like the callback hell we have in some subsystems. Better would be to eliminate that. 2.4 was much nicer in this regard and there has been quite a lot of unnecessary complications in this area when the kernel went to 2.6.

i have no problems with interpreting occasional non-exact backtraces, but it is certainly non-obvious, and when you are looking at backtraces en masse, the unnecessary repetitive task can get really distracting and frustrating.

Exact backtraces on the other hand almost immediately create a unique and reliable "visual fingerprint", and if you have looked at enough of them, you almost recognize them just from looking at the shape of them. It's a completely different 'experience'. (and userspace developers will laugh out loud at us now i suspect ...)

Quoted 9 lines
Quoted 4 lines

Also, the printout of the backtrace is butt-ugly and as un-ergonomic to the human eye as it gets - who came up with that "two-maybe-one function entries per-line" nonsense? [Whoever did it he never had to look at (and make sense of) hundreds of stacktraces in a row.]

The original goal was to make it fit as much as possible on the screen when you don't have a serial/net/fireconsole. But arguably it's less and less useful because the kernel has gotten so huge that most backtraces are very long and scroll away anyways.

yeah.

Ingo

#30 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 3 lines

As you suggested a few weeks ago the real solution would be a dwarf parser. Maybe ia64's could be taken?

Ia64's would be a lot of work.

Quoted 1 line

(and it works fine across irq/exception stacks too.)

It didn't work at all through the old locks/semaphore stubs.

Quoted 4 lines
Quoted 2 lines

I think i386 only gained it very recently, so it can't be _that_ big a problem.

i certainly used exact backtraces on i386 for many many years.

The patch into mainline went in in mid 2004.

http://www.kernel.org/git/?p=linux/kernel/git/tglx/history.git;a=commit;h=066479e379f387e5b1da0f1149fe0b97bac58888

-Andi

#31 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andi Kleen
Followup to:  <[email protected]>
By author:    Andi Kleen <[email protected]>
In newsgroup: linux.dev.kernel
Quoted 4 lines
Quoted 2 lines

x86 is a legacy architecture now anyway, right? ;)

I wish everybody would agree on that @)

It's going to live on for a very long time, though. Intel is still shipping some very fast 64-bit-deficient silicon. Once that's gone, it's going to live on for decades in the embedded world.

-hpa

#32 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 11 lines · H. Peter Anvin
Followup to:  <[email protected]>
By author:    Andi Kleen <[email protected]>
In newsgroup: linux.dev.kernel
Quoted 4 lines
Quoted 2 lines

x86 is a legacy architecture now anyway, right? ;)

I wish everybody would agree on that @)

It's going to live on for a very long time, though. Intel is still shipping some very fast 64-bit-deficient silicon. Once that's gone, it's going to live on for decades in the embedded world.

There's also a surprising number of people still running their shiny new x86-64's in 32 bit mode. (I suspect because they're reliant upon applications that aren't 64-bit clean that for whatever reason don't run in 32-bit emulation).

I'd be surprised if any of the major distros stopped shipping a 32-bit x86 release for several years.

Dave

Signature · 1 line
http://www.codemonkey.org.uk
#33 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 13 lines · Dave Jones
Quoted 11 lines · H. Peter Anvin
Followup to:  <[email protected]>
By author:    Andi Kleen <[email protected]>
In newsgroup: linux.dev.kernel
Quoted 4 lines
Quoted 2 lines

x86 is a legacy architecture now anyway, right? ;)

I wish everybody would agree on that @)

It's going to live on for a very long time, though. Intel is still shipping some very fast 64-bit-deficient silicon. Once that's gone, it's going to live on for decades in the embedded world.

Quoted 8 lines

There's also a surprising number of people still running their shiny new x86-64's in 32 bit mode. (I suspect because they're reliant upon applications that aren't 64-bit clean that for whatever reason don't run in 32-bit emulation).

I'd be surprised if any of the major distros stopped shipping a 32-bit x86 release for several years.

Yup. In fact, I wish Fedora had the 32-bit Firefox, at least as an option. I keep having to add the 32-bit repos for that reason alone :(

-hpa

#34 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andrew Morton
Quoted 10 lines · Andrew Morton
Quoted 3 lines

But testing on a 64bit box - even with numa emulation - would be much better because on 32bit ZONE_NORMAL often is node 0 only and you won't get much numaness for kernel objects.

That's an excellent point - most developers who are likely to want to test NUMA have x86_64 boxes and x86_64 has NUMA-emulation-on-SMP. I'd semi-forgotten that it existed.

This rather weakens the reasons for retaining support for NUMA-on-non-summit-x86. Ingo?

my 64-bit boxes (half of the testbed) are busy ones used for daily testing that i just cannot keep running for days doing stress-tests. Neither can they wait 10 minutes to boot up an allyesconfig kernel. So at least as long as my testconfig is concerned, 32-bit boxes and i386 NUMA still has some place.

(and i'm not at all arguing it's a big thing - it's a minor thing. But i absolutely resist Andi's approach on conceptual grounds. It's backwards, for the reasons i outlined before. Had Andi's patch been in place the zone alignment bug had probably not been found - simple as that. Reducing choice artificially is the kind of thing that decreases the kernel's quality. Improving the quality of the kernel starts with making sure everyone understands how to achieve it - and Andi is one of the largest and most important contributors so i'd really like to make sure he understands my point :-)

Ingo

#35 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andrew Morton
Quoted 6 lines · Andrew Morton
Quoted 1 line

(which has nothing to do with x86_64 anyway)

True.

I guess the concern here is that we don't want people building these frankenkernels and then sending us bug reports against them.

sure - lets simply turn it into a printk, as per the patch below.

it's not like we are being swamped with these bugreports, it seems i was the only one who tried. So lets not over-react it. (and the panic was the worst possible thing we could do.)

Ingo

---

warn users that running CONFIG_NUMA on non-x440 boxes is barely tested.

 arch/i386/kernel/srat.c |    9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
arch/i386/kernel/srat.c +3 -6
Index: linux/arch/i386/kernel/srat.c
===================================================================
@@ -267,12 +267,9 @@ int __init get_memcfg_from_srat(void)
 	int i = 0;
 
 	extern int use_cyclone;
-	if (use_cyclone == 0) {
-		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
-		early_printk(s);
-		panic(s);
-	}
+	/* Make sure user sees something */
+	if (use_cyclone == 0)
+		printk(KERN_WARN "WARNING: Not an IBM x440/NUMAQ and CONFIG_NUMA enabled!\n");
 
 	if (ACPI_FAILURE(acpi_find_root_pointer(ACPI_PHYSICAL_ADDRESSING,
 						rsdp_address))) {
#36 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 33 lines · Ingo Molnar
Quoted 6 lines · Andrew Morton
Quoted 1 line

(which has nothing to do with x86_64 anyway)

True.

I guess the concern here is that we don't want people building these frankenkernels and then sending us bug reports against them.

sure - lets simply turn it into a printk, as per the patch below.

it's not like we are being swamped with these bugreports, it seems i was the only one who tried. So lets not over-react it. (and the panic was the worst possible thing we could do.)

Ingo

---

warn users that running CONFIG_NUMA on non-x440 boxes is barely tested.

 arch/i386/kernel/srat.c |    9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

Index: linux/arch/i386/kernel/srat.c
===================================================================

arch/i386/kernel/srat.c int __init get_memcfg_from_srat(void) · line 269

Quoted 3 lines
arch/i386/kernel/srat.c +0 -0
@@ -267,12 +267,9 @@ int __init get_memcfg_from_srat(void)
 	int i = 0;
 
 	extern int use_cyclone;

argh.

If we didn't do this lazy-ass put-the-declaration-in-the-C-file thing, we'd have noticed that the declaration of use_cyclone is in include/asm-i386/mach-summit/mach_mpparse.h.

use_cyclone isn't defined if !CONFIG_X86_CYCLONE_TIMER. arch/i386/kernel/srat.c is only compiled if X86_SUMMIT || X86_GENERICARCH.

<tries to break it>

I have a config here which has NUMA=y, ACPI_SRAT=y, X86_SUMMIT=n (and X86_SUMMIT_NUMA=y!!) and, for some reason it set X86_CYCLONE_TIMER=y, which is dumb. But it will manage to link with this patch applied.

But still, referencing a variable which is implemented in arch/i386/kernel/timers/timer_cyclone.c from within arch/i386/kernel/srat.c is asking for trouble, no?

#37 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 4 lines

But still, referencing a variable which is implemented in arch/i386/kernel/timers/timer_cyclone.c from within arch/i386/kernel/srat.c is asking for trouble, no?

Probably, but where would you define it instead? It kind of belongs into timer_cyclone.c

-Andi

#38 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 7 lines · Andi Kleen
Quoted 4 lines

But still, referencing a variable which is implemented in arch/i386/kernel/timers/timer_cyclone.c from within arch/i386/kernel/srat.c is asking for trouble, no?

Probably, but where would you define it instead? It kind of belongs into timer_cyclone.c

there should be a asm-i386/cyclone.h that properly defines this. (and to make things more consistent, it could make it 0 even if !CONFIG_X86_CYCLONE_TIMER)

Ingo

#39 lore raw

Re: [PATCH] x86 NUMA panic compile error

In reply to Andrew Morton
Quoted 3 lines · Andrew Morton

If we didn't do this lazy-ass put-the-declaration-in-the-C-file thing, we'd have noticed that the declaration of use_cyclone is in include/asm-i386/mach-summit/mach_mpparse.h.

updated patch below. Or lets drop the original patch that adds the panic?

