2005-05-25 20:56:11

by Andrew Morton

[permalink] [raw]
Subject: 2.6.12-rc5-mm1


ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/

- Added Oracle's clustering filesystem driver, via git-ocfs.

OCFS2, a shared disk cluster file system. See
Documentation/filesystems/ocfs2.txt Additionally a users guide is
available at:
http://oss.oracle.com/projects/ocfs2-tools/dist/documentation/users_guide.txt

- New Xtensa architecture: Tensilica Xtensa CPU series.

- Added the Red Hat distributed lock manager for people to look at.

- Various new git trees. The remaining holdouts are:

bk-acpi.patch
bk-drm.patch
bk-drm-via.patch
bk-input.patch
bk-nfs.patch
bk-watchdog.patch

- The CPU scheduler is probably busted on the less-common architectures.
For now, those architectures will need to emulate
sched-remove-set_tsk_need_resched-from-init_idle-v2-ia64-fix.patch

- Added the s390 team's execute-in-place driver. Not sure that I like all
the code duplication in this though.

- Again, if there are patches in here which you think should be merged in
2.6.12, please point them out to me.

- CPU scheduler udpates, kexec/kdump updates, i2o updates, v4l updates, etc.



Changes since 2.6.12-rc4-mm2:


-fix-for-bttv-driver-v0915-for-leadtek-winfast-vc100-xp-capture-cards.patch
-serio-resume-fix.patch
-alps-printk-tidy.patch
-alps-resume-fix.patch
-serport-oops-fix.patch
-serio-id-attributes.patch
-fix-impossible-vmallocchunk.patch
-ide-proc-destroy-error.patch
-6300esb-tco-timer-support.patch
-uml-remove-elfh.patch
-uml-critical-change-memcpy-to-memmove.patch
-md-fix-splitting-of-md-linear-request-that-cross-a-device-boundary.patch
-md-set-the-unplug_fn-and-issue_flush_fn-for-md-devices-after-committed-to-creation.patch
-mm-fix-rss-counter-being-incremented-when-unmapping.patch
-mm-acct-accounting-fix.patch
-linux-kernel-elf-core-dump-privilege-elevation.patch
-x86_64-reduce-nmi-watchdog-stack-usage.patch
-x86_64-readd-missing-tests-in-entrys.patch
-x86_64-add-a-guard-page-at-the-end-of-the-47bit-address.patch
-x86_64-fix-defaults-for-physical-core-id-in.patch
-x86_64-increase-number-of-io-apics.patch
-x86_64-dont-look-up-struct-page-pointer-of-physical.patch
-x86_64-update-tsc-sync-algorithm.patch
-x86_64-remove-x86_apicid-field.patch
-x86_64-dont-print-the-internal-k8c-flag-in.patch
-x86_64-remove-unique-apic-io-apic-id-check.patch
-x86_64-add-pmtimer-support.patch
-x86_64-check-if-ptrace-rip-is-canonical.patch
-x86_64-fix-canonical-checking-for-segment-registers-in.patch
-x86_64-when-checking-vmalloc-mappings-dont-use.patch
-x86_64-fix-oem-hpet-check.patch
-x86_64-make-vsyscallc-compile-without-config_sysctl.patch
-x86_64-collected-nmi-watchdog-fixes.patch
-x86_64-collected-nmi-watchdog-fixes-warning-fix.patch
-x86_64-dont-assume-bsp-has-id-0-in-new-smp-bootup.patch
-x86_64-update-defconfig.patch
-mm-nommuc-try-to-fix-__vmalloc.patch
-drivers-input-keyboard-atkbdc-fix-off-by-one-errors.patch
-s390-dasd-set-online-failure.patch
-swapout-oops-fix.patch
-packet-driver-ioctl-fix.patch
-packet-driver-ioctl-fix-fix.patch
-crypto-fix-null-encryption-compression.patch
-cdrw-dvd-packet-writing-data-corruption-fix.patch
-spurious-interrupt-fix.patch
-libata-flush-comreset-set-and-clear.patch
-add-scsi-changer-driver.patch
-add-scsi-changer-driver-gregkh-driver-fix.patch
-block_read_full_page-get_block-error-fix.patch
-do_swap_page-can-map-random-data-if-swap-read-fails.patch
-wireless-3crwe154g72-kconfig-help-fix.patch
-smc91c92_cs-reduce-stack-usage-in-smc91c92_event.patch
-typo-in-tulip-driver.patch
-a-new-10gb-ethernet-driver-by-chelsio-communications.patch
-selinux-fix-avc_alloc_node-oom-with-no-policy-loaded.patch
-mips-add-resource-management-to-pmu.patch
-alpha-osf_sys-use-helper-functions-to-convert-between-tv-and-jiffies.patch
-sysfs-for-ipmi-for-new-mm-kernels.patch
-fix-pci-mmap-on-ppc-and-ppc64.patch
-pcmcia-ds-handle-any-error-code.patch
-kill-asm-ioctl32h.patch
-profilec-schedule-parsing-fix.patch
-ieee1394-feature-removal-notices.patch
-drivers-ieee1394-pcilynxc-remove-dead-options.patch
-drivers-ieee1394-ieee1394_transactionsc-possible-cleanups.patch
-ieee1394-remove-null-checks-prior-to-kfree-in-ieee1394-kfree-handles-null-pointers-fin.patch
-drivers-ieee1394-pcilynxc-use-the-dma_32bit_mask-constant.patch
-ieee1394-single-buffer-fixes-to-video1394.patch
-ieee1394-fix-cross_bound-check-for-null-iso-packets.patch
-ieee1394-fix-premature-expiry-of-async-packets.patch
-dvb-b2c2-flexcop-driver-refactoring-part-1-drop-old-b2c2-usb-stuff.patch
-dvb-b2c2-flexcop-driver-refactoring-part-2-add-modular-flexcop-driver.patch
-dvb-flexcop-fix-usb-transfer-handling.patch
-dvb-flexcop-add-acknowledgements.patch
-dvb-flexcop-fix-mac-address-reading.patch
-dvb-flexcop-fixed-interrupt-sharing.patch
-dvb-flexcop-use-hw-pid-filter.patch
-dvb-flexcop-fix-module-refcount-handling.patch
-dvb-flexcop-readme-update.patch
-dvb-flexcop-i2c-read-fixes.patch
-dvb-flexcop-diseqc-fix.patch
-dvb-support-for-tt-hauppauge-nexus-s-rev-23.patch
-dvb-saa7146-no-need-to-initialize-static-global-variables-to-0.patch
-dvb-dvb_frontend-fix-module-param.patch
-dvb-av7110-audio-out-fix.patch
-dvb-add-support-for-knc-1-cards.patch
-dvb-remove-unnecessary-casts-in-dvb-core.patch
-dvb-dvb_net-handle-ipv6-and-llc-snap.patch
-dvb-av7110-fix-video_set_display_format.patch
-dvb-av7110-fix-ntsc-pal-switching.patch
-dvb-av7110-fix-comment.patch
-dvb-av7110-fix-indentation.patch
-dvb-nxt6000-support-frontend-status-reads.patch
-dvb-tda1004x-formatting-cleanups.patch
-dvb-stv0299-fix-fe_dishnetwork_send_legacy_cmd.patch
-dvb-remove-unnecessary-casts-in-frontends.patch
-dvb-dib3000-add-null-pointer-check.patch
-dvb-ves1820-remove-unnecessary-msleep.patch
-dvb-mt352-embed-struct-mt352_config-in-mt352_state.patch
-dvb-tda1004x-dont-use-bitfields.patch
-dvb-tda1004x-allow-n_i2c-to-be-overridden-by-the-card-driver.patch
-dvb-tda10046-support-for-different-firmware-versions.patch
-dvb-dvb-pllh-prevent-multiple-inclusion.patch
-dvb-make-needlessly-global-code-static-or-drop-it.patch
-dvb-frontends-misc-minor-cleanups.patch
-dvb-modified-dvb_register_adapter-to-avoid-kmalloc-kfree.patch
-dvb-bt8xx-update-documentation.patch
-dvb-dst-reorganize-twinhan-dst-driver-to-support-ci.patch
-dvb-dst-add-support-for-twinhan-200103a.patch
-dvb-dst-fixed-tuning-problem.patch
-dvb-dst-fix-for-descrambling-failure.patch
-dvb-dst-misc-fixes.patch
-dvb-bt8xx-updated-documentation.patch
-dvb-dst-fix-a-bug-in-the-module-parameter.patch
-dvb-dst-fixed-ci-debug-output.patch
-dvb-bt8xx-whitespace-cleanup.patch
-dvb-budget-av-ci-fixes.patch
-fusion-kfree-cleanup.patch

Merged

+linus.patch

Linus latest

+v4l-bttv-i2c-oops-fix.patch
+ipmi-build-fix.patch
+x86_64-config_bug=n-fixes.patch
+ib-allow-null-sa_query-callbacks.patch
+ib-fix-potential-ib_umad-leak.patch
+ib-fix-endianness-of-path-record-mtu-field.patch
+make-sure-therm_adt746x-only-handles-known-hardware.patch
+therm_adt746x-show-correct-sensor-locations.patch

Fixes for 2.6.12

+namei-fixes-01-19.patch
+namei-fixes-02-19.patch
+namei-fixes-03-19.patch
+namei-fixes-04-19.patch
+namei-fixes-05-19.patch
+namei-fixes-06-19.patch
+namei-fixes-07-19.patch
+namei-fixes-08-19.patch
+namei-fixes-09-19.patch
+namei-fixes-10-19.patch
+namei-fixes-11-19.patch
+namei-fixes-12-19.patch
+namei-fixes-13-19.patch
+namei-fixes-14-19.patch
+namei-fixes-15-19.patch
+namei-fixes-16-19.patch
+namei-fixes-17-19.patch
+namei-fixes-18-19.patch
+namei-fixes-19-19.patch

VFS name handling fixes

+ipmi-and-acpi=offht-acpi-get-firmware-failurepatch.patch
+ipmi-and-acpi=offht-ipmi_si_intf-acpi-disabled.patch

IPMI/ACPI fixes

-cpufreq-CPUFREQ-14-powernow-k8-dual-core-on2.6.12.patch
+cpufreq-CPUFREQ-14-powernow-k8-dualcore.patch
-cpufreq-CPUFREQ-20-powernow-k8-static-cpu_sharedcore_mask.patch
+cpufreq-CPUFREQ-21-ondemand-cleanups.patch
+cpufreq-CPUFREQ-22-ondemand-store-idle-ticks-for-all-cpus.patch
+cpufreq-CPUFREQ-23-ondemand-idle_tick-cleanup.patch
+cpufreq-CPUFREQ-24-ondemand-automatic-downscaling.patch
+cpufreq-CPUFREQ-25-ondemand-default-sampling-downfactor.patch
+cpufreq-CPUFREQ-26-longhaul-disable-mastering.patch
+cpufreq-CPUFREQ-27-longhaul-magic-port-frobbing.patch
+cpufreq-CPUFREQ-28-longhaul-transition-latency.patch
+cpufreq-CPUFREQ-29-longhaul-icky-evil-nasty-ide-dma-wait.patch
+cpufreq-CPUFREQ-30-speedstep-lib-typos.patch

cpufreq updates

-bk-cryptodev.patch
+git-cryptodev.patch

The cryptodev tree is now in git

-gregkh-driver-driver-pm-diag-update.patch
-gregkh-driver-driver-remove-detach_state.patch
-gregkh-driver-attr_void.patch
+gregkh-driver-driver-model-documentation-update.patch
+gregkh-driver-libfs-add-simple-attribute-files.patch
+gregkh-driver-driver-fix-error-handling-in-bus_add_device.patch
+gregkh-driver-driver-device_attr-01.patch
+gregkh-driver-driver-device_attr-02.patch
+gregkh-driver-driver-device_attr-03.patch
+gregkh-driver-driver-device_attr-04.patch
+gregkh-driver-driver-device_attr-05.patch
+gregkh-driver-driver-device_attr-06.patch
+gregkh-driver-driver-device_attr-07.patch
+gregkh-driver-driver-device_attr-08.patch
+gregkh-driver-driver-device_attr-09.patch
+gregkh-driver-driver-device_attr-10.patch
+gregkh-driver-driver-device_attr-11.patch
+gregkh-driver-driver-device_attr-12.patch
+gregkh-driver-driver-device_attr-i2c-sysfs.h.patch
+gregkh-driver-driver-device_attr-i2c-adm1026.patch

driver core tree updates

+ipmi-class_simple-fixes.patch

Fix ipmi driver for driver core updates

-gregkh-i2c-w1-ds18xx_sensors.patch
-gregkh-i2c-w1-new_rom_family.patch
+gregkh-i2c-i2c-adm9240-cleanup.patch
+gregkh-i2c-i2c-jiffies.h.patch
+gregkh-i2c-i2c-macro-abuse-cleanup.patch
+gregkh-i2c-i2c-via686a-code-cleanup.patch
+gregkh-i2c-i2c-adm1021-remove_die_code.patch
+gregkh-i2c-i2c-Kconfig-corrections.patch
+gregkh-i2c-i2c-macro-abuse-cleanup-via686a.patch
+gregkh-i2c-i2c-driver-device_attr-fixup.patch
+gregkh-i2c-i2c-spelling-fixes-more-01.patch
+gregkh-i2c-i2c-spelling-fixes-more-02.patch
+gregkh-i2c-i2c-spelling-fixes-more-03.patch
+gregkh-i2c-i2c-spelling-fixes-more-04.patch
+gregkh-i2c-i2c-mpc-race-fix.patch
+gregkh-i2c-i2c-mailing-list-move.patch
+gregkh-i2c-w1-ds18xx_sensors.patch
+gregkh-i2c-w1-new_rom_family.patch
+gregkh-i2c-w1-cleanups.patch
+gregkh-i2c-w1-new-family-structure.patch
+gregkh-i2c-w1-build-fixups.patch
+gregkh-i2c-w1-remove-dup-family-id.patch

i2c updates

-git-libata.patch

Empty.