Ingo

---

 arch/i386/kernel/srat.c |   11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)
arch/i386/kernel/srat.c +5 -6
Index: linux/arch/i386/kernel/srat.c
===================================================================
@@ -266,13 +266,12 @@ int __init get_memcfg_from_srat(void)
 	int tables = 0;
 	int i = 0;
 
+#ifdef CONFIG_X86_CYCLONE_TIMER
 	extern int use_cyclone;
-	if (use_cyclone == 0) {
-		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
-		early_printk(s);
-		panic(s);
-	}
+	/* Make sure user sees something */
+	if (use_cyclone == 0)
+#endif
+		printk(KERN_WARN "WARNING: Not an IBM x440/NUMAQ and CONFIG_NUMA enabled!\n");
 
 	if (ACPI_FAILURE(acpi_find_root_pointer(ACPI_PHYSICAL_ADDRESSING,
 						rsdp_address))) {
#40 lore raw

Re: [PATCH] x86 NUMA panic compile error

Quoted 2 lines · Ingo Molnar

updated patch below. Or lets drop the original patch that adds the panic?

another update: s/KERN_WARN/KERN_WARNING ...

Ingo

---

re-enable dummy NUMA on i386.

 arch/i386/kernel/srat.c |   11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)
arch/i386/kernel/srat.c +5 -6
Index: linux/arch/i386/kernel/srat.c
===================================================================
@@ -266,13 +266,12 @@ int __init get_memcfg_from_srat(void)
 	int tables = 0;
 	int i = 0;
 
+#ifdef CONFIG_X86_CYCLONE_TIMER
 	extern int use_cyclone;
-	if (use_cyclone == 0) {
-		/* Make sure user sees something */
-		static const char s[] __initdata = "Not an IBM x440/NUMAQ. Don't use i386 CONFIG_NUMA anywhere else."
-		early_printk(s);
-		panic(s);
-	}
+	/* Make sure user sees something */
+	if (use_cyclone == 0)
+#endif
+		printk(KERN_WARNING "WARNING: Not an IBM x440/NUMAQ and CONFIG_NUMA enabled!\n");
 
 	if (ACPI_FAILURE(acpi_find_root_pointer(ACPI_PHYSICAL_ADDRESSING,
 						rsdp_address))) {
#41 lore raw
In reply to Andrew Morton

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
<c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
<c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
<c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

Regards,
Michal

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


-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642
#42 lore raw
In reply to Andrew Morton

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
<c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
<c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
<c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#43 lore raw
Quoted 17 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

Quoted 8 lines
snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

#44 lore raw
In reply to Michal Piotrowski
Quoted 17 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

Quoted 8 lines
snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

#45 lore raw

At Mon, 15 May 2006 10:04:29 -0700,

Quoted 25 lines · Andrew Morton
Quoted 23 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Quoted 6 lines
Quoted 2 lines

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

Ouch, I checked only spinlocks but forgot rwlocks. Will fix them.

Thanks.

Takashi

#46 lore raw

At Mon, 15 May 2006 19:30:24 +0200,

Quoted 28 lines · I wrote:

At Mon, 15 May 2006 10:04:29 -0700,

Quoted 25 lines · Andrew Morton
Quoted 23 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

Ouch, I checked only spinlocks but forgot rwlocks. Will fix them.

The patch below should fix them (already merged in ALSA HG repo).

[ALSA] Fix rwlock around snd_iprintf() in sound core

Fixed rwlock around snd_iprintf() in sound core part. Replaced with mutex.

Also, make mutex and flags static variables with addition of snd_card_locked() function (just for sound.c).

---
commit f4e4b6a7b3b404edb1acd72ae3cb26e6246b04bb
tree 6e529ed077718299bf7dff400cedc62767d7bdbe
parent 3a5be5e3bc2d702b9f8428465729895b691c9059
author Takashi Iwai <[email protected]> Mon, 15 May 2006 00:00:00 +0200 committer Takashi Iwai <[email protected]> Mon, 15 May 2006 19:51:18 +0200

 include/sound/core.h |    3 +--
 sound/core/init.c    |   51 ++++++++++++++++++++++++++++++--------------------
 sound/core/sound.c   |    7 +------
 3 files changed, 33 insertions(+), 28 deletions(-)
include/sound/core.h +1 -2
index 5135147..5d184be 100644
@@ -233,9 +233,8 @@ int copy_from_user_toio(volatile void __
 
 /* init.c */
 
-extern unsigned int snd_cards_lock;
 extern struct snd_card *snd_cards[SNDRV_CARDS];
-extern rwlock_t snd_card_rwlock;
+int snd_card_locked(int card);
 #if defined(CONFIG_SND_MIXER_OSS) || defined(CONFIG_SND_MIXER_OSS_MODULE)
 #define SND_MIXER_OSS_NOTIFY_REGISTER	0
 #define SND_MIXER_OSS_NOTIFY_DISCONNECT	1
sound/core/init.c +31 -20

147 lines not on this page · raw message

sound/core/sound.c +1 -6
index 8313f97..02c8cc4 100644
@@ -81,14 +81,9 @@ extern struct class *sound_class;
  */
 void snd_request_card(int card)
 {
-	int locked;
-
 	if (! current->fs->root)
 		return;
-	read_lock(&snd_card_rwlock);
-	locked = snd_cards_lock & (1 << card);
-	read_unlock(&snd_card_rwlock);
-	if (locked)
+	if (snd_card_locked(card))
 		return;
 	if (card < 0 || card >= cards_limit)
 		return;

-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

#47 lore raw
In reply to Takashi Iwai

At Mon, 15 May 2006 19:30:24 +0200,

Quoted 28 lines · I wrote:

At Mon, 15 May 2006 10:04:29 -0700,

Quoted 25 lines · Andrew Morton
Quoted 23 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

Ouch, I checked only spinlocks but forgot rwlocks. Will fix them.

The patch below should fix them (already merged in ALSA HG repo).

[ALSA] Fix rwlock around snd_iprintf() in sound core

Fixed rwlock around snd_iprintf() in sound core part. Replaced with mutex.

Also, make mutex and flags static variables with addition of snd_card_locked() function (just for sound.c).

---
commit f4e4b6a7b3b404edb1acd72ae3cb26e6246b04bb
tree 6e529ed077718299bf7dff400cedc62767d7bdbe
parent 3a5be5e3bc2d702b9f8428465729895b691c9059
author Takashi Iwai <[email protected]> Mon, 15 May 2006 00:00:00 +0200 committer Takashi Iwai <[email protected]> Mon, 15 May 2006 19:51:18 +0200

 include/sound/core.h |    3 +--
 sound/core/init.c    |   51 ++++++++++++++++++++++++++++++--------------------
 sound/core/sound.c   |    7 +------
 3 files changed, 33 insertions(+), 28 deletions(-)
include/sound/core.h +1 -2
index 5135147..5d184be 100644
@@ -233,9 +233,8 @@ int copy_from_user_toio(volatile void __
 
 /* init.c */
 
-extern unsigned int snd_cards_lock;
 extern struct snd_card *snd_cards[SNDRV_CARDS];
-extern rwlock_t snd_card_rwlock;
+int snd_card_locked(int card);
 #if defined(CONFIG_SND_MIXER_OSS) || defined(CONFIG_SND_MIXER_OSS_MODULE)
 #define SND_MIXER_OSS_NOTIFY_REGISTER	0
 #define SND_MIXER_OSS_NOTIFY_DISCONNECT	1
sound/core/init.c +31 -20

147 lines not on this page · raw message

sound/core/sound.c +1 -6
index 8313f97..02c8cc4 100644
@@ -81,14 +81,9 @@ extern struct class *sound_class;
  */
 void snd_request_card(int card)
 {
-	int locked;
-
 	if (! current->fs->root)
 		return;
-	read_lock(&snd_card_rwlock);
-	locked = snd_cards_lock & (1 << card);
-	read_unlock(&snd_card_rwlock);
-	if (locked)
+	if (snd_card_locked(card))
 		return;
 	if (card < 0 || card >= cards_limit)
 		return;
#48 lore raw

sound/core/init.c struct snd_shutdown_f_ops { · line 41

Quoted 3 lines · Takashi Iwai
sound/core/init.c +1 -1
-unsigned int snd_cards_lock = 0;	/* locked for registering/using */
+static unsigned int snd_cards_lock = 0;	/* locked for registering/using */

May as well remove the `= 0' while you're there. It adds four bytes to the module unnecessarily.

sound/core/sound.c extern struct class *sound_class; · line 85

Quoted 6 lines

void snd_request_card(int card)
{

sound/core/sound.c +0 -2
-	int locked;
-
 	if (! current->fs->root)
 		return;

I wonder what that's there for.

-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

#49 lore raw
In reply to Takashi Iwai

sound/core/init.c struct snd_shutdown_f_ops { · line 41

Quoted 3 lines · Takashi Iwai
sound/core/init.c +1 -1
-unsigned int snd_cards_lock = 0;	/* locked for registering/using */
+static unsigned int snd_cards_lock = 0;	/* locked for registering/using */

May as well remove the `= 0' while you're there. It adds four bytes to the module unnecessarily.

sound/core/sound.c extern struct class *sound_class; · line 85

Quoted 6 lines

void snd_request_card(int card)
{

sound/core/sound.c +0 -2
-	int locked;
-
 	if (! current->fs->root)
 		return;

I wonder what that's there for.

#50 lore raw

At Mon, 15 May 2006 11:11:11 -0700,

Quoted 8 lines · Andrew Morton

sound/core/init.c struct snd_shutdown_f_ops { · line 41

Quoted 3 lines · Takashi Iwai
sound/core/init.c +1 -1
-unsigned int snd_cards_lock = 0;	/* locked for registering/using */
+static unsigned int snd_cards_lock = 0;	/* locked for registering/using */

May as well remove the `= 0' while you're there. It adds four bytes to the module unnecessarily.

Yep. Will remove it.

Quoted 8 lines

sound/core/sound.c extern struct class *sound_class; · line 85

Quoted 6 lines

void snd_request_card(int card)
{

sound/core/sound.c +0 -2
-	int locked;
-
 	if (! current->fs->root)
 		return;

I wonder what that's there for.

This is almost for backward compatibility only. We have already module aliases, so it can be cleaned up.

Takashi

-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

#51 lore raw
In reply to Andrew Morton

At Mon, 15 May 2006 11:11:11 -0700,

Quoted 8 lines · Andrew Morton

sound/core/init.c struct snd_shutdown_f_ops { · line 41

Quoted 3 lines · Takashi Iwai
sound/core/init.c +1 -1
-unsigned int snd_cards_lock = 0;	/* locked for registering/using */
+static unsigned int snd_cards_lock = 0;	/* locked for registering/using */

May as well remove the `= 0' while you're there. It adds four bytes to the module unnecessarily.

Yep. Will remove it.

Quoted 8 lines

sound/core/sound.c extern struct class *sound_class; · line 85

Quoted 6 lines

void snd_request_card(int card)
{

sound/core/sound.c +0 -2
-	int locked;
-
 	if (! current->fs->root)
 		return;

I wonder what that's there for.

This is almost for backward compatibility only. We have already module aliases, so it can be cleaned up.

Takashi

#52 lore raw
In reply to Takashi Iwai

Hi,

Quoted 33 lines · Takashi Iwai

At Mon, 15 May 2006 19:30:24 +0200,

Quoted 31 lines · I wrote:

At Mon, 15 May 2006 10:04:29 -0700,

Quoted 28 lines · Andrew Morton
Quoted 25 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

Quoted 9 lines
Quoted 5 lines
Quoted 3 lines

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

Ouch, I checked only spinlocks but forgot rwlocks. Will fix them.

The patch below should fix them (already merged in ALSA HG repo).

Problem fixed. Thanks!

Regards,
Michal

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


-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642
#53 lore raw
In reply to Takashi Iwai

Hi,

Quoted 33 lines · Takashi Iwai

At Mon, 15 May 2006 19:30:24 +0200,

Quoted 31 lines · I wrote:

At Mon, 15 May 2006 10:04:29 -0700,

Quoted 28 lines · Andrew Morton
Quoted 25 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

Quoted 9 lines
Quoted 5 lines
Quoted 3 lines

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

Ouch, I checked only spinlocks but forgot rwlocks. Will fix them.

The patch below should fix them (already merged in ALSA HG repo).

Problem fixed. Thanks!

Regards,
Michal

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

At Mon, 15 May 2006 10:04:29 -0700,

Quoted 25 lines · Andrew Morton
Quoted 23 lines · Michal Piotrowski

Hi,

Quoted 4 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

git-alsa.patch

BUG: sleeping function called from invalid context at

/usr/src/linux-mm/sound/core/info.c:117
in_atomic():1, irqs_disabled():0
<c1003ef9> show_trace+0xd/0xf <c100440c> dump_stack+0x17/0x19
<c10178ce> __might_sleep+0x93/0x9d <f988eeb5> snd_iprintf+0x1b/0x84 [snd]
<f988d808> snd_card_module_info_read+0x34/0x4e [snd] <f988f197>

snd_info_entry_open+0x20f/0x2cc [snd]
 <c1067a17> __dentry_open+0x133/0x260   <c1067bb7> nameidata_to_filp+0x1c/0x2e
 <c1067bf7> do_filp_open+0x2e/0x35   <c1068bf2> do_sys_open+0x54/0xd7
 <c1068ca1> sys_open+0x16/0x18   <c11dab67> sysenter_past_esp+0x54/0x75

Non-volatile memory driver v1.2

Quoted 6 lines
Quoted 2 lines

Here is dmesg http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

heh, yes, Takashi baited his hook and pulled in a big one.

snd_card_module_info_read() calls snd_iprintf() under read_lock()..

Ouch, I checked only spinlocks but forgot rwlocks. Will fix them.

Thanks.

Takashi

-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

#55 lore raw
In reply to Andrew Morton
Quoted 1 line

- Added the ecryptfs filesystem

CC [M] fs/ecryptfs/super.o
fs/ecryptfs/super.c: In function `ecryptfs_statfs':
fs/ecryptfs/super.c:129: warning: passing arg 1 of `vfs_statfs' from incompatible pointer type fs/ecryptfs/super.c: At top level:
fs/ecryptfs/super.c:207: warning: initialization from incompatible pointer type

* ->statfs wants vfsmount as first argument
* ecryptfs_statfs() is inlined

#56 lore raw
Quoted 8 lines · Alexey Dobriyan
Quoted 2 lines

- Added the ecryptfs filesystem

CC [M] fs/ecryptfs/super.o
fs/ecryptfs/super.c: In function `ecryptfs_statfs':
fs/ecryptfs/super.c:129: warning: passing arg 1 of `vfs_statfs' from incompatible pointer type fs/ecryptfs/super.c: At top level:
fs/ecryptfs/super.c:207: warning: initialization from incompatible pointer type

hm, I wonder why I didn't notice that.

Quoted 2 lines

* ->statfs wants vfsmount as first argument
* ecryptfs_statfs() is inlined

yup. Fixed a bunch of those, let one slip through.

I don't immediately see how to fix this one, actually:

static inline int ecryptfs_statfs(struct super_block *sb, struct kstatfs *buf)

{
	return vfs_statfs(ecryptfs_superblock_to_lower(sb), buf);
}

Once we've run ecryptfs_superblock_to_lower() to get the "lower superblock", we need to turn that back into a vfsmount for vfs_statfs()..

(and that function needn't have been inlined - it's only ever called indirectly)

I think I'll be dropping the fs-cache patches (again) fairly soon. They're intrusive, quite some effort to carry, they're not getting adequate review (afaict) and there doesn't seem to be a lot of demand for them, sorry.

#57 lore raw
Quoted 27 lines · Andrew Morton
Quoted 8 lines · Alexey Dobriyan
Quoted 2 lines

- Added the ecryptfs filesystem

CC [M] fs/ecryptfs/super.o
fs/ecryptfs/super.c: In function `ecryptfs_statfs':
fs/ecryptfs/super.c:129: warning: passing arg 1 of `vfs_statfs' from incompatible pointer type fs/ecryptfs/super.c: At top level:
fs/ecryptfs/super.c:207: warning: initialization from incompatible pointer type

hm, I wonder why I didn't notice that.

Quoted 2 lines

* ->statfs wants vfsmount as first argument
* ecryptfs_statfs() is inlined

yup. Fixed a bunch of those, let one slip through.

I don't immediately see how to fix this one, actually:

static inline int ecryptfs_statfs(struct super_block *sb, struct kstatfs *buf)

{
	return vfs_statfs(ecryptfs_superblock_to_lower(sb), buf);
}

Once we've run ecryptfs_superblock_to_lower() to get the "lower superblock", we need to turn that back into a vfsmount for vfs_statfs()..

Quoted 6 lines

(and that function needn't have been inlined - it's only ever called indirectly)

I think I'll be dropping the fs-cache patches (again) fairly soon. They're intrusive, quite some effort to carry, they're not getting adequate review (afaict) and there doesn't seem to be a lot of demand for them, sorry.

Silly me. GFS2 is guilty too

fs/gfs2/ops_super.c:371: warning: initialization from incompatible pointer type

static int gfs2_statfs(struct super_block *sb, struct kstatfs *buf)

#58 lore raw
In reply to Andrew Morton
Quoted 10 lines · Andrew Morton

I don't immediately see how to fix this one, actually:

static inline int ecryptfs_statfs(struct super_block *sb, struct kstatfs *buf)

{
	return vfs_statfs(ecryptfs_superblock_to_lower(sb), buf);
}

Once we've run ecryptfs_superblock_to_lower() to get the "lower superblock", we need to turn that back into a vfsmount for vfs_statfs()..

Assuming we have all our other ducks in a row, this should work. I still need to finish the build and run my tests to say for certain, but here is a tentative fix.

---

fs/ecryptfs/super.c 2006-05-15 14:27:45.000000000 -0500 renamed from fs/ecryptfs/super.c 2006-05-15 14:16:15.000000000 -0500 +5 -2
Index: linux-2.6.17-rc4-mm1-ecryptfs/fs/ecryptfs/super.c
===================================================================
@@ -124,9 +124,12 @@
  * Get the filesystem statistics. Currently, we let this pass right through
  * to the lower filesystem and take no action ourselves.
  */
-static inline int ecryptfs_statfs(struct super_block *sb, struct kstatfs *buf)
+static int ecryptfs_statfs(struct vfsmount *vfs_mnt, struct kstatfs *buf)
 {
-       return vfs_statfs(ecryptfs_superblock_to_lower(sb), buf);
+       struct vfsmount *lower_mnt;
+
+       lower_mnt = ecryptfs_superblock_to_private(vfs_mnt->mnt_sb)->lower_mnt;
+       return vfs_statfs(lower_mnt, buf);
 }
 
 /**
#59 lore raw
In reply to Andrew Morton

Hi,

Quoted 5 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

+kbuild-export-symbol-usage-report-generator.patch

When I try make exportcheck with O=/dir I get this error [michal@ltg01-fedora linux-mm]$ make O=../linux-mm-obj/ exportcheck EXPORTFILE=../linux-mm-obj/export.dat
Using /usr/src/linux-mm as source for kernel
[..]
Can't open perl script "/usr/src/linux-mm-obj/scripts/export_report.pl": No such file or directory
make[2]: *** [__modpost] Error 2
make[1]: *** [modules] Error 2
make: *** [exportcheck] Error 2

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#60 lore raw
Quoted 24 lines · Michal Piotrowski

Hi,

Quoted 5 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

[snip]

Quoted 1 line

+kbuild-export-symbol-usage-report-generator.patch

When I try make exportcheck with O=/dir I get this error [michal@ltg01-fedora linux-mm]$ make O=../linux-mm-obj/ exportcheck EXPORTFILE=../linux-mm-obj/export.dat
Using /usr/src/linux-mm as source for kernel
[..]
Can't open perl script "/usr/src/linux-mm-obj/scripts/export_report.pl": No such file or directory
make[2]: *** [__modpost] Error 2
make[1]: *** [modules] Error 2
make: *** [exportcheck] Error 2

Quoted 3 lines

Regards,
Michal

cc added.