-bk-mtd.patch
+git-mtd.patch

The MTD tree is now in git

+git-netdev-chelsio.patch
+git-netdev-ieee80211.patch
+git-netdev-wifi.patch

More of Jeff's net device trees

+git-ocfs.patch

Oracle cluster filesystem

-gregkh-pci-pci-hotplug-shpc-power-fix.patch
-gregkh-pci-pci-pciehp-downstream-port-fix.patch
-gregkh-pci-pci-cpci-update.patch
-gregkh-pci-pci-remove-pci_visit_dev.patch
-gregkh-pci-pci-modalias-sysfs.patch
-gregkh-pci-pci-modalias-hotplug.patch
+gregkh-pci-pci-fix-pci-mmap-on-ppc-and-ppc64.patch
+gregkh-pci-pci-driver-device_attr-fixup.patch

PCI tree updates

+git-scsi-misc-build-fix.patch
+git-scsi-misc-sbp2-warning-fix.patch

Fixes for git-scsi-misc.patch

+gregkh-usb-speedtch-prep.patch

Makes Greg's USB tree apply

-gregkh-usb-usb-usbnet-fixes.patch
-gregkh-usb-usb-ehci-suspend-stop-timer.patch
-gregkh-usb-usb-modalias-sysfs.patch
-gregkh-usb-usb-cypress_m8-add-lt-20-support.patch
+gregkh-usb-usb-ftdi_sio-new-id.patch
+gregkh-usb-usb-serial-generic-init-fix.patch
+gregkh-usb-usb-fix-gadget-build-error.patch
+gregkh-usb-usb-driver-device_attr-fixup.patch
+gregkh-usb-usb-storage-trumpion.patch
+gregkh-usb-usb-modalias-shrink.patch

USB tree updates

+usb-option-card-driver.patch

New USB driver

+zd1201-build-fix.patch

Bring back this USB-vs-netdev fix

+ppc64-sparsemem-memory-model-fix-2.patch
+remove-direct-ref-to-contig_page_data-for-x86-64.patch
+add-x86-64-kconfig-options-for-sparsemem.patch
+reorganize-x86-64-numa-and-discontigmem-config-options.patch
+add-x86-64-specific-support-for-sparsemem.patch
+add-x86-64-specific-support-for-sparsemem-tidy.patch

More sparsemem updates

+avoiding-mmap-fragmentation-fix.patch
+avoiding-mmap-fragmentation-revert-unneeded-64-bit-changes.patch
+avoiding-mmap-fragmentation-fix-2.patch
+mmap-topdown-fix-for-large-stack-limit-large-allocation.patch
+mm-remove-pg_highmem.patch
+mm-remove-pg_highmem-tidy.patch
+vm-try_to_free_pages-unused-argument.patch

Various MM fixes

+ppp_mppe-add-ppp-mppe-encryption-module-kconfig-fix.patch

Fix ppp_mppe-add-ppp-mppe-encryption-module.patch

+use-pci_set_dma_mask-instead-of-direct-assignment-of-dma-mask.patch
+cs89x0c-support-for-philips-pnx0105-network-adapter.patch
+cs89x0c-support-for-philips-pnx0105-network-adapter-tidy.patch

Net driver fixes

+ppc32-fix-some-minor-issues-related-to-fsl-book-e-kgdb.patch
+ppc32-fix-alsa-powermac-driver-on-old-machines.patch
+ppc32-add-via-ide-support-to-mpc8555-cds-platform.patch
+ppc32-support-for-82xx-pqii-on-chip-pci-bridge.patch

ppc32 updates

+ppc64-quieten-rtas-printks.patch

ppc64 fix

-x86-port-lockless-mce-preparation.patch
-x86-port-lockless-mce-implementation.patch
-x86-port-lockless-mce-implementation-fix.patch
-x86-port-lockless-mce-implementation-fix-2.patch

Dropped

+m32r-build-fix-for-asm-m32r-topologyh.patch

m32r fix

+ppc64-pcibus_to_node-fix.patch
+fix-pcibus_to_node-for-x86_64.patch

Fixes for pcibus_to_node patch in -mm

+i386-selectable-frequency-of-the-timer-interrupt.patch
+ia64-selectable-timer-interrupt-frequency.patch
+i386-collect-host-bridge-resources.patch
+x86-avoid-wasting-irqs-for-pci-devices.patch
+via-82c586b-irq-routing-fix.patch
+x86-include-asm-uaccessh-in-asm-checksumh.patch
+x86-remove-i386_ksymsc-almost.patch

x86 updates

+x86_64-avoid-wasting-irqs.patch
+x86_64-collect-host-bridge-resources.patch

x86_64 updates

+ioc4-core-driver-rewrite.patch
+ioc4-config-split.patch
+ioc4-pci-bus-speed-detection.patch

IOC driver rewrite

-suspend-resume-smp-support-fix-3.patch

Dropped

-iounmap-debugging.patch

This broke - dropped

+turn-soft-lock-off-when-panicking.patch

Quash the softlockup detector after a panic

-rt-lsm.patch

Dropped

+cfq-cfq-elevator_insert_back-fix.patch
+cfq-cfq_io_context-leak-fix.patch
+cfq-remove-serveral-unused-fields-from-cfq-data-structures.patch

CFQ fixes

+timers-introduce-try_to_del_timer_sync.patch
+posix-timers-use-try_to_del_timer_sync.patch

posx timers cleanups

+kprobes-function-return-probes-fix-5.patch
+move-kprobe-arming-into-arch-specific-code.patch
+kprobes-moves-lock-unlock-to-non-arch-kprobe_flush_task.patch
+kprobes-ia64-kdebug-die-notification.patch
+kprobes-ia64-kdebug-die-notification-fix.patch
+kprobes-ia64-arch-specific-handling-of-kprobes.patch
+kprobes-ia64-arch-specific-handling-of-kprobes-fix.patch
+kprobes-ia64-architecture-specific-support.patch
+kprobes-ia64-support-kprobe-on-branch-call-instructions.patch
+kprobes-temporary-disarming-of-reentrant-probe.patch
+kprobes-temporary-disarming-of-reentrant-probe-for-i386.patch
+kprobes-temporary-disarming-of-reentrant-probe-for-x86_64.patch
+kprobes-temporary-disarming-of-reentrant-probe-for-ppc64.patch
+kprobes-temporary-disarming-of-reentrant-probe-for-sparc64.patch

Much kprobes work

+remove-eventpoll-macro-obfuscation.patch

epoll cleanup

+optimize-sys_times-for-a-single-thread-process.patch
+optimize-sys_times-for-a-single-thread-process-update.patch
+optimize-sys_times-for-a-single-thread-process-update-2.patch

sys_times() hack^wspeedup

+turn-off-sibling-call-optimization-w-frame-pointers.patch

Improve debuggability

+add-skip_hangcheck_timer.patch

Add a way of whutting up the hangcheck timer

+ipcsem-remove-superflous-decrease-variable-from-sys_semtimedop.patch

IPC cleanup

+reiserfs-add-checking-of-journal_begin-return-value.patch
+quota-improve-credits-estimates.patch
+quota-ext3-improve-quota-credit-estimates.patch
+quota-reiserfs-improve-quota-credit-estimates.patch

quota fixes

+namespacec-fix-bind-mount-from-foreign-namespace.patch
+namespacec-fix-mnt_namespace-clearing.patch
+namespacec-fix-race-in-mark_mounts_for_expiry.patch
+namespacec-cleanup-in-mark_mounts_for_expiry.patch
+namespacec-split-mark_mounts_for_expiry.patch
+namespacec-fix-expiring-of-detached-mount.patch

namespace.c fixes

+xtensa-tensilica-xtensa-cpu-arch-maintainer-record.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-1.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-2.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-3.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-4.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-5.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-6.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-7.patch
+xtensa-architecture-support-for-tensilica-xtensa-part-8.patch

New architecture

+make-reiserfs-bug-on-too-big-transaction.patch

reiserfs sanity check

+ipmi-doc-updates-for-ipmi.patch
+ipmi-ipmi-timer-shutdown-cleanup.patch
+ipmi-add-ipmi-power-cycle-capability.patch
+ipmi-use-completions-not-semaphores-in-the-ipmi-powerdown-code.patch
+ipmi-add-32-bit-ioctl-translations-for-64-bit-platforms.patch

IPMI driver updates

+tpm-replace-odd-LPC-init-function.patch

TPM driver update

+dlm-core-locking.patch
+dlm-lockspaces-callbacks-directory.patch
+dlm-communication.patch
+dlm-recovery.patch
+dlm-configuration.patch
+dlm-device-interface.patch
+dlm-device-interface-fix.patch
+dlm-debug-fs.patch
+dlm-build.patch

RH distrubuted lock manager

+connector-netlink-id-fix.patch
+connector-remove-socket-number-parameter.patch
+fork-connector-send-status-to-userspace.patch

Connector driver updates

+inotify-44-update-2.patch

inotify fixes

+i2o-bugfixes-and-compability-enhancements.patch
+i2o-first-code-cleanup-of-spare-warnings-and-unused.patch
+i2o-new-sysfs-attributes-and-adaptec-specific-block.patch
+i2o-new-sysfs-attributes-and-adaptec-specific-block-fix.patch
+i2o-adaptec-specific-sg_io-access-firmware-access-through.patch
+i2o-second-code-cleanup-of-sparse-warnings-and-unneeded.patch
+i2o-lindent-run-and-replacement-of-printk-through-osm.patch
+i2o-limit-max-sector-workaround-for-promise-controllers.patch

i2o driver updates

+drop-obsolete-dibusb-driver.patch
+add-generalized-dvb-usb-driver.patch
+dvb-usb-fix-init-error-checking.patch
+dvb_frontend-use-time_after.patch
+flexcop-add-bcm3510-atsc-frontend-support-for-air2pc-card.patch
+flexcop-add-bcm3510-atsc-frontend-support-for-air2pc-card-fix.patch

DVB updates

+pcmcia-add-a-few-more-ids-for-pcnet_cs.patch

More pcmcia device IDs

+pcmcia-move-pcmcia-ioctl-to-a-separate-file-fix.patch

Fix pcmcia-move-pcmcia-ioctl-to-a-separate-file.patch

+pcmcia-properly-handle-all-errors-of-register_chrdev.patch

pcmcia error handling fix

-numa-aware-slab-allocator-v3.patch
-numa-aware-slab-allocator-v2-tidy.patch
-numa-aware-slab-allocator-v3-cleanup.patch
-ppc64-numa-nodes-hack.patch

Dropped

+perfctr-ppc64-wraparound-fixes.patch
+perfctr-x86-update-with-k8-multicore-fixes-take-2.patch

perfctr fixes

+sched-micro-optimize-task-requeueing-in-schedule.patch
+dynamic-sched-domains-sched-changes.patch
+dynamic-sched-domains-sched-changes-fix.patch
+dynamic-sched-domains-cpuset-changes.patch
+dynamic-sched-domains-ia64-changes.patch
+sched-implement-nice-support-across-physical-cpus-on-smp.patch
+sched-change_prio_bias_only_if_queued.patch
+sched-account_rt_tasks_in_prio_bias.patch
+consolidate-preempt-options-into-kernel-kconfigpreempt.patch
+sched-remove-set_tsk_need_resched-from-init_idle-v2.patch
+sched-remove-set_tsk_need_resched-from-init_idle-v2-ia64-fix.patch
+sched-voluntary-kernel-preemption.patch

CPU scheduler updates

+bttv-support-for-adlink-rtv24-capture-card.patch
+bttv-support-for-adlink-rtv24-capture-card-tidy.patch
+bttv-support-for-adlink-rtv24-capture-card-more-tidy.patch
+v4l-saa7134-ntsc-vbi-fix.patch
+v4l-pal-m-chroma-subcarrier-frequency-fix.patch

v4l updates

+kexec-kexec-on-panic-fix-with-nmi-watchdog-enabled.patch
+kdump-documentation-update-to-add-gdb-macros.patch
+kdump-use-real-pt_regs-from-exception.patch
+kdump-use-real-pt_regs-from-exception-fix.patch
+kdump-use-real-pt_regs-from-exception-fix-fix.patch
+kdump-save-trap-information-for-later-analysis.patch

keec/kdump updates and fixes

+reiser4-mm-remove-pg_highmem-fix.patch

Fix reiser4 for the PG_highmem removal

+vga-to-fbcon-fix.patch

vgacon fix

+docbook-update-comments.patch

kerneldoc fixlets

+fuse-dont-allow-restarting-of-system-calls.patch

FUSE update

+xip-bdev-execute-in-place-3rd-version.patch
+xip-fs-mm-execute-in-place-3rd-version.patch
+xip-fs-mm-execute-in-place-3rd-version-fix.patch
+xip-ext2-execute-in-place-3rd-version.patch
+xip-ext2-execute-in-place-3rd-version-fixes.patch
+xip-madvice-fadvice-execute-in-place-3rd-version.patch
+xip-description.patch

Execute-in-place driver



number of patches in -mm: 1270
number of changesets in external trees: 155
number of patches in -mm only: 1264
total patches: 1419



All 1270 patches:

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/patch-list



2005-05-25 21:52:14

by Brice Goglin

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Andrew Morton a ?crit :
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/

Hi Andrew,

It looks like dlm assumes that CONFIG_DLM_DEBUG is set.
The attached patch fixes this.