#61 lore raw
In reply to Andrew Morton
Quoted 3 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

Hi Andrew,

While going through my patches currently in -mm I noticed a problem. I have to hang my head i shame and admit to introducing a bug in what should have been a bugFIX patch :-(

In the isdn-unsafe-interaction-between-isdn_write-and-isdn_writebuf_stub.patch patch, which is currently in -mm there's a bug.

This bit :
- copy_from_user(skb_put(skb, len), buf, len);
+ if (!copy_from_user(skb_put(skb, len), buf, len))
should really be :
- copy_from_user(skb_put(skb, len), buf, len);
+ if (copy_from_user(skb_put(skb, len), buf, len))
Somehow a stray "!" crept in there.

Sorry about that. Here's a tiny fix-it-up patch to be applied on top :

drivers/isdn/i4l/isdn_common.c +1 -1
@@ -1952,7 +1952,7 @@ isdn_writebuf_stub(int drvidx, int chan,
 	if (!skb)
 		return -ENOMEM;
 	skb_reserve(skb, hl);
-	if (!copy_from_user(skb_put(skb, len), buf, len))
+	if (copy_from_user(skb_put(skb, len), buf, len))
 		return -EFAULT;
 	ret = dev->drv[drvidx]->interface->writebuf_skb(drvidx, chan, 1, skb);
 	if (ret <= 0)
#62 lore raw
In reply to Andrew Morton

Hi,

Quoted 3 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

When I try to "modprobe -r i2c_i801" modprobe hangs

[michal@ltg01-fedora ~]$ ps aux | grep mod
root 5943 0.0 0.0 1648 432 tty1 D+ 20:15 0:00
modprobe -r i2c_i801
michal 15499 0.0 0.0 1836 496 pts/4 S+ 20:33 0:00 grep mod

Here is strace log
http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/strace.txt
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

2.6.17-rc3-mm1 was fine. I don't see nothing abnormal in dmesg.

Regards,
Michal

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

Hi,

Quoted 3 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

When I try to "modprobe -r i2c_i801" modprobe hangs

[michal@ltg01-fedora ~]$ ps aux | grep mod
root 5943 0.0 0.0 1648 432 tty1 D+ 20:15 0:00
modprobe -r i2c_i801
michal 15499 0.0 0.0 1836 496 pts/4 S+ 20:33 0:00 grep mod

Quoted 6 lines

Here is strace log
http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/strace.txt
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

2.6.17-rc3-mm1 was fine. I don't see nothing abnormal in dmesg.

Are you able to get a sysrq-P and/or sysrq-T trace out of it?

#64 lore raw
Quoted 16 lines · Andrew Morton
Quoted 15 lines · Michal Piotrowski

Hi,

Quoted 3 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

When I try to "modprobe -r i2c_i801" modprobe hangs

[michal@ltg01-fedora ~]$ ps aux | grep mod
root 5943 0.0 0.0 1648 432 tty1 D+ 20:15 0:00
modprobe -r i2c_i801
michal 15499 0.0 0.0 1836 496 pts/4 S+ 20:33 0:00 grep mod

Quoted 9 lines
Quoted 6 lines

Here is strace log
http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/strace.txt
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

2.6.17-rc3-mm1 was fine. I don't see nothing abnormal in dmesg.

Are you able to get a sysrq-P and/or sysrq-T trace out of it?

Something like
echo "t" > /proc/sysrq-trigger
modprobe -r i2c_i801
echo "t" > /proc/sysrq-trigger
?
http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg2

I'm not sure why when I do "echo "p" > /proc/sysrq-trigger" I get only this [root@ltg01-fedora ~]# echo "p" > /proc/sysrq-trigger
SysRq : Show Regs

selinux?

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#65 lore raw
Quoted 28 lines · Michal Piotrowski
Quoted 16 lines · Andrew Morton
Quoted 15 lines · Michal Piotrowski

Hi,

Quoted 3 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

When I try to "modprobe -r i2c_i801" modprobe hangs

[michal@ltg01-fedora ~]$ ps aux | grep mod
root 5943 0.0 0.0 1648 432 tty1 D+ 20:15 0:00
modprobe -r i2c_i801
michal 15499 0.0 0.0 1836 496 pts/4 S+ 20:33 0:00 grep mod

Quoted 9 lines
Quoted 6 lines

Here is strace log
http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/strace.txt
Here is config http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-config

2.6.17-rc3-mm1 was fine. I don't see nothing abnormal in dmesg.

Are you able to get a sysrq-P and/or sysrq-T trace out of it?

Quoted 6 lines

Something like
echo "t" > /proc/sysrq-trigger
modprobe -r i2c_i801
echo "t" > /proc/sysrq-trigger
?
http://www.stardust.webpages.pl/files/mm/2.6.17-rc4-mm1/mm-dmesg2

Great, thanks. Here's the relevant part:

modprobe D 00000019 2740 2163 2129 (NOTLB)

       f0915e60 f1d7b694 8422805f 00000019 f0915e08 c10819a3 f1d7b694 f0915e18 
       00000007 f7e5f374 f7e5f250 f7f0f110 c741cf00 8429b934 00000019 000738d5 
       00000001 f1e6ac54 c101732e f0915e48 c10168d4 f887dba0 00000246 00000001 
Call Trace:
 <c11d768e> wait_for_completion+0x8e/0x108   <c1180fae> i2c_del_adapter_nolock+0x255/0x277
 <c1180fe7> i2c_del_adapter+0x17/0x28   <f887c023> i801_remove+0xd/0x2f [i2c_i801]
 <c10eeb69> pci_device_remove+0x19/0x2c   <c11395a7> __device_release_driver+0x63/0x79
 <c1139889> driver_detach+0x94/0xc4   <c1138d58> bus_remove_driver+0x5d/0x79
 <c11399a2> driver_unregister+0xb/0x18   <c10eecd4> pci_unregister_driver+0x13/0xa5
 <f887c9a5> i2c_i801_exit+0xd/0xf [i2c_i801]   <c103d3a8> sys_delete_module+0x19e/0x1d6
 <c11dab67> sysenter_past_esp+0x54/0x75  

I'd assume that Kumar's i2c-add-support-for-virtual-i2c-adapters.patch is the culprit.

Quoted 5 lines

I'm not sure why when I do "echo "p" > /proc/sysrq-trigger" I get only this [root@ltg01-fedora ~]# echo "p" > /proc/sysrq-trigger
SysRq : Show Regs

selinux?

The p' command isn't effective when using /proc/sysrq-trigger. p' will show the backtrace of the currently-running process (on the current cpu). That task is always "echo". Boring. So `p' is really only useful when invoked from interrupt context, via sysrq-p.

#66 lore raw
Quoted 15 lines · Andrew Morton

Great, thanks. Here's the relevant part:

modprobe D 00000019 2740 2163 2129 (NOTLB)

       f0915e60 f1d7b694 8422805f 00000019 f0915e08 c10819a3 f1d7b694 f0915e18
       00000007 f7e5f374 f7e5f250 f7f0f110 c741cf00 8429b934 00000019 000738d5
       00000001 f1e6ac54 c101732e f0915e48 c10168d4 f887dba0 00000246 00000001
Call Trace:
 <c11d768e> wait_for_completion+0x8e/0x108   <c1180fae> i2c_del_adapter_nolock+0x255/0x277
 <c1180fe7> i2c_del_adapter+0x17/0x28   <f887c023> i801_remove+0xd/0x2f [i2c_i801]
 <c10eeb69> pci_device_remove+0x19/0x2c   <c11395a7> __device_release_driver+0x63/0x79
 <c1139889> driver_detach+0x94/0xc4   <c1138d58> bus_remove_driver+0x5d/0x79
 <c11399a2> driver_unregister+0xb/0x18   <c10eecd4> pci_unregister_driver+0x13/0xa5
 <f887c9a5> i2c_i801_exit+0xd/0xf [i2c_i801]   <c103d3a8> sys_delete_module+0x19e/0x1d6
 <c11dab67> sysenter_past_esp+0x54/0x75
Quoted 2 lines

I'd assume that Kumar's i2c-add-support-for-virtual-i2c-adapters.patch is the culprit.

Unfortunately it's not this patch.
I'll check all Kumar's patches.

Regards,
Michal

Signature · 3 lines
Michal K. K. Piotrowski
LTG - Linux Testers Group
(http://www.stardust.webpages.pl/ltg/wiki/)
#67 lore raw
Quoted 5 lines
Quoted 2 lines

I'd assume that Kumar's i2c-add-support-for-virtual-i2c-adapters.patch is the culprit.

Unfortunately it's not this patch.
I'll check all Kumar's patches.

There are no differences in i2c patches between 2.6.17-rc3-mm1 and 2.6.17-rc4-mm1. As you said the bug is new in 2.6.17-rc4-mm1, this suggests that the bug is not in the i2c patches. Could be that these patches need to be updated to reflect a recent change in another tree though.