Regards,
Brice


Attachments:
fix-dlm-without-debug.patch (476.00 B)

2005-05-25 21:58:15

by Brice Goglin

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Andrew Morton a ?crit :
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/

Hi Andrew,

drivers/pcmcia/ds.c defines pcmcia_store_allow_func_id_match
without the new "struct device_attribute *attr" argument.
The attached patch fixes this.

Signed-off-by: Brice Goglin <[email protected]>

Regards,
Brice


Attachments:
fix-pcmcia-ds-device-attribute.patch (577.00 B)

2005-05-25 22:37:22

by Alexandre Buisse

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/

Hi Andrew,

missing extern in drivers/dlm/lvb_table.h.
The definition was moved into drivers/dlm/lock.c.
The attached patch fixes this.


Signed-off-by: Alexandre Buisse <[email protected]>

Regards,
Alexandre


Attachments:
fix-dlm-extern-lvb_table.patch (2.26 kB)

2005-05-25 22:52:33

by Jesper Juhl

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Wed, 25 May 2005, Andrew Morton wrote:

>
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/
>
[...]
>
> - Again, if there are patches in here which you think should be merged in
> 2.6.12, please point them out to me.
>

I'd say the following patches of mine might are candidates :

atm-nicstar-remove-a-bunch-of-pointless-casts-of-null.patch
cosmetic-fixes-for-example-programs-in-documentation-cdrom-sbpcd.patch
dont-do-pointless-null-checks-and-casts-before-kfree.patch
get-rid-of-redundant-null-checks-before-kfree-in-arch-i386.patch
kfree-cleanups-for-drivers-firmware.patch
kfree-cleanups-in-ixjc.patch
remove-pointless-null-check-before-kfree-in-sony535c.patch
remove-redundant-null-check-before-before-kfree-in.patch
remove-redundant-null-checks-before-kfree-in-sound-and.patch
streamline-preempt_count-type-across-archs.patch
preempt_count-is-int-remove-cast-and-dont-assign-to.patch

They are all quite simple and have not been the cause of any trouble in
-mm, so we might as well get them merged.


--
Jesper Juhl


2005-05-26 01:17:55

by Matthew Dobson

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

I get the following when trying to build rc5-mm1:

arch/i386/kernel/setup.c:99: error: parse error before "acpi_sci_flags"
arch/i386/kernel/setup.c:99: warning: type defaults to `int' in declaration
of `acpi_sci_flags'
arch/i386/kernel/setup.c:99: warning: data definition has no type or
storage class
arch/i386/kernel/setup.c: In function `parse_cmdline_early':
arch/i386/kernel/setup.c:836: error: request for member `trigger' in
something not a structure or union
arch/i386/kernel/setup.c:839: error: request for member `trigger' in
something not a structure or union
arch/i386/kernel/setup.c:842: error: request for member `polarity' in
something not a structure or union
arch/i386/kernel/setup.c:845: error: request for member `polarity' in
something not a structure or union
make[1]: *** [arch/i386/kernel/setup.o] Error 1
make[1]: *** Waiting for unfinished jobs....

Looks like when #ifdef CONFIG_ACPI was added to include/linux/acpi.h it
broke things b/c acpi_sci_flags is needed for CONFIG_ACPI_BOOT, which
depends on X86_HT || ACPI. Thus, you can have ACPI=n, and X86_HT=y &
ACPI_BOOT=y and it won't build. This patch at least gets it to build for me...

-Matt


Attachments:
fix_acpi_breakage.patch (679.00 B)

2005-05-26 02:43:28

by Ed Tomlinson

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Wednesday 25 May 2005 16:49, you wrote:
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/
>
> - Added Oracle's clustering filesystem driver, via git-ocfs.
>
> ? ? OCFS2, a shared disk cluster file system. ?See
> ? ? Documentation/filesystems/ocfs2.txt Additionally a users guide is
> ? ? available at:
> ? ? http://oss.oracle.com/projects/ocfs2-tools/dist/documentation/users_guide.txt
>
> - New Xtensa architecture: Tensilica Xtensa CPU series.
>
> - Added the Red Hat distributed lock manager for people to look at.
>
> - Various new git trees. ?The remaining holdouts are:

Got the following when I tried to use sound. Anyone else see problems in alsa land?

May 25 21:17:38 grover kernel: [ 131.322558] PGD 2ab13067 PUD 2ab1b067 PMD 2a3f1067 PTE 0
May 25 21:17:38 grover kernel: [ 131.351464] CPU 0
May 25 21:17:38 grover kernel: [ 131.358041] Modules linked in: radeon nfsd exportfs lockd sunrpc sg parport_pc lp parport ipv6 sd_
mod evdev mousedev tsdev usbhid usb_storage snd_intel8x0 snd_ac97_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore sn
d_page_alloc ehci_hcd ohci_hcd eth1394 sata_nv libata forcedeth
ohci1394 powernow_k8 freq_table processor cpufreq_userspace w83627hf eeprom i2c_sensor i2c_isa i2c_nforce2 i2c_core sr_mod sbp2 scs
i_mod ieee1394 rtc unix
May 25 21:17:38 grover kernel: [ 131.487598] Pid: 5481, comm: artsd Not tainted 2.6.12-rc5-mm1
May 25 21:17:38 grover kernel: [ 131.506496] RIP: 0010:[__nosave_end+129759479/2131247104] <ffffffff8813b8f7>{:snd_pcm:snd_pcm_mmap
_data_close+7}
May 25 21:17:38 grover kernel: [ 131.535142] RSP: 0018:ffff81002ab9dee0 EFLAGS: 00010286
May 25 21:17:38 grover kernel: [ 131.553183] RAX: 00002aaaadb59000 RBX: ffff810029c64988 RCX: fffffffffffffff2
May 25 21:17:38 grover kernel: [ 131.576657] RDX: ffff810029c64988 RSI: ffff81002d672ce0 RDI: ffff81002b339f50
May 25 21:17:38 grover kernel: [ 131.600129] RBP: ffff81002b3051c0 R08: ffff810029c649a8 R09: ffff81002ab9dec8
May 25 21:17:38 grover kernel: [ 131.623602] R10: 0000000000000002 R11: 0000000000000206 R12: ffff81002b339f50
May 25 21:17:38 grover kernel: [ 131.647075] R13: ffff81002b339f50 R14: ffff81002e7f2a08 R15: 00002aaaadb67000
May 25 21:17:38 grover kernel: [ 131.670548] FS: 00002aaaad6d8f50(0000) GS:ffffffff80550840(0000) knlGS:0000000000000000
May 25 21:17:38 grover kernel: [ 131.697165] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
May 25 21:17:38 grover kernel: [ 131.716064] CR2: 00002aaaadb59098 CR3: 000000002aafd000 CR4: 00000000000006e0
May 25 21:17:38 grover kernel: [ 131.739537] Process artsd (pid: 5481, threadinfo ffff81002ab9c000, task ffff81002ada97f0)
May 25 21:17:38 grover kernel: [ 131.766439] Stack: ffffffff8016942d 0000000000000000 0000000000000000 ffff81002e7f2a00
May 25 21:17:38 grover kernel: [ 131.792174] ffffffff8016a936 ffff81002d672df0 ffff81002d672e08 ffff81002d672df0
May 25 21:17:38 grover kernel: [ 131.818481] ffff81002e7f2a00 0000000000560600
May 25 21:17:38 grover kernel: [ 131.835065] Call Trace:<ffffffff8016942d>{remove_vm_struct+125} <ffffffff8016a936>{do_munmap+550}
May 25 21:17:38 grover kernel: [ 131.864282] <ffffffff8016b0fd>{sys_munmap+77} <ffffffff8010ead6>{system_call+126}
May 25 21:17:38 grover kernel: [ 131.890928]
May 25 21:17:38 grover kernel: [ 131.898387]
May 25 21:17:38 grover kernel: [ 131.898388] Code: 48 8b 80 98 00 00 00 ff 88 08 01 00 00 c3 66 66 66 90 66 66

Hope this helps,
Ed Tomlinson

2005-05-26 03:42:15

by Andrew Morton

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1


(Added alsa-devel to cc)

Ed Tomlinson <[email protected]> wrote:
>
> On Wednesday 25 May 2005 16:49, you wrote:
> >
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/
> >
> > - Added Oracle's clustering filesystem driver, via git-ocfs.
> >
> > ? ? OCFS2, a shared disk cluster file system. ?See
> > ? ? Documentation/filesystems/ocfs2.txt Additionally a users guide is
> > ? ? available at:
> > ? ? http://oss.oracle.com/projects/ocfs2-tools/dist/documentation/users_guide.txt
> >
> > - New Xtensa architecture: Tensilica Xtensa CPU series.
> >
> > - Added the Red Hat distributed lock manager for people to look at.
> >
> > - Various new git trees. ?The remaining holdouts are:
>
> Got the following when I tried to use sound. Anyone else see problems in alsa land?
>
> May 25 21:17:38 grover kernel: [ 131.322558] PGD 2ab13067 PUD 2ab1b067 PMD 2a3f1067 PTE 0
> May 25 21:17:38 grover kernel: [ 131.351464] CPU 0
> May 25 21:17:38 grover kernel: [ 131.358041] Modules linked in: radeon nfsd exportfs lockd sunrpc sg parport_pc lp parport ipv6 sd_
> mod evdev mousedev tsdev usbhid usb_storage snd_intel8x0 snd_ac97_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore sn
> d_page_alloc ehci_hcd ohci_hcd eth1394 sata_nv libata forcedeth
> ohci1394 powernow_k8 freq_table processor cpufreq_userspace w83627hf eeprom i2c_sensor i2c_isa i2c_nforce2 i2c_core sr_mod sbp2 scs
> i_mod ieee1394 rtc unix
> May 25 21:17:38 grover kernel: [ 131.487598] Pid: 5481, comm: artsd Not tainted 2.6.12-rc5-mm1
> May 25 21:17:38 grover kernel: [ 131.506496] RIP: 0010:[__nosave_end+129759479/2131247104] <ffffffff8813b8f7>{:snd_pcm:snd_pcm_mmap
> _data_close+7}
> May 25 21:17:38 grover kernel: [ 131.535142] RSP: 0018:ffff81002ab9dee0 EFLAGS: 00010286
> May 25 21:17:38 grover kernel: [ 131.553183] RAX: 00002aaaadb59000 RBX: ffff810029c64988 RCX: fffffffffffffff2
> May 25 21:17:38 grover kernel: [ 131.576657] RDX: ffff810029c64988 RSI: ffff81002d672ce0 RDI: ffff81002b339f50
> May 25 21:17:38 grover kernel: [ 131.600129] RBP: ffff81002b3051c0 R08: ffff810029c649a8 R09: ffff81002ab9dec8
> May 25 21:17:38 grover kernel: [ 131.623602] R10: 0000000000000002 R11: 0000000000000206 R12: ffff81002b339f50
> May 25 21:17:38 grover kernel: [ 131.647075] R13: ffff81002b339f50 R14: ffff81002e7f2a08 R15: 00002aaaadb67000
> May 25 21:17:38 grover kernel: [ 131.670548] FS: 00002aaaad6d8f50(0000) GS:ffffffff80550840(0000) knlGS:0000000000000000
> May 25 21:17:38 grover kernel: [ 131.697165] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> May 25 21:17:38 grover kernel: [ 131.716064] CR2: 00002aaaadb59098 CR3: 000000002aafd000 CR4: 00000000000006e0
> May 25 21:17:38 grover kernel: [ 131.739537] Process artsd (pid: 5481, threadinfo ffff81002ab9c000, task ffff81002ada97f0)
> May 25 21:17:38 grover kernel: [ 131.766439] Stack: ffffffff8016942d 0000000000000000 0000000000000000 ffff81002e7f2a00
> May 25 21:17:38 grover kernel: [ 131.792174] ffffffff8016a936 ffff81002d672df0 ffff81002d672e08 ffff81002d672df0
> May 25 21:17:38 grover kernel: [ 131.818481] ffff81002e7f2a00 0000000000560600
> May 25 21:17:38 grover kernel: [ 131.835065] Call Trace:<ffffffff8016942d>{remove_vm_struct+125} <ffffffff8016a936>{do_munmap+550}
> May 25 21:17:38 grover kernel: [ 131.864282] <ffffffff8016b0fd>{sys_munmap+77} <ffffffff8010ead6>{system_call+126}
> May 25 21:17:38 grover kernel: [ 131.890928]
> May 25 21:17:38 grover kernel: [ 131.898387]
> May 25 21:17:38 grover kernel: [ 131.898388] Code: 48 8b 80 98 00 00 00 ff 88 08 01 00 00 c3 66 66 66 90 66 66
>
> Hope this helps,
> Ed Tomlinson