I'll try to reproduce the bug here.

Signature · 1 line
Jean Delvare
#68 lore raw

Quoting myself:

Quoted 1 line

I'll try to reproduce the bug here.

I can reproduce it, with both i2c-i801 and i2c-parport, so it's not related to a specific driver. I'm currently performing a bisection on 2.6.17-rc4-mm1 to try and isolate the culprit. It seems to point to gregkh-driver-*. i2c patches are innocent for sure, including Kumar's ones.

Signature · 1 line
Jean Delvare
#69 lore raw

Michal Piotrowski reported:

Quoted 1 line
Quoted 1 line

When I try to "modprobe -r i2c_i801" modprobe hangs

Quoting myself:

Quoted 5 lines

I can reproduce it, with both i2c-i801 and i2c-parport, so it's not related to a specific driver. I'm currently performing a bisection on 2.6.17-rc4-mm1 to try and isolate the culprit. It seems to point to gregkh-driver-*. i2c patches are innocent for sure, including Kumar's ones.

And the winner is... gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

Stephen, Greg?

Signature · 1 line
Jean Delvare
#70 lore raw
Quoted 14 lines · Jean Delvare

Michal Piotrowski reported:

Quoted 2 lines
Quoted 2 lines

When I try to "modprobe -r i2c_i801" modprobe hangs

Quoting myself:

Quoted 6 lines

I can reproduce it, with both i2c-i801 and i2c-parport, so it's not related to a specific driver. I'm currently performing a bisection on 2.6.17-rc4-mm1 to try and isolate the culprit. It seems to point to gregkh-driver-*. i2c patches are innocent for sure, including Kumar's ones.

Quoted 6 lines

And the winner is... gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

Stephen, Greg?

Look at the error return from class_device_register. I bet it was working before because the class_device_register wasn't checking something, now it is and that exposes a bug that has existed in i2c since sysfs support was added.

I can make up a more verbose version if that helps

#71 lore raw
In reply to Jean Delvare

Quoting myself:

Quoted 4 lines

And the winner is... gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

Stephen, Greg?

Indeed this patch breaks class_device_add in the success case...

Andrew, maybe you want to put this in the hot-fixes directory for 2.6.17-rc1-mm4, as the problem hits all drivers which register with a class.

Fix class_device_add success case after gregkh-driver-driver-core-class_device_add-needs-error-checks.patch broke it. class_dev was no more put and class_name was no more freed before leaving. The former caused locks on driver removal (class_dev usage count could never be 0.)

This fix should be folded into gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

---
 drivers/base/class.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/base/class.c +1 -1
@@ -620,7 +620,7 @@
 	}
 	up(&parent_class->sem);
 
-	return 0;
+	goto out1;
 
  out8:
 	if (class_dev->dev)
Signature · 1 line
Jean Delvare
#72 lore raw
Quoted 29 lines · Jean Delvare

Quoting myself:

Quoted 4 lines

And the winner is... gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

Stephen, Greg?

Indeed this patch breaks class_device_add in the success case...

Andrew, maybe you want to put this in the hot-fixes directory for 2.6.17-rc1-mm4, as the problem hits all drivers which register with a class.

Fix class_device_add success case after gregkh-driver-driver-core-class_device_add-needs-error-checks.patch broke it. class_dev was no more put and class_name was no more freed before leaving. The former caused locks on driver removal (class_dev usage count could never be 0.)

This fix should be folded into gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

---
 drivers/base/class.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/base/class.c line 624

Quoted 6 lines
drivers/base/class.c +1 -1
@@ -620,7 +620,7 @@
 	}
 	up(&parent_class->sem);
 
-	return 0;
+	goto out1;
 

Thanks for catching this. I've merged this with Stephen's original patch.

thanks again,

greg k-h

#73 lore raw
Quoted 40 lines · On Tue, 16 May 2006 09:08:42 -0700 Greg KH <[email protected]> wrote:
Quoted 29 lines · Jean Delvare

Quoting myself:

Quoted 4 lines

And the winner is... gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

Stephen, Greg?

Indeed this patch breaks class_device_add in the success case...

Andrew, maybe you want to put this in the hot-fixes directory for 2.6.17-rc1-mm4, as the problem hits all drivers which register with a class.

Fix class_device_add success case after gregkh-driver-driver-core-class_device_add-needs-error-checks.patch broke it. class_dev was no more put and class_name was no more freed before leaving. The former caused locks on driver removal (class_dev usage count could never be 0.)

This fix should be folded into gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

---
 drivers/base/class.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/base/class.c line 624

Quoted 6 lines
drivers/base/class.c +1 -1
@@ -620,7 +620,7 @@
 	}
 	up(&parent_class->sem);
 
-	return 0;
+	goto out1;
 

Thanks for catching this. I've merged this with Stephen's original patch.

Quoted 3 lines

thanks again,

greg k-h

ditto, hmm, one more need for a test suite for basic driver stuff

#74 lore raw
Quoted 43 lines · Stephen Hemminger
Quoted 40 lines · On Tue, 16 May 2006 09:08:42 -0700 Greg KH <[email protected]> wrote:
Quoted 29 lines · Jean Delvare

Quoting myself:

Quoted 4 lines

And the winner is... gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

Stephen, Greg?

Indeed this patch breaks class_device_add in the success case...

Andrew, maybe you want to put this in the hot-fixes directory for 2.6.17-rc1-mm4, as the problem hits all drivers which register with a class.

Fix class_device_add success case after gregkh-driver-driver-core-class_device_add-needs-error-checks.patch broke it. class_dev was no more put and class_name was no more freed before leaving. The former caused locks on driver removal (class_dev usage count could never be 0.)

This fix should be folded into gregkh-driver-driver-core-class_device_add-needs-error-checks.patch

---
 drivers/base/class.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/base/class.c line 624

Quoted 6 lines
drivers/base/class.c +1 -1
@@ -620,7 +620,7 @@
 	}
 	up(&parent_class->sem);
 
-	return 0;
+	goto out1;
 

Thanks for catching this. I've merged this with Stephen's original patch.

Quoted 5 lines
Quoted 3 lines

thanks again,

greg k-h

ditto, hmm, one more need for a test suite for basic driver stuff

If you know of some way to create one, I would be glad to have it...

thanks,

greg k-h

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

- The cachefs patches are back (local-disk-based caching of network filesystem files) (David Howells)

the blind hacks below are needed to make CONFIG_CACHEFILES build.

Ingo

fs/cachefiles/cf-namei.c +11 -11
Index: linux/fs/cachefiles/cf-namei.c
===================================================================
@@ -124,31 +124,31 @@ try_again:
 	}
 
 	/* do the multiway lock magic */
-	trap = lock_rename(cache->graveyard, dir);
+	trap = lock_rename(cache->graveyard, dir, 0);
 
 	/* do some checks before getting the grave dentry */
 	if (rep->d_parent != dir) {
 		/* the entry was probably culled when we dropped the parent dir
 		 * lock */
-		unlock_rename(cache->graveyard, dir);
+		unlock_rename(cache->graveyard, dir, 0);
 		_leave(" = 0 [culled?]");
 		return 0;
 	}
 
 	if (!S_ISDIR(cache->graveyard->d_inode->i_mode)) {
-		unlock_rename(cache->graveyard, dir);
+		unlock_rename(cache->graveyard, dir, 0);
 		cachefiles_io_error(cache, "Graveyard no longer a directory");
 		return -EIO;
 	}
 
 	if (trap == rep) {
-		unlock_rename(cache->graveyard, dir);
+		unlock_rename(cache->graveyard, dir, 0);
 		cachefiles_io_error(cache, "May not make directory loop");
 		return -EIO;
 	}
 
 	if (d_mountpoint(rep)) {
-		unlock_rename(cache->graveyard, dir);
+		unlock_rename(cache->graveyard, dir, 0);
 		cachefiles_io_error(cache, "Mountpoint in cache");
 		return -EIO;
 	}
@@ -160,7 +160,7 @@ try_again:
 
 		grave = d_alloc(cache->graveyard, &name);
 		if (!grave) {
-			unlock_rename(cache->graveyard, dir);
+			unlock_rename(cache->graveyard, dir, 0);
 			_leave(" = -ENOMEM");
 			return -ENOMEM;
 		}
@@ -168,7 +168,7 @@ try_again:
 		alt = cache->graveyard->d_inode->i_op->lookup(
 			cache->graveyard->d_inode, grave, NULL);
 		if (IS_ERR(alt)) {
-			unlock_rename(cache->graveyard, dir);
+			unlock_rename(cache->graveyard, dir, 0);
 			dput(grave);
 
 			if (PTR_ERR(alt) == -ENOMEM) {
@@ -188,7 +188,7 @@ try_again:
 	}
 
 	if (grave->d_inode) {
-		unlock_rename(cache->graveyard, dir);
+		unlock_rename(cache->graveyard, dir, 0);
 		dput(grave);
 		grave = NULL;
 		cond_resched();
@@ -196,7 +196,7 @@ try_again:
 	}
 
 	if (d_mountpoint(grave)) {
-		unlock_rename(cache->graveyard, dir);
+		unlock_rename(cache->graveyard, dir, 0);
 		dput(grave);
 		cachefiles_io_error(cache, "Mountpoint in graveyard");
 		return -EIO;
@@ -204,7 +204,7 @@ try_again:
 
 	/* target should not be an ancestor of source */
 	if (trap == grave) {
-		unlock_rename(cache->graveyard, dir);
+		unlock_rename(cache->graveyard, dir, 0);
 		dput(grave);
 		cachefiles_io_error(cache, "May not make directory loop");
 		return -EIO;
@@ -231,7 +231,7 @@ try_again:
 
 	fsnotify_oldname_free(old_name);
 
-	unlock_rename(cache->graveyard, dir);
+	unlock_rename(cache->graveyard, dir, 0);
 	dput(grave);
 	_leave(" = 0");
 	return 0;

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

#78 lore raw

2.6.17-rc4-mm1: no help text for MTD_NAND_CS553

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-mtd.patch
...
git trees
...