2005-05-26 05:30:52

by Yani Ioannou

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Hi Andrew,

This patch updates all the device attribute callbacks that weren't
updated with the new parameter, I guess because they weren't in Greg's
tree (including drivers/pcmcia/ds.c). Without the patch these
callbacks are probably broken (and generate a warning along the lines
of "assignment from incompatible pointer type").

Please see http://lkml.org/lkml/2005/5/19/40 for the scripts I used to
update the attributes automatically.

Signed-off-by: Yani Ioannou <[email protected]>

Thanks,
Yani

char/tpm/tpm.c | 2 +-
char/tpm/tpm.h | 8 ++++----
message/i2o/bus-osm.c | 2 +-
message/i2o/exec-osm.c | 4 ++--
pcmcia/ds.c | 2 +-
5 files changed, 9 insertions(+), 9 deletions(-)


Attachments:
(No filename) (743.00 B)
patch-linux-2.6.12-rc5-mm1-devattrupdate.diff (3.89 kB)
Download all attachments

2005-05-26 06:37:42

by Martin Bligh

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Build failure on numaq:
http://ftp.kernel.org/pub/linux/kernel/people/mbligh/config/abat/numaq

In file included from include/linux/sched.h:12,
from arch/i386/kernel/asm-offsets.c:7:
include/linux/jiffies.h:42:3: #error You lose.
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:213:31: division by zero in #if
include/linux/jiffies.h:257:30: division by zero in #if
In file included from include/linux/sched.h:12,
from arch/i386/kernel/asm-offsets.c:7:
include/linux/jiffies.h: In function `jiffies_to_msecs':
include/linux/jiffies.h:262: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h:262: error: (Each undeclared identifier is reported only once
include/linux/jiffies.h:262: error: for each function it appears in.)
include/linux/jiffies.h:268:36: division by zero in #if
include/linux/jiffies.h: In function `jiffies_to_usecs':
include/linux/jiffies.h:273: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h:281:30: division by zero in #if
include/linux/jiffies.h: In function `msecs_to_jiffies':
include/linux/jiffies.h:286: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h:294:36: division by zero in #if
include/linux/jiffies.h: In function `usecs_to_jiffies':
include/linux/jiffies.h:299: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h: In function `timespec_to_jiffies':
include/linux/jiffies.h:318: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h:324: error: `SHIFT_HZ' undeclared (first use in this function)
include/linux/jiffies.h: In function `jiffies_to_timespec':
include/linux/jiffies.h:337: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h: In function `timeval_to_jiffies':
include/linux/jiffies.h:359: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h:363: error: `SHIFT_HZ' undeclared (first use in this function)
include/linux/jiffies.h: In function `jiffies_to_timeval':
include/linux/jiffies.h:375: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h:385:6: division by zero in #if
include/linux/jiffies.h: In function `jiffies_to_clock_t':
include/linux/jiffies.h:386: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h: In function `clock_t_to_jiffies':
include/linux/jiffies.h:397: error: `CONFIG_HZ' undeclared (first use in this function)
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h:416:6: division by zero in #if
include/linux/jiffies.h: In function `jiffies_64_to_clock_t':
include/linux/jiffies.h:417: error: `CONFIG_HZ' undeclared (first use in this function)
make[1]: *** [arch/i386/kernel/asm-offsets.s] Error 1
make: *** [arch/i386/kernel/asm-offsets.s] Error 2
05/25/05-20:57:45 Build the kernel. Failed rc = 2
05/25/05-20:57:45 build: kernel build Failed rc = 1

2005-05-26 06:41:17

by Martin Bligh

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1



--"Martin J. Bligh" <[email protected]> wrote (on Wednesday, May 25, 2005 23:37:26 -0700):

> Build failure on numaq:
> http://ftp.kernel.org/pub/linux/kernel/people/mbligh/config/abat/numaq

I take that back ... it's actually every ia32 machine I have that
fails to build ...

M.

> In file included from include/linux/sched.h:12,
> from arch/i386/kernel/asm-offsets.c:7:
> include/linux/jiffies.h:42:3: #error You lose.
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:213:31: division by zero in #if
> include/linux/jiffies.h:257:30: division by zero in #if
> In file included from include/linux/sched.h:12,
> from arch/i386/kernel/asm-offsets.c:7:
> include/linux/jiffies.h: In function `jiffies_to_msecs':
> include/linux/jiffies.h:262: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h:262: error: (Each undeclared identifier is reported only once
> include/linux/jiffies.h:262: error: for each function it appears in.)
> include/linux/jiffies.h:268:36: division by zero in #if
> include/linux/jiffies.h: In function `jiffies_to_usecs':
> include/linux/jiffies.h:273: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h:281:30: division by zero in #if
> include/linux/jiffies.h: In function `msecs_to_jiffies':
> include/linux/jiffies.h:286: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h:294:36: division by zero in #if
> include/linux/jiffies.h: In function `usecs_to_jiffies':
> include/linux/jiffies.h:299: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h: In function `timespec_to_jiffies':
> include/linux/jiffies.h:318: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h:324: error: `SHIFT_HZ' undeclared (first use in this function)
> include/linux/jiffies.h: In function `jiffies_to_timespec':
> include/linux/jiffies.h:337: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h: In function `timeval_to_jiffies':
> include/linux/jiffies.h:359: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h:363: error: `SHIFT_HZ' undeclared (first use in this function)
> include/linux/jiffies.h: In function `jiffies_to_timeval':
> include/linux/jiffies.h:375: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h:385:6: division by zero in #if
> include/linux/jiffies.h: In function `jiffies_to_clock_t':
> include/linux/jiffies.h:386: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h: In function `clock_t_to_jiffies':
> include/linux/jiffies.h:397: error: `CONFIG_HZ' undeclared (first use in this function)
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h:416:6: division by zero in #if
> include/linux/jiffies.h: In function `jiffies_64_to_clock_t':
> include/linux/jiffies.h:417: error: `CONFIG_HZ' undeclared (first use in this function)
> make[1]: *** [arch/i386/kernel/asm-offsets.s] Error 1
> make: *** [arch/i386/kernel/asm-offsets.s] Error 2
> 05/25/05-20:57:45 Build the kernel. Failed rc = 2
> 05/25/05-20:57:45 build: kernel build Failed rc = 1


2005-05-26 06:48:35

by Andrew Morton

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

"Martin J. Bligh" <[email protected]> wrote:
>
> Build failure on numaq:
> http://ftp.kernel.org/pub/linux/kernel/people/mbligh/config/abat/numaq
>
> In file included from include/linux/sched.h:12,
> from arch/i386/kernel/asm-offsets.c:7:
> include/linux/jiffies.h:42:3: #error You lose.

You lost! CONFIG_HZ didn't get set.

Something obviously went wrong in the magic in kernel/Kconfig.hz. Wanna do
`grep HZ .config' and see if you can work out why it broke?

Roman, is there a better way of doing the below?


From: Christoph Lameter <[email protected]>

Make the timer frequency selectable. The timer interrupt may cause bus
and memory contention in large NUMA systems since the interrupt occurs
on each processor HZ times per second.

Signed-off-by: Christoph Lameter <[email protected]>
Signed-off-by: Shai Fultheim <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
---

arch/i386/Kconfig | 2 +
arch/x86_64/Kconfig | 2 +
include/asm-i386/param.h | 4 ++-
include/asm-x86_64/param.h | 6 +++--
kernel/Kconfig.hz | 46 +++++++++++++++++++++++++++++++++++++++++++++
5 files changed, 57 insertions(+), 3 deletions(-)

diff -puN arch/i386/Kconfig~i386-selectable-frequency-of-the-timer-interrupt arch/i386/Kconfig
--- 25/arch/i386/Kconfig~i386-selectable-frequency-of-the-timer-interrupt 2005-05-23 19:41:04.000000000 -0700
+++ 25-akpm/arch/i386/Kconfig 2005-05-23 19:41:04.000000000 -0700
@@ -1116,6 +1116,8 @@ config APM_REAL_MODE_POWER_OFF
a work-around for a number of buggy BIOSes. Switch this option on if
your computer crashes instead of powering off properly.

+source kernel/Kconfig.hz
+
endmenu

source "arch/i386/kernel/cpu/cpufreq/Kconfig"
diff -puN arch/x86_64/Kconfig~i386-selectable-frequency-of-the-timer-interrupt arch/x86_64/Kconfig
--- 25/arch/x86_64/Kconfig~i386-selectable-frequency-of-the-timer-interrupt 2005-05-23 19:41:04.000000000 -0700
+++ 25-akpm/arch/x86_64/Kconfig 2005-05-23 19:41:04.000000000 -0700
@@ -401,6 +401,8 @@ config SECCOMP

If unsure, say Y. Only embedded should say N here.

+source kernel/Kconfig.hz
+
endmenu

#
diff -puN include/asm-i386/param.h~i386-selectable-frequency-of-the-timer-interrupt include/asm-i386/param.h
--- 25/include/asm-i386/param.h~i386-selectable-frequency-of-the-timer-interrupt 2005-05-23 19:41:04.000000000 -0700
+++ 25-akpm/include/asm-i386/param.h 2005-05-23 19:41:04.000000000 -0700
@@ -1,8 +1,10 @@
+#include <linux/config.h>
+
#ifndef _ASMi386_PARAM_H
#define _ASMi386_PARAM_H

#ifdef __KERNEL__
-# define HZ 1000 /* Internal kernel timer frequency */
+# define HZ CONFIG_HZ /* Internal kernel timer frequency */
# define USER_HZ 100 /* .. some user interfaces are in "ticks" */
# define CLOCKS_PER_SEC (USER_HZ) /* like times() */
#endif
diff -puN include/asm-x86_64/param.h~i386-selectable-frequency-of-the-timer-interrupt include/asm-x86_64/param.h
--- 25/include/asm-x86_64/param.h~i386-selectable-frequency-of-the-timer-interrupt 2005-05-23 19:41:04.000000000 -0700
+++ 25-akpm/include/asm-x86_64/param.h 2005-05-23 19:41:04.000000000 -0700
@@ -1,9 +1,11 @@
+#include <linux/config.h>
+
#ifndef _ASMx86_64_PARAM_H
#define _ASMx86_64_PARAM_H

#ifdef __KERNEL__
-# define HZ 1000 /* Internal kernel timer frequency */
-# define USER_HZ 100 /* .. some user interfaces are in "ticks */
+# define HZ CONFIG_HZ /* Internal kernel timer frequency */
+# define USER_HZ 100 /* .. some user interfaces are in "ticks */
#define CLOCKS_PER_SEC (USER_HZ) /* like times() */
#endif

diff -puN /dev/null kernel/Kconfig.hz
--- /dev/null 2003-09-15 06:40:47.000000000 -0700
+++ 25-akpm/kernel/Kconfig.hz 2005-05-23 19:41:04.000000000 -0700
@@ -0,0 +1,46 @@
+#
+# Timer Interrupt Frequency Configuration
+#
+
+choice
+ prompt "Timer frequency"
+ default HZ_250
+ help
+ Allows the configuration of the timer frequency. It is customary
+ to have the timer interrupt run at 1000 HZ but 100 HZ may be more
+ beneficial for servers and NUMA systems that do not need to have
+ a fast response for user interaction and that may experience bus
+ contention and cacheline bounces as a result of timer interrupts.
+ Note that the timer interrupt occurs on each processor in an SMP
+ environment leading to NR_CPUS * HZ number of timer interrupts
+ per second.
+
+
+ config HZ_100
+ bool "100 HZ"
+ help
+ 100 HZ is a typical choice for servers, SMP and NUMA systems
+ with lots of processors that may show reduced performance if
+ too many timer interrupts are occurring.
+
+ config HZ_250
+ bool "250 HZ"
+ help
+ 250 HZ is a good compromise choice allowing server performance
+ while also showing good interactive responsiveness even
+ on SMP and NUMA systems.
+
+ config HZ_1000
+ bool "1000 HZ"
+ help
+ 1000 HZ is the preferred choice for desktop systems and other
+ systems requiring fast interactive responses to events.
+
+endchoice
+
+config HZ
+ int
+ default 100 if HZ_100
+ default 250 if HZ_250
+ default 1000 if HZ_1000
+
_

2005-05-26 07:05:54

by Martin Bligh

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1



--Andrew Morton <[email protected]> wrote (on Wednesday, May 25, 2005 23:47:17 -0700):

> "Martin J. Bligh" <[email protected]> wrote:
>>
>> Build failure on numaq:
>> http://ftp.kernel.org/pub/linux/kernel/people/mbligh/config/abat/numaq
>>
>> In file included from include/linux/sched.h:12,
>> from arch/i386/kernel/asm-offsets.c:7:
>> include/linux/jiffies.h:42:3: #error You lose.
>
> You lost! CONFIG_HZ didn't get set.
>
> Something obviously went wrong in the magic in kernel/Kconfig.hz. Wanna do
> `grep HZ .config' and see if you can work out why it broke?

Tis conspicious by it's absence.

mbligh@kernel:~/linux-2.6.12-rc5-mm1$ grep HZ .config
mbligh@kernel:~/linux-2.6.12-rc5-mm1$

I'll poke at it in the morning, with the benfits of less wine, and more
sleep

M.

2005-05-26 07:14:41

by Martin Bligh

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1



--"Martin J. Bligh" <[email protected]> wrote (on Thursday, May 26, 2005 00:05:33 -0700):

>
>
> --Andrew Morton <[email protected]> wrote (on Wednesday, May 25, 2005 23:47:17 -0700):
>
>> "Martin J. Bligh" <[email protected]> wrote:
>>>
>>> Build failure on numaq:
>>> http://ftp.kernel.org/pub/linux/kernel/people/mbligh/config/abat/numaq
>>>
>>> In file included from include/linux/sched.h:12,
>>> from arch/i386/kernel/asm-offsets.c:7:
>>> include/linux/jiffies.h:42:3: #error You lose.
>>
>> You lost! CONFIG_HZ didn't get set.
>>
>> Something obviously went wrong in the magic in kernel/Kconfig.hz. Wanna do
>> `grep HZ .config' and see if you can work out why it broke?
>
> Tis conspicious by it's absence.
>
> mbligh@kernel:~/linux-2.6.12-rc5-mm1$ grep HZ .config
> mbligh@kernel:~/linux-2.6.12-rc5-mm1$
>
> I'll poke at it in the morning, with the benfits of less wine, and more
> sleep
>
> M.

source kernel/Kconfig.hz is under:
menu "APM (Advanced Power Management) BIOS Support"
depends on PM && !X86_VISWS

So it's screwed if you don't have PM defined, it seems.

2005-05-26 07:23:50

by Martin Bligh

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1



--"Martin J. Bligh" <[email protected]> wrote (on Thursday, May 26, 2005 00:14:34 -0700):

>
>
> --"Martin J. Bligh" <[email protected]> wrote (on Thursday, May 26, 2005 00:05:33 -0700):
>
>>
>>
>> --Andrew Morton <[email protected]> wrote (on Wednesday, May 25, 2005 23:47:17 -0700):
>>
>>> "Martin J. Bligh" <[email protected]> wrote:
>>>>
>>>> Build failure on numaq:
>>>> http://ftp.kernel.org/pub/linux/kernel/people/mbligh/config/abat/numaq
>>>>
>>>> In file included from include/linux/sched.h:12,
>>>> from arch/i386/kernel/asm-offsets.c:7:
>>>> include/linux/jiffies.h:42:3: #error You lose.
>>>
>>> You lost! CONFIG_HZ didn't get set.
>>>
>>> Something obviously went wrong in the magic in kernel/Kconfig.hz. Wanna do
>>> `grep HZ .config' and see if you can work out why it broke?
>>
>> Tis conspicious by it's absence.
>>
>> mbligh@kernel:~/linux-2.6.12-rc5-mm1$ grep HZ .config
>> mbligh@kernel:~/linux-2.6.12-rc5-mm1$
>>
>> I'll poke at it in the morning, with the benfits of less wine, and more
>> sleep
>>
>> M.
>
> source kernel/Kconfig.hz is under:
> menu "APM (Advanced Power Management) BIOS Support"
> depends on PM && !X86_VISWS
>
> So it's screwed if you don't have PM defined, it seems.

Ironically once I work around that it fails with:

arch/i386/kernel/built-in.o(.init.text+0xfe4): In function `sys_set_thread_area':
arch/i386/kernel/process.c:869: undefined reference to `acpi_read_root_resources'
drivers/built-in.o(.text+0x1fa3d): In function `acpi_pci_root_add':
drivers/acpi/pci_root.c:275: undefined reference to `pci_acpi_scan_root'
make: *** [.tmp_vmlinux1] Error 1

instead. Oh well. bedtime.

M.

2005-05-26 07:26:17

by Andrew Morton

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

"Martin J. Bligh" <[email protected]> wrote:
>
> source kernel/Kconfig.hz is under:
> menu "APM (Advanced Power Management) BIOS Support"
> depends on PM && !X86_VISWS
>
> So it's screwed if you don't have PM defined, it seems.

Ah, OK. Something like this:

--- 25/arch/i386/Kconfig~i386-selectable-frequency-of-the-timer-interrupt-fix 2005-05-26 00:22:55.000000000 -0700
+++ 25-akpm/arch/i386/Kconfig 2005-05-26 00:22:55.000000000 -0700
@@ -1116,8 +1116,6 @@ config APM_REAL_MODE_POWER_OFF
a work-around for a number of buggy BIOSes. Switch this option on if
your computer crashes instead of powering off properly.

-source kernel/Kconfig.hz
-
endmenu

source "arch/i386/kernel/cpu/cpufreq/Kconfig"
@@ -1275,6 +1273,8 @@ source "crypto/Kconfig"

source "lib/Kconfig"

+source kernel/Kconfig.hz
+
#
# Use the generic interrupt handling code in kernel/irq/:
#
diff -puN arch/x86_64/Kconfig~i386-selectable-frequency-of-the-timer-interrupt-fix arch/x86_64/Kconfig
--- 25/arch/x86_64/Kconfig~i386-selectable-frequency-of-the-timer-interrupt-fix 2005-05-26 00:22:55.000000000 -0700
+++ 25-akpm/arch/x86_64/Kconfig 2005-05-26 00:22:55.000000000 -0700
@@ -401,8 +401,6 @@ config SECCOMP

If unsure, say Y. Only embedded should say N here.

-source kernel/Kconfig.hz
-
endmenu

#
@@ -517,3 +515,5 @@ source "security/Kconfig"
source "crypto/Kconfig"

source "lib/Kconfig"
+
+source kernel/Kconfig.hz
_



and this:


--- 25/arch/ia64/Kconfig~ia64-selectable-timer-interrupt-frequency-fix 2005-05-26 00:23:18.000000000 -0700
+++ 25-akpm/arch/ia64/Kconfig 2005-05-26 00:23:39.000000000 -0700
@@ -157,8 +157,6 @@ config IA64_PAGE_SIZE_64KB

endchoice

-source kernel/Kconfig.hz
-
config IA64_BRL_EMU
bool
depends on ITANIUM
@@ -445,3 +443,5 @@ source "arch/ia64/Kconfig.debug"
source "security/Kconfig"

source "crypto/Kconfig"
+
+source kernel/Kconfig.hz
_

2005-05-26 07:43:33

by J.A. Magallon

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1


On 05.26, Andrew Morton wrote:
>
> (Added alsa-devel to cc)
>
> Ed Tomlinson <[email protected]> wrote:
> >
> > Got the following when I tried to use sound. Anyone else see problems in alsa land?
> >

Me too. As beep-media-player ends playing a mp3 track, oops !
Decoded below, for if it gives additional info:

ksymoops 2.4.9 on i686 2.6.11-jam20. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.6.11-jam20/ (default)
-m /boot/System.map-2.6.11-jam20 (default)

Warning: You did not tell me where to find symbol information. I will
assume that the log matches the kernel and modules that are running
right now and I'll use the default options above for symbol resolution.
If the current kernel and/or modules do not match the log, you can get
more accurate output by telling me the kernel version and where to find
map, modules, ksyms etc. ksymoops -h explains the options.

Error (regular_file): read_ksyms stat /proc/ksyms failed
No modules in ksyms, skipping objects
No ksyms, skipping lsmod
May 26 09:35:15 werewolf kernel: <1>Unable to handle kernel paging request at virtual address a5a5c060
May 26 09:35:15 werewolf kernel: f0a08f70
May 26 09:35:15 werewolf kernel: *pde = 0295d067
May 26 09:35:15 werewolf kernel: Oops: 0000 [#2]
May 26 09:35:15 werewolf kernel: CPU: 0
May 26 09:35:15 werewolf kernel: EIP: 0060:[pg0+1079750512/1337656320] Tainted: P VLI
May 26 09:35:15 werewolf kernel: EIP: 0060:[<f0a08f70>] Tainted: P VLI
Using defaults from ksymoops -t elf32-i386 -a i386
May 26 09:35:15 werewolf kernel: EFLAGS: 00210282 (2.6.11-jam20)
May 26 09:35:15 werewolf kernel: eax: a5a5c000 ebx: eebba67c ecx: b2101f08 edx: f0a08f6d
May 26 09:35:15 werewolf kernel: esi: eebba6a4 edi: b26b7c78 ebp: c6da6680 esp: b2d50f6c
May 26 09:35:15 werewolf kernel: ds: 007b es: 007b ss: 0068
May 26 09:35:15 werewolf kernel: Stack: b0147ce3 00000000 eec3d580 b26b7c78 a5a5c000 b0149469 a5a4c000 eec3d580
May 26 09:35:15 werewolf kernel: b01497ce a5a4c000 a5a5c000 be93d120 eec3d580 eec3d5b0 00000002 b2d50000
May 26 09:35:15 werewolf kernel: b014983a a5a4c000 000000b8 b0102b0b a5a4c000 00010000 a6d131b4 000000b8
May 26 09:35:15 werewolf kernel: Call Trace:
May 26 09:35:15 werewolf kernel: [<b0147ce3>] remove_vm_struct+0x46/0x68
May 26 09:35:15 werewolf kernel: [<b0149469>] unmap_vma_list+0xe/0x17
May 26 09:35:15 werewolf kernel: [<b01497ce>] do_munmap+0xcb/0xff
May 26 09:35:15 werewolf kernel: [<b014983a>] sys_munmap+0x38/0x50
May 26 09:35:15 werewolf kernel: [<b0102b0b>] sysenter_past_esp+0x54/0x75
May 26 09:35:15 werewolf kernel: Code: e1 05 03 0d b0 c4 43 b0 eb c9 c7 42 44 9c 4d a1 f0 89 42 50 81 4a 14 00 00 08 00 8b 40 60 f0 ff 80 a4 00 00 00 31 c0 c3 8b 40 50 <8b> 40 60 f0 ff 88 a4 00 00 00 c3 8b 40 50 8b 40 60 f0 ff 80 a4


>>EIP; f0a08f70 <pg0+405baf70/4fbb0400> <=====

>>eax; a5a5c000 <phys_startup_32+a595c000/b0000000>
>>ebx; eebba67c <pg0+3e76c67c/4fbb0400>
>>ecx; b2101f08 <pg0+1cb3f08/4fbb0400>
>>edx; f0a08f6d <pg0+405baf6d/4fbb0400>
>>esi; eebba6a4 <pg0+3e76c6a4/4fbb0400>
>>edi; b26b7c78 <pg0+2269c78/4fbb0400>
>>ebp; c6da6680 <pg0+16958680/4fbb0400>
>>esp; b2d50f6c <pg0+2902f6c/4fbb0400>

Trace; b0147ce3 <remove_vm_struct+46/68>
Trace; b0149469 <unmap_vma_list+e/17>
Trace; b01497ce <do_munmap+cb/ff>
Trace; b014983a <sys_munmap+38/50>
Trace; b0102b0b <sysenter_past_esp+54/75>

This architecture has variable length instructions, decoding before eip
is unreliable, take these instructions with a pinch of salt.