This patch adds an MTD_NAND_CS553 option without a help text.

Please add at least a small help text.

TIA
Adrian

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

Re: 2.6.17-rc4-mm1: no help text for MTD_NAND_CS553

Quoted 3 lines · Adrian Bunk

This patch adds an MTD_NAND_CS553 option without a help text.

Please add at least a small help text.

Done; thanks.

config MTD_NAND_CS553X
tristate "NAND support for CS5535/CS5536 (AMD Geode companion chip)" depends on MTD_NAND && X86_PC && PCI
help
The CS553x companion chips for the AMD Geode processor include NAND flash controllers with built-in hardware ECC capabilities; enabling this option will allow you to use these. The driver will check the MSRs to verify that the controller is enabled for NAND, and currently requires that the controller be in MMIO mode.

If you say "m", the module will be called "cs553x_nand.ko".

Signature · 1 line
dwmw2
#80 lore raw

PATCH 1/1 drivers/mtd/devices/docprobe.c: correct #if's

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-mtd.patch
...
git trees
...

If we correct the names of the config options, the code might actually work as intended...

---

 drivers/mtd/devices/docprobe.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/mtd/devices/docprobe.c renamed from drivers/mtd/devices/docprobe.c.old +3 -3
@@ -231,21 +231,21 @@
 
 static int docfound;
 
-#ifdef CONFIG_DOC2000
+#ifdef CONFIG_MTD_DOC2000
 extern void DoC2k_init(struct mtd_info *);
 #define doc2k_initfunc (&DoC2k_init)
 #else 
 #define doc2k_initfunc NULL
 #endif
 
-#ifdef CONFIG_DOC2001
+#ifdef CONFIG_MTD_DOC2001
 extern void DoCMil_init(struct mtd_info *);
 #define docmil_initfunc (&DoCMil_init)
 #else 
 #define docmil_initfunc NULL
 #endif
 
-#ifdef CONFIG_DOC2001PLUS
+#ifdef CONFIG_MTD_DOC2001PLUS
 extern void DoCMilPlus_init(struct mtd_info *);
 #define docmplus_initfunc (&DoCMilPlus_init)
 #else 
#81 lore raw

Re: [-mm patch] drivers/mtd/devices/docprobe.c: correct #if's

Quoted 2 lines · Adrian Bunk

If we correct the names of the config options, the code might actually work as intended...

S'true. Applied; thanks.

I've dug out one or two of these devices from the toy cupboard -- I want to make sure we get the new version of the driver, using the generic NAND code, working with all of them... and then we can drop the old drivers altogether.

Signature · 1 line
dwmw2
#82 lore raw

PATCH 1/1 make dvb/b2c2/flexcop-fe-tuner.c:alps_tdee4_stv0297_tuner_set_params() static

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-dvb.patch
...
git trees
...

This patch makes the needlessly global
alps_tdee4_stv0297_tuner_set_params() static.

drivers/media/dvb/b2c2/flexcop-fe-tuner.c renamed from drivers/media/dvb/b2c2/flexcop-fe-tuner.c.old +2 -1
@@ -354,7 +354,8 @@
 	.demod_address = 0x0e,
 };
 
-int alps_tdee4_stv0297_tuner_set_params (struct dvb_frontend* fe, struct dvb_frontend_parameters *fep)
+static int alps_tdee4_stv0297_tuner_set_params(struct dvb_frontend* fe,
+					       struct dvb_frontend_parameters *fep)
 {
 	struct flexcop_device *fc = fe->dvb->priv;
 	u8 buf[4];
#83 lore raw

PATCH 1/1 drivers/media/video/pwc/: make code static

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-dvb.patch
...
git trees
...

This patch makes the following needlessly global code static:
- pwc-ctrl.c: pwc_get_leds()
- pwc_preferred_compression

---

 drivers/media/video/pwc/pwc-ctrl.c |    2 +-
 drivers/media/video/pwc/pwc-if.c   |    2 +-
 drivers/media/video/pwc/pwc.h      |    2 --
 3 files changed, 2 insertions(+), 4 deletions(-)
drivers/media/video/pwc/pwc.h renamed from drivers/media/video/pwc/pwc.h.old +0 -2
@@ -274,7 +274,6 @@
 #if CONFIG_PWC_DEBUG
 extern int pwc_trace;
 #endif
-extern int pwc_preferred_compression;
 extern int pwc_mbufs;
 
 /** functions in pwc-if.c */
@@ -308,7 +307,6 @@
 extern int pwc_get_saturation(struct pwc_device *pdev, int *value);
 extern int pwc_set_saturation(struct pwc_device *pdev, int value);
 extern int pwc_set_leds(struct pwc_device *pdev, int on_value, int off_value);
-extern int pwc_get_leds(struct pwc_device *pdev, int *on_value, int *off_value);
 extern int pwc_get_cmos_sensor(struct pwc_device *pdev, int *sensor);
 extern int pwc_restore_user(struct pwc_device *pdev);
 extern int pwc_save_user(struct pwc_device *pdev);
drivers/media/video/pwc/pwc-ctrl.c renamed from drivers/media/video/pwc/pwc-ctrl.c.old +1 -1
@@ -925,7 +925,7 @@
 	return SendControlMsg(SET_STATUS_CTL, LED_FORMATTER, 2);
 }
 
-int pwc_get_leds(struct pwc_device *pdev, int *on_value, int *off_value)
+static int pwc_get_leds(struct pwc_device *pdev, int *on_value, int *off_value)
 {
 	unsigned char buf[2];
 	int ret;
drivers/media/video/pwc/pwc-if.c renamed from drivers/media/video/pwc/pwc-if.c.old +1 -1
@@ -133,7 +133,7 @@
 #endif
 static int power_save = 0;
 static int led_on = 100, led_off = 0; /* defaults to LED that is on while in use */
-       int pwc_preferred_compression = 1; /* 0..3 = uncompressed..high */
+static int pwc_preferred_compression = 1; /* 0..3 = uncompressed..high */
 static struct {
 	int type;
 	char serial_number[30];
#84 lore raw

PATCH 1/1 make variables static after klibc merge

In reply to Andrew Morton
Quoted 5 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.17-rc4/2.6.17-rc4-mm1/

- This tree contains a large number of new bugs^H^H^H^Hpatches.

Quoted 8 lines

- klibc (Kernel libc), as git-klibc.patch (Peter Anvin)
...
Changes since 2.6.17-rc3-mm1:
...
git-klibc.patch
...
git trees
...

We can now make the following variables static:
- drivers/md/md.c: mdp_major
- init/main.c: envp_init[]

---

 drivers/md/md.c |    2 +-
 init/main.c     |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/md/md.c renamed from drivers/md/md.c.old +1 -1
@@ -2563,7 +2563,7 @@
 	.default_attrs	= md_default_attrs,
 };
 
-int mdp_major = 0;
+static int mdp_major = 0;
 
 static struct kobject *md_probe(dev_t dev, int *part, void *data)
 {
init/main.c renamed from init/main.c.old +1 -1
@@ -161,7 +161,7 @@
 __setup("maxcpus=", maxcpus);
 
 static char * argv_init[MAX_INIT_ARGS+2] = { "init", NULL, };
-char * envp_init[MAX_INIT_ENVS+2] = { "HOME=/", "TERM=linux", NULL, };
+static char * envp_init[MAX_INIT_ENVS+2] = { "HOME=/", "TERM=linux", NULL, };
 static const char *panic_later, *panic_param;
 
 extern struct obs_kernel_param __setup_start[], __setup_end[];
#85 lore raw

PATCH 1/1 make drivers/mtd/nand/cs553x_nand.c:cs553x_init() static

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-mtd.patch
...
git trees
...

This patch makes the needlessly global cs553x_init() static.

drivers/mtd/nand/cs553x_nand.c renamed from drivers/mtd/nand/cs553x_nand.c.old +1 -1
@@ -267,7 +267,7 @@
 	return err;
 }
 
-int __init cs553x_init(void)
+static int __init cs553x_init(void)
 {
 	int err = -ENXIO;
 	int i;
#86 lore raw

Re: [-mm patch] make drivers/mtd/nand/cs553x_nand.c:cs553x_init() static

Quoted 1 line · Adrian Bunk

This patch makes the needlessly global cs553x_init() static.

NAK. That problem was inherited from another board driver which was copied and modified. And that driver inherited it in turn... let's not piss about with them one at a time; let's just fix them _all_ at once, instead. And let's remove the bizarre use of '#ifdef MODULE' around the cleanup functions while we're at it.

http://git.infradead.org/?p=mtd-2.6.git;a=commitdiff;h=cead4dbc03ba6eb2e35bac04439b76a0cc2286ce

Signature · 1 line
dwmw2
#87 lore raw

PATCH 1/1 fs/nfs/: make code static

In reply to Andrew Morton

This patch makes needlessly global code static.

---

 fs/nfs/client.c         |    6 ++++--
 fs/nfs/internal.h       |    2 --
 fs/nfs/namespace.c      |   12 +++++++++---
 include/linux/nfs_fs.h  |    4 ----
 include/linux/nfs_xdr.h |    1 -
 5 files changed, 13 insertions(+), 12 deletions(-)
fs/nfs/internal.h renamed from fs/nfs/internal.h.old +0 -2
@@ -38,8 +38,6 @@
 };
 
 /* client.c */
-extern struct rpc_program nfs_program;
-
 extern void nfs_put_client(struct nfs_client *);
 extern struct nfs_client *nfs_find_client(struct sockaddr_in *, int);
 extern struct nfs_server *nfs_create_server(struct nfs_mount_data *,
include/linux/nfs_xdr.h renamed from include/linux/nfs_xdr.h.old +0 -1
@@ -838,6 +838,5 @@
 extern struct rpc_version	nfs_version4;
 
 extern struct rpc_version	nfsacl_version3;
-extern struct rpc_program	nfsacl_program;
 
 #endif
fs/nfs/client.c renamed from fs/nfs/client.c.old +4 -2
@@ -75,7 +75,7 @@
 #endif
 };
 
-struct rpc_program nfs_program = {
+static struct rpc_program nfs_program = {
 	.name			= "nfs",
 	.number			= NFS_PROGRAM,
 	.nrvers			= ARRAY_SIZE(nfs_version),
@@ -90,12 +90,14 @@
 
 
 #ifdef CONFIG_NFS_V3_ACL
+static struct rpc_program	nfsacl_program;
+
 static struct rpc_stat		nfsacl_rpcstat = { &nfsacl_program };
 static struct rpc_version *	nfsacl_version[] = {
 	[3]			= &nfsacl_version3,
 };
 
-struct rpc_program		nfsacl_program = {
+static struct rpc_program	nfsacl_program = {
 	.name			= "nfsacl",
 	.number			= NFS_ACL_PROGRAM,
 	.nrvers			= ARRAY_SIZE(nfsacl_version),
include/linux/nfs_fs.h renamed from include/linux/nfs_fs.h.old +0 -4
@@ -312,10 +312,6 @@
 extern struct nfs_open_context *get_nfs_open_context(struct nfs_open_context *ctx);
 extern void put_nfs_open_context(struct nfs_open_context *ctx);
 extern struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, int mode);
-extern struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent,
-					const struct dentry *dentry,
-					struct nfs_fh *fh,
-					struct nfs_fattr *fattr);
 
 /* linux/net/ipv4/ipconfig.c: trims ip addr off front of name, too. */
 extern u32 root_nfs_parse_addr(char *name); /*__init*/
fs/nfs/namespace.c renamed from fs/nfs/namespace.c.old +9 -3
@@ -26,6 +26,11 @@
 static DECLARE_WORK(nfs_automount_task, nfs_expire_automounts, &nfs_automount_list);
 int nfs_mountpoint_expiry_timeout = 500 * HZ;
 
+static struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent,
+					const struct dentry *dentry,
+					struct nfs_fh *fh,
+					struct nfs_fattr *fattr);
+
 /*
  * nfs_path - reconstruct the path given an arbitrary dentry
  * @base - arbitrary string to prepend to the path
@@ -206,9 +211,10 @@
  * @fattr - attributes for new root inode
  *
  */
-struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent,
-		const struct dentry *dentry, struct nfs_fh *fh,
-		struct nfs_fattr *fattr)
+static struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent,
+					const struct dentry *dentry,
+					struct nfs_fh *fh,
+					struct nfs_fattr *fattr)
 {
 	struct nfs_clone_mount mountdata = {
 		.sb = mnt_parent->mnt_sb,
#88 lore raw

PATCH 1/1 arch/i386/oprofile/: make functions static

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
+x86_64-mm-remove-un-set_nmi_callback-and-reserve-release_lapic_nmi-functions.patch
...
x86_64 tree updates
...

This patch makes the following needlessly global functions static:
- nmi_int.c: profile_exceptions_notify()
- nmi_timer_int.c: profile_timer_exceptions_notify()

---

 arch/i386/oprofile/nmi_int.c       |    4 ++--
 arch/i386/oprofile/nmi_timer_int.c |    4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)
arch/i386/oprofile/nmi_int.c renamed from arch/i386/oprofile/nmi_int.c.old +2 -2
@@ -82,8 +82,8 @@
 #define exit_driverfs() do { } while (0)
 #endif /* CONFIG_PM */
 
-int profile_exceptions_notify(struct notifier_block *self,
-				unsigned long val, void *data)
+static int profile_exceptions_notify(struct notifier_block *self,
+				     unsigned long val, void *data)
 {
 	struct die_args *args = (struct die_args *)data;
 	int ret = NOTIFY_DONE;
arch/i386/oprofile/nmi_timer_int.c renamed from arch/i386/oprofile/nmi_timer_int.c.old +2 -2
@@ -19,8 +19,8 @@
 #include <asm/ptrace.h>
 #include <asm/kdebug.h>
  
-int profile_timer_exceptions_notify(struct notifier_block *self,
-				unsigned long val, void *data)
+static int profile_timer_exceptions_notify(struct notifier_block *self,
+					   unsigned long val, void *data)
 {
 	struct die_args *args = (struct die_args *)data;
 	int ret = NOTIFY_DONE;
#89 lore raw

Re: [-mm patch] arch/i386/oprofile/: make functions static

Quoted 10 lines · Adrian Bunk
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
+x86_64-mm-remove-un-set_nmi_callback-and-reserve-release_lapic_nmi-functions.patch
...
x86_64 tree updates
...

Quoted 3 lines

This patch makes the following needlessly global functions static:
- nmi_int.c: profile_exceptions_notify()
- nmi_timer_int.c: profile_timer_exceptions_notify()

Merged. Thanks

-Andi

Applied
#90 lore raw

Re: [-mm patch] arch/i386/oprofile/: make functions static

In reply to Adrian Bunk
Quoted 10 lines · Adrian Bunk
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
+x86_64-mm-remove-un-set_nmi_callback-and-reserve-release_lapic_nmi-functions.patch
...
x86_64 tree updates
...

Quoted 3 lines

This patch makes the following needlessly global functions static:
- nmi_int.c: profile_exceptions_notify()
- nmi_timer_int.c: profile_timer_exceptions_notify()

Applied.

-Andi

#91 lore raw

PATCH 1/1 fs/ocfs2/dlm/: cleanups

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-ocfs2.patch
...
git trees
...

This patch #if 0's the no longer used dlm_dump_lock_resources().

Since this makes dlmdebug.h empty, this patch also removes this header.

Additionally, the needlessly global dlm_is_node_recovered() is made static.

---

 fs/ocfs2/dlm/dlmdebug.c    |    4 ++--
 fs/ocfs2/dlm/dlmdebug.h    |   30 ------------------------------
 fs/ocfs2/dlm/dlmdomain.c   |    1 -
 fs/ocfs2/dlm/dlmmaster.c   |    1 -
 fs/ocfs2/dlm/dlmrecovery.c |    2 +-
 5 files changed, 3 insertions(+), 35 deletions(-)
fs/ocfs2/dlm/dlmdebug.h deleted +0 -30
@@ -1,30 +0,0 @@
-/* -*- mode: c; c-basic-offset: 8; -*-
- * vim: noexpandtab sw=8 ts=8 sts=0:
- *
- * dlmdebug.h
- *
- * Copyright (C) 2004 Oracle.  All rights reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public
- * License as published by the Free Software Foundation; either
- * version 2 of the License, or (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public
- * License along with this program; if not, write to the
- * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
- * Boston, MA 021110-1307, USA.
- *
- */
-
-#ifndef DLMDEBUG_H
-#define DLMDEBUG_H
-
-void dlm_dump_lock_resources(struct dlm_ctxt *dlm);
-
-#endif
fs/ocfs2/dlm/dlmdebug.c renamed from fs/ocfs2/dlm/dlmdebug.c.old +2 -2
@@ -37,10 +37,8 @@
 
 #include "dlmapi.h"
 #include "dlmcommon.h"
-#include "dlmdebug.h"
 
 #include "dlmdomain.h"
-#include "dlmdebug.h"
 
 #define MLOG_MASK_PREFIX ML_DLM
 #include "cluster/masklog.h"
@@ -120,6 +118,7 @@
 }
 EXPORT_SYMBOL_GPL(dlm_print_one_lock);
 
+#if 0
 void dlm_dump_lock_resources(struct dlm_ctxt *dlm)
 {
 	struct dlm_lock_resource *res;
@@ -142,6 +141,7 @@
 	}
 	spin_unlock(&dlm->spinlock);
 }
+#endif  /*  0  */
 
 static const char *dlm_errnames[] = {
 	[DLM_NORMAL] =			"DLM_NORMAL",
fs/ocfs2/dlm/dlmdomain.c renamed from fs/ocfs2/dlm/dlmdomain.c.old +0 -1
@@ -41,7 +41,6 @@
 #include "dlmapi.h"
 #include "dlmcommon.h"
 
-#include "dlmdebug.h"
 #include "dlmdomain.h"
 
 #include "dlmver.h"
fs/ocfs2/dlm/dlmmaster.c renamed from fs/ocfs2/dlm/dlmmaster.c.old +0 -1
@@ -47,7 +47,6 @@
 
 #include "dlmapi.h"
 #include "dlmcommon.h"
-#include "dlmdebug.h"
 #include "dlmdomain.h"
 
 #define MLOG_MASK_PREFIX (ML_DLM|ML_DLM_MASTER)
fs/ocfs2/dlm/dlmrecovery.c renamed from fs/ocfs2/dlm/dlmrecovery.c.old +1 -1
@@ -357,7 +357,7 @@
 
 /* returns true if node is no longer in the domain
  * could be dead or just not joined */
-int dlm_is_node_recovered(struct dlm_ctxt *dlm, u8 node)
+static int dlm_is_node_recovered(struct dlm_ctxt *dlm, u8 node)
 {
 	int recovered;
 	spin_lock(&dlm->spinlock);
#92 lore raw

Re: [-mm patch] fs/ocfs2/dlm/: cleanups

Hi Adrian,

Quoted 6 lines · Adrian Bunk

This patch #if 0's the no longer used dlm_dump_lock_resources().