Code; f0a08f45 <pg0+405baf45/4fbb0400>
00000000 <_EIP>:
Code; f0a08f45 <pg0+405baf45/4fbb0400>
0: e1 05 loope 7 <_EIP+0x7>
Code; f0a08f47 <pg0+405baf47/4fbb0400>
2: 03 0d b0 c4 43 b0 add 0xb043c4b0,%ecx
Code; f0a08f4d <pg0+405baf4d/4fbb0400>
8: eb c9 jmp ffffffd3 <_EIP+0xffffffd3>
Code; f0a08f4f <pg0+405baf4f/4fbb0400>
a: c7 42 44 9c 4d a1 f0 movl $0xf0a14d9c,0x44(%edx)
Code; f0a08f56 <pg0+405baf56/4fbb0400>
11: 89 42 50 mov %eax,0x50(%edx)
Code; f0a08f59 <pg0+405baf59/4fbb0400>
14: 81 4a 14 00 00 08 00 orl $0x80000,0x14(%edx)
Code; f0a08f60 <pg0+405baf60/4fbb0400>
1b: 8b 40 60 mov 0x60(%eax),%eax
Code; f0a08f63 <pg0+405baf63/4fbb0400>
1e: f0 ff 80 a4 00 00 00 lock incl 0xa4(%eax)
Code; f0a08f6a <pg0+405baf6a/4fbb0400>
25: 31 c0 xor %eax,%eax
Code; f0a08f6c <pg0+405baf6c/4fbb0400>
27: c3 ret
Code; f0a08f6d <pg0+405baf6d/4fbb0400>
28: 8b 40 50 mov 0x50(%eax),%eax

This decode from eip onwards should be reliable

Code; f0a08f70 <pg0+405baf70/4fbb0400>
00000000 <_EIP>:
Code; f0a08f70 <pg0+405baf70/4fbb0400> <=====
0: 8b 40 60 mov 0x60(%eax),%eax <=====
Code; f0a08f73 <pg0+405baf73/4fbb0400>
3: f0 ff 88 a4 00 00 00 lock decl 0xa4(%eax)
Code; f0a08f7a <pg0+405baf7a/4fbb0400>
a: c3 ret
Code; f0a08f7b <pg0+405baf7b/4fbb0400>
b: 8b 40 50 mov 0x50(%eax),%eax
Code; f0a08f7e <pg0+405baf7e/4fbb0400>
e: 8b 40 60 mov 0x60(%eax),%eax
Code; f0a08f81 <pg0+405baf81/4fbb0400>
11: f0 lock
Code; f0a08f82 <pg0+405baf82/4fbb0400>
12: ff .byte 0xff
Code; f0a08f83 <pg0+405baf83/4fbb0400>
13: 80 .byte 0x80
Code; f0a08f84 <pg0+405baf84/4fbb0400>
14: a4 movsb %ds:(%esi),%es:(%edi)


1 warning and 1 error issued. Results may not be reliable.

--
J.A. Magallon <jamagallon()able!es> \ Software is like sex:
werewolf!able!es \ It's better when it's free
Mandriva Linux release 2006.0 (Cooker) for i586
Linux 2.6.11-jam20 (gcc 4.0.0 (4.0.0-3mdk for Mandriva Linux release 2006.0))


2005-05-26 07:45:17

by J.A. Magallon

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1


On 05.25, Andrew Morton wrote:
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/
>
>
> - Again, if there are patches in here which you think should be merged in
> 2.6.12, please point them out to me.
>

I collected this from lkml:

--- linux-2.6.12-rc4-mm2/arch/i386/kernel/cpu/intel_cacheinfo.c.old 2005-05-17 00:05:28.000000000 +0200
+++ linux-2.6.12-rc4-mm2/arch/i386/kernel/cpu/intel_cacheinfo.c 2005-05-17 00:05:49.000000000 +0200
@@ -118,7 +118,7 @@
};

#define MAX_CACHE_LEAVES 4
-static unsigned short __devinitdata num_cache_leaves;
+static unsigned short num_cache_leaves;

static int __devinit cpuid4_cache_lookup(int index, struct _cpuid4_info *this_leaf)
{

--
J.A. Magallon <jamagallon()able!es> \ Software is like sex:
werewolf!able!es \ It's better when it's free
Mandriva Linux release 2006.0 (Cooker) for i586
Linux 2.6.11-jam20 (gcc 4.0.0 (4.0.0-3mdk for Mandriva Linux release 2006.0))


2005-05-26 07:53:47

by Andrew Morton

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

"J.A. Magallon" <[email protected]> wrote:
>
> I collected this from lkml:
>
> --- linux-2.6.12-rc4-mm2/arch/i386/kernel/cpu/intel_cacheinfo.c.old 2005-05-17 00:05:28.000000000 +0200
> +++ linux-2.6.12-rc4-mm2/arch/i386/kernel/cpu/intel_cacheinfo.c 2005-05-17 00:05:49.000000000 +0200
> @@ -118,7 +118,7 @@
> };
>
> #define MAX_CACHE_LEAVES 4
> -static unsigned short __devinitdata num_cache_leaves;
> +static unsigned short num_cache_leaves;

Got it, thanks.

2005-05-26 07:59:45

by Andrew Morton

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

"J.A. Magallon" <[email protected]> wrote:
>
>
> On 05.26, Andrew Morton wrote:
> >
> > (Added alsa-devel to cc)
> >
> > Ed Tomlinson <[email protected]> wrote:
> > >
> > > Got the following when I tried to use sound. Anyone else see problems in alsa land?
> > >
>
> Me too. As beep-media-player ends playing a mp3 track, oops !

hm, OK, you're also on x86_64. What sound card and driver?

> Decoded below, for if it gives additional info:

Actually, no, ksymoops removes info. Please just send the kernel's oops
output directly. (Make sure that CONFIG_KALLSYMS=y).

2005-05-26 08:57:42

by Mikael Pettersson

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

2.6.12-rc5-mm1 includes Andrea's seccomp-disable-tsc patch,
which I believe is broken on SMP. In process.c we find:

/*
+ * This function selects if the context switch from prev to next
+ * has to tweak the TSC disable bit in the cr4.
+ */
+static void disable_tsc(struct thread_info *prev,
+ struct thread_info *next)
+{
+ if (unlikely(has_secure_computing(prev) ||
+ has_secure_computing(next))) {
+ /* slow path here */
+ if (has_secure_computing(prev) &&
+ !has_secure_computing(next)) {
+ clear_in_cr4(X86_CR4_TSD);
+ } else if (!has_secure_computing(prev) &&
+ has_secure_computing(next))
+ set_in_cr4(X86_CR4_TSD);
+ }
+}

which it calls from __switch_to().

The problem is that {set,clear}_in_cr4() both update a single
global mmu_cr4_features variable, which is asynchronously written
to all CPUs by {,__}flush_tlb_all(). Hence, the CR4.TSD setting
is at best probabilistic.

I spotted this because perfctr used to flip CR4.PCE in __switch_to()
ages ago, but I had to abandon that when kernel 2.3.40 changed to
the current scheme with a global mmu_cr4_features.
(Another reason was that CR4 writes were and still are very slow.)

/Mikael

2005-05-26 11:09:58

by Roman Zippel

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Hi,

On Thu, 26 May 2005, Andrew Morton wrote:

> "Martin J. Bligh" <[email protected]> wrote:
> >
> > source kernel/Kconfig.hz is under:
> > menu "APM (Advanced Power Management) BIOS Support"
> > depends on PM && !X86_VISWS
> >
> > So it's screwed if you don't have PM defined, it seems.
>
> Ah, OK. Something like this:

Only i386 has this problem, the others put it into the "Processor type and
features" menu, where it doesn't get the extra dependencies, so the
easiest solution might to move it for i386 there as well.

bye, Roman

2005-05-26 13:04:20

by Andrea Arcangeli

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Thu, May 26, 2005 at 10:57:27AM +0200, Mikael Pettersson wrote:
> 2.6.12-rc5-mm1 includes Andrea's seccomp-disable-tsc patch,
> which I believe is broken on SMP. In process.c we find:
>
> /*
> + * This function selects if the context switch from prev to next
> + * has to tweak the TSC disable bit in the cr4.
> + */
> +static void disable_tsc(struct thread_info *prev,
> + struct thread_info *next)
> +{
> + if (unlikely(has_secure_computing(prev) ||
> + has_secure_computing(next))) {
> + /* slow path here */
> + if (has_secure_computing(prev) &&
> + !has_secure_computing(next)) {
> + clear_in_cr4(X86_CR4_TSD);
> + } else if (!has_secure_computing(prev) &&
> + has_secure_computing(next))
> + set_in_cr4(X86_CR4_TSD);
> + }
> +}
>
> which it calls from __switch_to().
>
> The problem is that {set,clear}_in_cr4() both update a single
> global mmu_cr4_features variable, which is asynchronously written
> to all CPUs by {,__}flush_tlb_all(). Hence, the CR4.TSD setting
> is at best probabilistic.

You're right. This won't destabilize the kernel (and it couldn't trigger
in my testing) but it may lead to the tsc to be erroneously disabled on
a non-seccomp task, after a change_page_attr (i.e. insmod) or similar
events using flush_tlb_all (like rmmod). I didn't notice this race
condition sorry (I now recall why we overwrite the cr4 flag in the
global tlb flush: just to flush the global pagetables, since vmalloc
uses global pagetables too).

> I spotted this because perfctr used to flip CR4.PCE in __switch_to()
> ages ago, but I had to abandon that when kernel 2.3.40 changed to
> the current scheme with a global mmu_cr4_features.
> (Another reason was that CR4 writes were and still are very slow.)

Speed is not an issue, cr4 would never be tweaked unless you use
seccomp, the cr4 flip is in an extreme slow path.

I think there are two ways to solve this race:

1) why don't we read the cr4 from the cpu? would movl %%cr4, %%eax
generate a general protection fault? Can the cr4 be read in ring 0?
Why to read it from memory if we've that information in the cpu already?
2) If there's a good reason to read if from memory, then what about
making the mmu_cr4_features per-cpu? That way would solve the problem
too.

Both solutions relay on the fact that the pagetable flush cannot happen
from irqs, so as long as set/clear_in_cr4 is never called from irq (like
in switch_to with this patch), no locking would be required. If the
pagetable flush can be called from irqs (or alternatively if the
set/clear_in_cr4 can be called from irqs) then we'd need to use
test_and_set/clear_bit instead of normal C code despite the structure
being per-cpu, or despite reading the data from the cpu instead of
reading it from memory. In theory we could add a BUG_ON(in_interrupt())
to both, since both set/clear_in_cr4 and the flush_tlb_all aren't fast
paths (probably it worth it just in case).

Comments welcome, thanks!

2005-05-26 13:54:59

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Hi,

On Thursday, 26 of May 2005 09:58, Andrew Morton wrote:
> "J.A. Magallon" <[email protected]> wrote:
> >
> >
> > On 05.26, Andrew Morton wrote:
> > >
> > > (Added alsa-devel to cc)
> > >
> > > Ed Tomlinson <[email protected]> wrote:
> > > >
> > > > Got the following when I tried to use sound. Anyone else see problems in alsa land?
> > > >
> >
> > Me too. As beep-media-player ends playing a mp3 track, oops !
>
> hm, OK, you're also on x86_64. What sound card and driver?

I've got the following on a dual-Opteron box with Tyan Thunder K8W (snd_intel8x0):

Unable to handle kernel paging request at 00002aaaaab2f098 RIP:
<ffffffff880b5da7>{:snd_pcm:snd_pcm_mmap_data_close+7}
PGD 178b1067 PUD 178b2067 PMD 178a3067 PTE 0
Oops: 0000 [1] SMP
CPU 0
Modules linked in: usbserial floppy nvram snd_pcm_oss snd_mixer_oss snd_intel8x0 snd_ac97_codec snd_pcm snd_timer snd soundcore snd_page_alloc ehci_hcd ohci_hcd usbcore joydev sg sr_mod ide_cd cdrom dm_mod parport_pc lp parport
Pid: 4201, comm: artsd Not tainted 2.6.12-rc5-mm1
RIP: 0010:[<ffffffff880b5da7>] <ffffffff880b5da7>{:snd_pcm:snd_pcm_mmap_data_close+7}
RSP: 0018:ffff810017149ed0 EFLAGS: 00010286
RAX: 00002aaaaab2f000 RBX: ffff81001cb2ed08 RCX: ffff810017646b18
RDX: ffff810017646b18 RSI: ffff81001841ead8 RDI: ffff81001841ea88
RBP: ffff81001841ea88 R08: ffff81001cb2ed28 R09: ffff810017149ec8
R10: ffffffffffffffff R11: 0000000000000202 R12: ffff81001cb2ed48
R13: ffff810017dc9418 R14: ffff81003b31e0b8 R15: 00002aaaaab3f000
FS: 00002aaaad039de0(0000) GS:ffffffff8054e840(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00002aaaaab2f098 CR3: 0000000038f5d000 CR4: 00000000000006e0
Process artsd (pid: 4201, threadinfo ffff810017148000, task ffff810017616810)
Stack: ffffffff801705ea 00002aaaaab2f000 0000000000000000 0000000000000000
ffff81003b31e0b0 ffff81001841ea88 ffffffff801718c6 ffff81001841e9b8
ffff81001841e9d0 ffff81001841e9b8
Call Trace:<ffffffff801705ea>{remove_vm_struct+106} <ffffffff801718c6>{do_munmap+550}
<ffffffff8017194d>{sys_munmap+77} <ffffffff8010dc6e>{system_call+126}


Code: 48 8b 80 98 00 00 00 f0 ff 88 08 01 00 00 c3 66 66 66 90 66
RIP <ffffffff880b5da7>{:snd_pcm:snd_pcm_mmap_data_close+7} RSP <ffff810017149ed0>
CR2: 00002aaaaab2f098


Greets,
Rafael


--
- Would you tell me, please, which way I ought to go from here?
- That depends a good deal on where you want to get to.
-- Lewis Carroll "Alice's Adventures in Wonderland"

2005-05-26 14:48:37

by Steven Cole

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Andrew Morton wrote:
> "Martin J. Bligh" <[email protected]> wrote:
>
>>source kernel/Kconfig.hz is under:
>> menu "APM (Advanced Power Management) BIOS Support"
>> depends on PM && !X86_VISWS
>>
>> So it's screwed if you don't have PM defined, it seems.
>
>
> Ah, OK. Something like this:
>

FWIW, I had the same error as Martin, and your patch fixed it
for me on my PIII.

Steven

2005-05-26 19:15:43

by Mikael Pettersson

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Andrea Arcangeli writes:
> Speed is not an issue, cr4 would never be tweaked unless you use
> seccomp, the cr4 flip is in an extreme slow path.
>
> I think there are two ways to solve this race:
>
> 1) why don't we read the cr4 from the cpu? would movl %%cr4, %%eax
> generate a general protection fault? Can the cr4 be read in ring 0?
> Why to read it from memory if we've that information in the cpu already?

Yes the kernel can read cr4. I believe the best solution is to modify
__flush_tlb_global() to read cr4's current value and mix it in when
toggling CR4.PGE and uploading mmu_cr4_features.
(But please make it conditional on CONFIG_SECCOMP.)

The code called from __switch_to() would have to set or clear cr4
locally only. That's easy using write_cr4() and read_cr4().

/Mikael

2005-05-26 20:52:24

by Andrew Morton

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

"Rafael J. Wysocki" <[email protected]> wrote:
>
> n Thursday, 26 of May 2005 09:58, Andrew Morton wrote:
> > "J.A. Magallon" <[email protected]> wrote:
> > >
> > >
> > > On 05.26, Andrew Morton wrote:
> > > >
> > > > (Added alsa-devel to cc)
> > > >
> > > > Ed Tomlinson <[email protected]> wrote:
> > > > >
> > > > > Got the following when I tried to use sound. Anyone else see problems in alsa land?
> > > > >
> > >
> > > Me too. As beep-media-player ends playing a mp3 track, oops !
> >
> > hm, OK, you're also on x86_64. What sound card and driver?
>
> I've got the following on a dual-Opteron box with Tyan Thunder K8W (snd_intel8x0):

OK, thanks. I guess we can set this problem aside for now, as this bug
isn't present in 2.6.12-rc5 (correct?).

I assume the problem is due to one of the ASLA patches in rc5-mm1, but it's
possible that it lies elsewhere. It would be great if someone could fire
up quilt and do a binary search...

To summarise, we have three people (one Opteron, one x86, one
unknown-x86_64) reporting this:


May 25 21:17:38 grover kernel: [ 131.322558] PGD 2ab13067 PUD 2ab1b067 PMD 2a3f1067 PTE 0
May 25 21:17:38 grover kernel: [ 131.351464] CPU 0
May 25 21:17:38 grover kernel: [ 131.358041] Modules linked in: radeon nfsd exportfs lockd sunrpc sg parport_pc lp parport ipv6 sd_
mod evdev mousedev tsdev usbhid usb_storage snd_intel8x0 snd_ac97_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore sn
d_page_alloc ehci_hcd ohci_hcd eth1394 sata_nv libata forcedeth
ohci1394 powernow_k8 freq_table processor cpufreq_userspace w83627hf eeprom i2c_sensor i2c_isa i2c_nforce2 i2c_core sr_mod sbp2 scs
i_mod ieee1394 rtc unix
May 25 21:17:38 grover kernel: [ 131.487598] Pid: 5481, comm: artsd Not tainted 2.6.12-rc5-mm1
May 25 21:17:38 grover kernel: [ 131.506496] RIP: 0010:[__nosave_end+129759479/2131247104] <ffffffff8813b8f7>{:snd_pcm:snd_pcm_mmap
_data_close+7}
May 25 21:17:38 grover kernel: [ 131.535142] RSP: 0018:ffff81002ab9dee0 EFLAGS: 00010286
May 25 21:17:38 grover kernel: [ 131.553183] RAX: 00002aaaadb59000 RBX: ffff810029c64988 RCX: fffffffffffffff2
May 25 21:17:38 grover kernel: [ 131.576657] RDX: ffff810029c64988 RSI: ffff81002d672ce0 RDI: ffff81002b339f50
May 25 21:17:38 grover kernel: [ 131.600129] RBP: ffff81002b3051c0 R08: ffff810029c649a8 R09: ffff81002ab9dec8
May 25 21:17:38 grover kernel: [ 131.623602] R10: 0000000000000002 R11: 0000000000000206 R12: ffff81002b339f50
May 25 21:17:38 grover kernel: [ 131.647075] R13: ffff81002b339f50 R14: ffff81002e7f2a08 R15: 00002aaaadb67000
May 25 21:17:38 grover kernel: [ 131.670548] FS: 00002aaaad6d8f50(0000) GS:ffffffff80550840(0000) knlGS:0000000000000000
May 25 21:17:38 grover kernel: [ 131.697165] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
May 25 21:17:38 grover kernel: [ 131.716064] CR2: 00002aaaadb59098 CR3: 000000002aafd000 CR4: 00000000000006e0
May 25 21:17:38 grover kernel: [ 131.739537] Process artsd (pid: 5481, threadinfo ffff81002ab9c000, task ffff81002ada97f0)
May 25 21:17:38 grover kernel: [ 131.766439] Stack: ffffffff8016942d 0000000000000000 0000000000000000 ffff81002e7f2a00
May 25 21:17:38 grover kernel: [ 131.792174] ffffffff8016a936 ffff81002d672df0 ffff81002d672e08 ffff81002d672df0
May 25 21:17:38 grover kernel: [ 131.818481] ffff81002e7f2a00 0000000000560600
May 25 21:17:38 grover kernel: [ 131.835065] Call Trace:<ffffffff8016942d>{remove_vm_struct+125} <ffffffff8016a936>{do_munmap+550}
May 25 21:17:38 grover kernel: [ 131.864282] <ffffffff8016b0fd>{sys_munmap+77} <ffffffff8010ead6>{system_call+126}
May 25 21:17:38 grover kernel: [ 131.890928]
May 25 21:17:38 grover kernel: [ 131.898387]
May 25 21:17:38 grover kernel: [ 131.898388] Code: 48 8b 80 98 00 00 00 ff 88 08 01 00 00 c3 66 66 66 90 66 66

2005-05-26 21:05:21

by Lee Revell

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Thu, 2005-05-26 at 13:45 -0700, Andrew Morton wrote:
> "Rafael J. Wysocki" <[email protected]> wrote:
> >
> > n Thursday, 26 of May 2005 09:58, Andrew Morton wrote:
> > > "J.A. Magallon" <[email protected]> wrote:
> > > >
> > > >
> > > > On 05.26, Andrew Morton wrote:
> > > > >
> > > > > (Added alsa-devel to cc)
> > > > >
> > > > > Ed Tomlinson <[email protected]> wrote:
> > > > > >
> > > > > > Got the following when I tried to use sound. Anyone else see problems in alsa land?
> > > > > >
> > > >
> > > > Me too. As beep-media-player ends playing a mp3 track, oops !
> > >
> > > hm, OK, you're also on x86_64. What sound card and driver?
> >
> > I've got the following on a dual-Opteron box with Tyan Thunder K8W (snd_intel8x0):
>
> OK, thanks. I guess we can set this problem aside for now, as this bug
> isn't present in 2.6.12-rc5 (correct?).
>
> I assume the problem is due to one of the ASLA patches in rc5-mm1, but it's
> possible that it lies elsewhere.

See the "Patch avoiding-mmap-fragmentation-fix-2.patch oops" thread.

Lee

2005-05-26 21:09:12

by Chris Wright

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

* Andrew Morton ([email protected]) wrote:
> I assume the problem is due to one of the ASLA patches in rc5-mm1, but it's
> possible that it lies elsewhere. It would be great if someone could fire
> up quilt and do a binary search...

I had started that, but then read a few accounts of backing out
avoiding-mmap-fragmentation-fix-2.patch fixing the problem. Both that
patch and alsa are using vm_private_data. Madness ensues.

2005-05-26 21:41:15

by J.A. Magallon

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1


On 05.26, Andrew Morton wrote:
> "J.A. Magallon" <[email protected]> wrote:
> >
> >
> > On 05.26, Andrew Morton wrote:
> > >
> > > (Added alsa-devel to cc)
> > >
> > > Ed Tomlinson <[email protected]> wrote:
> > > >
> > > > Got the following when I tried to use sound. Anyone else see problems in alsa land?
> > > >
> >
> > Me too. As beep-media-player ends playing a mp3 track, oops !
>
> hm, OK, you're also on x86_64. What sound card and driver?
>

Sorry for the late answer...

No, I'm ia32. Dual Xeon.

--
J.A. Magallon <jamagallon()able!es> \ Software is like sex:
werewolf!able!es \ It's better when it's free
Mandriva Linux release 2006.0 (Cooker) for i586
Linux 2.6.11-jam20 (gcc 4.0.0 (4.0.0-3mdk for Mandriva Linux release 2006.0))


2005-05-26 22:24:54

by Andrea Arcangeli

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Thu, May 26, 2005 at 09:15:26PM +0200, Mikael Pettersson wrote:
> (But please make it conditional on CONFIG_SECCOMP.)

Why should it be conditional? Isn't it faster to read it from the cpu
itself instead of reading it from slow ram? That info is cold in the
cache when the global tlb flush is invoked, while it's always in hot in
the cpu. Perhaps the cpu is dogslow at reading cr4 or what?

Note that the whole mmu_cr4 thing seems pretty flawed if used at
runtime, that's only necessary to initialize new cpus (head.S). So
basically at runtime nobody should read or write to mmu_cr4_features
global variable.

> The code called from __switch_to() would have to set or clear cr4
> locally only. That's easy using write_cr4() and read_cr4().

Agreed.

2005-05-27 02:47:41

by Andrea Arcangeli

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Here an update that passed successfully basic testing verifying task
gets killed with sig11 if rdtsc is used in seccomp mode on x86-64 (i686
still untested but fully symmetric).

From: Andrea Arcangeli <[email protected]>

I believe at least for seccomp it's worth to turn off the tsc, not just for
HT but for the L2 cache too. So it's up to you, either you turn it off
completely (which isn't very nice IMHO) or I recommend to apply this below
patch.

This has been tested successfully on x86-64 against current cogito
repository (i686 compiles so I didn't bother testing ;). People selling
the cpu through cpushare may appreciate this bit for a peace of mind.

There's no way to get any timing info anymore with this applied
(gettimeofday is forbidden of course). The seccomp environment is
completely deterministic so it can't be allowed to get timing info, it has
to be deterministic so in the future I can enable a computing mode that
does a parallel computing for each task with server side transparent
checkpointing and verification that the output is the same from all the 2/3
seller computers for each task, without the buyer even noticing (for now
the verification is left to the buyer client side and there's no
checkpointing, since that would require more kernel changes to track the
dirty bits but it'll be easy to extend once the basic mode is finished).