Since this makes dlmdebug.h empty, this patch also removes this header.

Additionally, the needlessly global dlm_is_node_recovered() is made static.

Yeah, that all looks good - thanks for the patch!

	--Mark

--
Mark Fasheh
Senior Software Developer, Oracle
mark.fasheh at oracle.com

#93 lore raw

PATCH 1/1 drivers/net/s2io.c: make bus_speed[] static

In reply to Andrew Morton
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-netdev-all.patch
...
git trees
...

This patch makes the needlessly global bus_speed[] static.

drivers/net/s2io.c renamed from drivers/net/s2io.c.old +1 -1
@@ -852,7 +852,7 @@
 	return 0;
 }
 
-int bus_speed[8] = {33, 133, 133, 200, 266, 133, 200, 266};
+static int bus_speed[8] = {33, 133, 133, 200, 266, 133, 200, 266};
 /**
  * s2io_print_pci_mode -
  */
#94 lore raw

Re: [-mm patch] drivers/net/s2io.c: make bus_speed[] static

Hi,

Quoted 1 line · Adrian Bunk

This patch makes the needlessly global bus_speed[] static.

Is there a reason why you don't also constify it while you are at it? Or is it because you want to do a series of static-only patches for now?

Andreas Mohr

#95 lore raw

Re: [-mm patch] drivers/net/s2io.c: make bus_speed[] static

Quoted 7 lines · Andreas Mohr

Hi,

Quoted 1 line · Adrian Bunk

This patch makes the needlessly global bus_speed[] static.

Is there a reason why you don't also constify it while you are at it? Or is it because you want to do a series of static-only patches for now?

The latter.

Feel free to send a patch that also makes it const.

Quoted 1 line

Andreas Mohr

cu
Adrian

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

Re: [-mm patch] drivers/net/s2io.c: make bus_speed[] static

In reply to Adrian Bunk
Quoted 10 lines · Adrian Bunk
Quoted 7 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
git-netdev-all.patch
...
git trees
...

Quoted 3 lines

This patch makes the needlessly global bus_speed[] static.

applied

#97 lore raw

2.6.17-rc4-mm1: please drop add-raw-driver-kconfig-entry-for-s390.patch

In reply to Andrew Morton
Quoted 6 lines · Andrew Morton

...
Changes since 2.6.17-rc3-mm1:
...
+add-raw-driver-kconfig-entry-for-s390.patch

Enable raw driver on s390

Since it seems the NAK's of Christoph and Martin weren't enough:
NAK++

How many more NAK's does it require to prevent offering a deprecated option on additional architectures?

This driver is declared obsolete since more than two years, and while it's worth a discussion how long to keep it for legacy users, merging a patch offering an obsolete driver for even more users is silly.

cu
Adrian

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

Re: 2.6.17-rc4-mm1: please drop add-raw-driver-kconfig-entry-for-s390.patch

Quoted 3 lines · Adrian Bunk

This driver is declared obsolete since more than two years, and while it's worth a discussion how long to keep it for legacy users, merging a patch offering an obsolete driver for even more users is silly.

If it is so deprecated, just drop it for 2.6.17. I mean, no major distro that counts will be based on 2.6.17, so noone will miss it from now on. nails with heads and stuff...

#99 lore raw

PATCH 1/1 the overdue removal of the obsolete raw driver

Quoted 9 lines · Olaf Hering
Quoted 3 lines · Adrian Bunk

This driver is declared obsolete since more than two years, and while it's worth a discussion how long to keep it for legacy users, merging a patch offering an obsolete driver for even more users is silly.

If it is so deprecated, just drop it for 2.6.17. I mean, no major distro that counts will be based on 2.6.17, so noone will miss it from now on. nails with heads and stuff...

Although I fear it will still not be accepted due to legacy users, I've attached it below.

cu
Adrian

<-- snip -->

This patch contains the overdue removal of the obsolete raw driver.

---

This patch was already sent on:
- 19 Jan 2006

 Documentation/feature-removal-schedule.txt |    8 
 drivers/char/Kconfig                       |   20 -
 drivers/char/Makefile                      |    1 
 drivers/char/mem.c                         |    1 
 drivers/char/raw.c                         |  343 ---------------------
 fs/compat_ioctl.c                          |   69 ----
 include/linux/compat_ioctl.h               |    3 
 include/linux/raw.h                        |   18 -
 8 files changed, 463 deletions(-)
Documentation/feature-removal-schedule.txt renamed from Documentation/feature-removal-schedule.txt.old +0 -8
@@ -20,8 +19,0 @@
-What:	RAW driver (CONFIG_RAW_DRIVER)
-When:	December 2005
-Why:	declared obsolete since kernel 2.6.3
-	O_DIRECT can be used instead
-Who:	Adrian Bunk <[email protected]>
-
----------------------------
-
drivers/char/Kconfig renamed from drivers/char/Kconfig.old +0 -20
@@ -939,26 +939,6 @@
 	tristate "NEC VR4100 series General-purpose I/O Unit support"
 	depends on CPU_VR41XX
 
-config RAW_DRIVER
-	tristate "RAW driver (/dev/raw/rawN) (OBSOLETE)"
-	help
-	  The raw driver permits block devices to be bound to /dev/raw/rawN. 
-	  Once bound, I/O against /dev/raw/rawN uses efficient zero-copy I/O. 
-	  See the raw(8) manpage for more details.
-
-          The raw driver is deprecated and will be removed soon.
-          Applications should simply open the device (eg /dev/hda1)
-          with the O_DIRECT flag.
-
-config MAX_RAW_DEVS
-	int "Maximum number of RAW devices to support (1-8192)"
-	depends on RAW_DRIVER
-	default "256"
-	help
-	  The maximum number of RAW devices that are supported.
-	  Default is 256. Increase this number in case you need lots of
-	  raw devices.
-
 config HPET
 	bool "HPET - High Precision Event Timer" if (X86 || IA64)
 	default n
drivers/char/Makefile renamed from drivers/char/Makefile.old +0 -1
@@ -42,3 +42,2 @@
-obj-$(CONFIG_RAW_DRIVER)	+= raw.o
 obj-$(CONFIG_SGI_SNSC)		+= snsc.o snsc_event.o
 obj-$(CONFIG_MMTIMER)		+= mmtimer.o
drivers/char/mem.c renamed from drivers/char/mem.c.old +0 -1
@@ -16,7 +16,6 @@
 #include <linux/mman.h>
 #include <linux/random.h>
 #include <linux/init.h>
-#include <linux/raw.h>
 #include <linux/tty.h>
 #include <linux/capability.h>
 #include <linux/smp_lock.h>
include/linux/compat_ioctl.h renamed from include/linux/compat_ioctl.h.old +0 -3
@@ -568,9 +568,6 @@
 COMPATIBLE_IOCTL(DEVFSDIOC_SET_EVENT_MASK)
 COMPATIBLE_IOCTL(DEVFSDIOC_RELEASE_EVENT_QUEUE)
 COMPATIBLE_IOCTL(DEVFSDIOC_SET_DEBUG_MASK)
-/* Raw devices */
-COMPATIBLE_IOCTL(RAW_SETBIND)
-COMPATIBLE_IOCTL(RAW_GETBIND)
 /* SMB ioctls which do not need any translations */
 COMPATIBLE_IOCTL(SMB_IOC_NEWCONN)
 /* NCP ioctls which do not need any translations */
fs/compat_ioctl.c renamed from fs/compat_ioctl.c.old +0 -69

87 lines not on this page · raw message

include/linux/raw.h deleted +0 -18
@@ -1,18 +0,0 @@
-#ifndef __LINUX_RAW_H
-#define __LINUX_RAW_H
-
-#include <linux/types.h>
-
-#define RAW_SETBIND	_IO( 0xac, 0 )
-#define RAW_GETBIND	_IO( 0xac, 1 )
-
-struct raw_config_request 
-{
-	int	raw_minor;
-	__u64	block_major;
-	__u64	block_minor;
-};
-
-#define MAX_RAW_MINORS CONFIG_MAX_RAW_DEVS
-
-#endif /* __LINUX_RAW_H */
drivers/char/raw.c deleted +0 -343

343 lines not on this page · raw message

#100 lore raw

Re: [2.6 patch] the overdue removal of the obsolete raw driver

Quoted 1 line · Adrian Bunk

-When: December 2005

Dude, you really screwed it up. ;-)

#101 lore raw

Re: 2.6.17-rc4-mm1: please drop add-raw-driver-kconfig-entry-for-s390.patch

In reply to Adrian Bunk
Quoted 2 lines · Adrian Bunk

Since it seems the NAK's of Christoph and Martin weren't enough:
NAK++

Then lets add an ACK++ as well

Quoted 3 lines

This driver is declared obsolete since more than two years, and while it's worth a discussion how long to keep it for legacy users, merging a patch offering an obsolete driver for even more users is silly.

The real world expects a raw driver. Application authors also expect the same functionality on all platforms.

Also in truth you can NACK and "obsolete" the raw devices all you like, nobody in the distribution world will drop it because it is part of "standard" unix/linux OS functionality and has been for longer than Linux even existed. If you want to fork Linux between "Linux the real world", and "Linux the abstract unused by anyone perfection" that's a fine way to ensure such a split happens.

You know the O_DIRECT stuff is better, I know the O_DIRECT stuff is better, but so was betamax and so is a Dvorak keyboard...

I happen to like "Linux the real world OS", if I wanted to run something more abstractly elegant then I'd download Plan 9.

ACK

Alan