Eliminating a cold-cache read of the cr4 global variable will save one
cacheline during the tlb flush while making the code per-cpu-safe at the
same time. Thanks to Mikael Pettersson for noticing the tlb flush wasn't
per-cpu-safe.

The global tlb flush can run from irq (IPI calling do_flush_tlb_all) but
it'll be transparent to the switch_to code since the IPI won't make any
change to the cr4 contents from the point of view of the interrupted
code and since it's now all per-cpu stuff, it will not race. So no need
to disable irqs in switch_to slow path.

Signed-off-by: Andrea Arcangeli <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
---

arch/i386/kernel/process.c | 21 +++++++++++++++++++++
arch/x86_64/kernel/process.c | 21 +++++++++++++++++++++
include/linux/seccomp.h | 6 ++++++
3 files changed, 48 insertions(+)

Index: arch/i386/kernel/process.c
===================================================================
--- 3ac9a34948049bff79a2b2ce49c0a3c84e35a748/arch/i386/kernel/process.c (mode:100644)
+++ uncommitted/arch/i386/kernel/process.c (mode:100644)
@@ -561,6 +561,25 @@
}

/*
+ * This function selects if the context switch from prev to next
+ * has to tweak the TSC disable bit in the cr4.
+ */
+static void disable_tsc(struct thread_info *prev,
+ struct thread_info *next)
+{
+ if (unlikely(has_secure_computing(prev) ||
+ has_secure_computing(next))) {
+ /* slow path here */
+ if (has_secure_computing(prev) &&
+ !has_secure_computing(next)) {
+ write_cr4(read_cr4() & ~X86_CR4_TSD);
+ } else if (!has_secure_computing(prev) &&
+ has_secure_computing(next))
+ write_cr4(read_cr4() | X86_CR4_TSD);
+ }
+}
+
+/*
* switch_to(x,yn) should switch tasks from x to y.
*
* We fsave/fwait so that an exception goes off at the right time
@@ -639,6 +658,8 @@
if (unlikely(prev->io_bitmap_ptr || next->io_bitmap_ptr))
handle_io_bitmap(next, tss);

+ disable_tsc(prev_p->thread_info, next_p->thread_info);
+
return prev_p;
}

Index: arch/x86_64/kernel/process.c
===================================================================
--- 3ac9a34948049bff79a2b2ce49c0a3c84e35a748/arch/x86_64/kernel/process.c (mode:100644)
+++ uncommitted/arch/x86_64/kernel/process.c (mode:100644)
@@ -439,6 +439,25 @@
}

/*
+ * This function selects if the context switch from prev to next
+ * has to tweak the TSC disable bit in the cr4.
+ */
+static void disable_tsc(struct thread_info *prev,
+ struct thread_info *next)
+{
+ if (unlikely(has_secure_computing(prev) ||
+ has_secure_computing(next))) {
+ /* slow path here */
+ if (has_secure_computing(prev) &&
+ !has_secure_computing(next)) {
+ write_cr4(read_cr4() & ~X86_CR4_TSD);
+ } else if (!has_secure_computing(prev) &&
+ has_secure_computing(next))
+ write_cr4(read_cr4() | X86_CR4_TSD);
+ }
+}
+
+/*
* This special macro can be used to load a debugging register
*/
#define loaddebug(thread,r) set_debug(thread->debugreg ## r, r)
@@ -556,6 +575,8 @@
}
}

+ disable_tsc(prev_p->thread_info, next_p->thread_info);
+
return prev_p;
}

Index: include/asm-i386/tlbflush.h
===================================================================
--- 3ac9a34948049bff79a2b2ce49c0a3c84e35a748/include/asm-i386/tlbflush.h (mode:100644)
+++ uncommitted/include/asm-i386/tlbflush.h (mode:100644)
@@ -22,16 +22,18 @@
*/
#define __flush_tlb_global() \
do { \
- unsigned int tmpreg; \
+ unsigned int tmpreg, cr4, cr4_orig; \
\
__asm__ __volatile__( \
- "movl %1, %%cr4; # turn off PGE \n" \
+ "movl %%cr4, %2; # turn off PGE \n" \
+ "movl %2, %1; \n" \
+ "andl %3, %1; \n" \
+ "movl %1, %%cr4; \n" \
"movl %%cr3, %0; \n" \
"movl %0, %%cr3; # flush TLB \n" \
"movl %2, %%cr4; # turn PGE back on \n" \
- : "=&r" (tmpreg) \
- : "r" (mmu_cr4_features & ~X86_CR4_PGE), \
- "r" (mmu_cr4_features) \
+ : "=&r" (tmpreg), "=&r" (cr4), "=&r" (cr4_orig) \
+ : "i" (~X86_CR4_PGE) \
: "memory"); \
} while (0)

Index: include/asm-x86_64/tlbflush.h
===================================================================
--- 3ac9a34948049bff79a2b2ce49c0a3c84e35a748/include/asm-x86_64/tlbflush.h (mode:100644)
+++ uncommitted/include/asm-x86_64/tlbflush.h (mode:100644)
@@ -22,16 +22,18 @@
*/
#define __flush_tlb_global() \
do { \
- unsigned long tmpreg; \
+ unsigned long tmpreg, cr4, cr4_orig; \
\
__asm__ __volatile__( \
- "movq %1, %%cr4; # turn off PGE \n" \
+ "movq %%cr4, %2; # turn off PGE \n" \
+ "movq %2, %1; \n" \
+ "andq %3, %1; \n" \
+ "movq %1, %%cr4; \n" \
"movq %%cr3, %0; # flush TLB \n" \
"movq %0, %%cr3; \n" \
"movq %2, %%cr4; # turn PGE back on \n" \
- : "=&r" (tmpreg) \
- : "r" (mmu_cr4_features & ~X86_CR4_PGE), \
- "r" (mmu_cr4_features) \
+ : "=&r" (tmpreg), "=&r" (cr4), "=&r" (cr4_orig) \
+ : "i" (~X86_CR4_PGE) \
: "memory"); \
} while (0)

Index: include/linux/seccomp.h
===================================================================
--- 3ac9a34948049bff79a2b2ce49c0a3c84e35a748/include/linux/seccomp.h (mode:100644)
+++ uncommitted/include/linux/seccomp.h (mode:100644)
@@ -19,6 +19,11 @@
__secure_computing(this_syscall);
}

+static inline int has_secure_computing(struct thread_info *ti)
+{
+ return unlikely(test_ti_thread_flag(ti, TIF_SECCOMP));
+}
+
#else /* CONFIG_SECCOMP */

#if (__GNUC__ > 2)
@@ -28,6 +33,7 @@
#endif

#define secure_computing(x) do { } while (0)
+#define has_secure_computing(x) 0

#endif /* CONFIG_SECCOMP */

2005-05-27 10:29:13

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Thursday, 26 of May 2005 22:45, Andrew Morton wrote:
> "Rafael J. Wysocki" <[email protected]> wrote:
> >
> > n Thursday, 26 of May 2005 09:58, Andrew Morton wrote:
> > > "J.A. Magallon" <[email protected]> wrote:
> > > >
> > > >
> > > > On 05.26, Andrew Morton wrote:
> > > > >
> > > > > (Added alsa-devel to cc)
> > > > >
> > > > > Ed Tomlinson <[email protected]> wrote:
> > > > > >
> > > > > > Got the following when I tried to use sound. Anyone else see problems in alsa land?
> > > > > >
> > > >
> > > > Me too. As beep-media-player ends playing a mp3 track, oops !
> > >
> > > hm, OK, you're also on x86_64. What sound card and driver?
> >
> > I've got the following on a dual-Opteron box with Tyan Thunder K8W (snd_intel8x0):
>
> OK, thanks. I guess we can set this problem aside for now, as this bug
> isn't present in 2.6.12-rc5 (correct?).

Yup.

> I assume the problem is due to one of the ASLA patches in rc5-mm1, but it's
> possible that it lies elsewhere.

Well, yes. Apparently, it goes away if you revert the following patch:

avoiding-mmap-fragmentation-fix-2.patch

Greets,
Rafael


--
- Would you tell me, please, which way I ought to go from here?
- That depends a good deal on where you want to get to.
-- Lewis Carroll "Alice's Adventures in Wonderland"

2005-05-27 17:40:35

by Chen, Kenneth W

[permalink] [raw]
Subject: RE: 2.6.12-rc5-mm1

Rafael J. Wysocki wrote on Friday, May 27, 2005 3:29 AM
> > I assume the problem is due to one of the ASLA patches in rc5-mm1, but it's
> > possible that it lies elsewhere.
>
> Well, yes. Apparently, it goes away if you revert the following patch:
>
> avoiding-mmap-fragmentation-fix-2.patch


avoiding-mmap-fragmentation-fix-2.patch has a major clash using
vm_private_data where alsa is also using. I just posted a patch,
please try that out. Thanks.

http://marc.theaimsgroup.com/?l=linux-mm&m=111721191501940&w=2

- Ken

2005-05-27 21:33:25

by Arnd Bergmann

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Middeweken 25 Mai 2005 22:49, Andrew Morton wrote:
>
> - New Xtensa architecture: Tensilica Xtensa CPU series.

I noticed this has another copy of all the backwards compatibility
syscalls in its arch/*/kernel/syscall.c file. This doesn't make
sense for a new architecture added to the tree.

Chris, are there any existing binaries that rely on your implementations
of old_mmap, sys_fork, sys_vfork, sys_olduname or sys_ipc and need to
work with future kernels? Otherwise, you should probably drop these.
For sys_ipc, you would need to add the subcalls directly to the table,
like parisc does.

Arnd <><

2005-05-27 22:25:56

by J.A. Magallon

[permalink] [raw]
Subject: Kill signed chars !!! [was Re: 2.6.12-rc5-mm1]

... and make gcc4 happy.

On 05.25, Andrew Morton wrote:
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/
>
>
> - Again, if there are patches in here which you think should be merged in
> 2.6.12, please point them out to me.
>

scripts/ is full of mismatches between char* params an signed char* arguments,
and viceversa. gcc4 now complaints loud about this. Patch below deletes all
those 'signed'. Anyways, which was the purpose of declaring 'signed char's
to store text ?

--- linux-2.6.12-rc5-mm1/scripts/kconfig/mconf.c.orig 2005-05-28 00:11:01.000000000 +0200
+++ linux-2.6.12-rc5-mm1/scripts/kconfig/mconf.c 2005-05-28 00:11:07.000000000 +0200
@@ -254,8 +254,8 @@
" USB$ => find all CONFIG_ symbols ending with USB\n"
"\n");

-static signed char buf[4096], *bufptr = buf;
-static signed char input_buf[4096];
+static char buf[4096], *bufptr = buf;
+static char input_buf[4096];
static char filename[PATH_MAX+1] = ".config";
static char *args[1024], **argptr = args;
static int indent;
--- linux-2.6.12-rc5-mm1/scripts/kconfig/confdata.c.orig 2005-05-28 00:11:23.000000000 +0200
+++ linux-2.6.12-rc5-mm1/scripts/kconfig/confdata.c 2005-05-28 00:11:33.000000000 +0200
@@ -27,10 +27,10 @@
NULL,
};

-static char *conf_expand_value(const signed char *in)
+static char *conf_expand_value(const char *in)
{
struct symbol *sym;
- const signed char *src;
+ const char *src;
static char res_value[SYMBOL_MAXLENGTH];
char *dst, name[SYMBOL_MAXLENGTH];

--- linux-2.6.12-rc5-mm1/scripts/kconfig/conf.c.orig 2005-05-28 00:12:18.000000000 +0200
+++ linux-2.6.12-rc5-mm1/scripts/kconfig/conf.c 2005-05-28 00:12:28.000000000 +0200
@@ -31,14 +31,14 @@
static int indent = 1;
static int valid_stdin = 1;
static int conf_cnt;
-static signed char line[128];
+static char line[128];
static struct menu *rootEntry;

static char nohelp_text[] = N_("Sorry, no help available for this option yet.\n");

-static void strip(signed char *str)
+static void strip(char *str)
{
- signed char *p = str;
+ char *p = str;
int l;

while ((isspace(*p)))
--- linux-2.6.12-rc5-mm1/scripts/basic/fixdep.c.orig 2005-05-28 00:00:02.000000000 +0200
+++ linux-2.6.12-rc5-mm1/scripts/basic/fixdep.c 2005-05-28 00:07:27.000000000 +0200
@@ -212,23 +212,23 @@
if (*p == '_')
*p = '/';
else
- *p = tolower((unsigned char)*p);
+ *p = tolower((int)*p);
}
printf(" $(wildcard include/config/%s.h) \\\n", s);
}

-void parse_config_file(signed char *map, size_t len)
+void parse_config_file(char *map, size_t len)
{
int *end = (int *) (map + len);
/* start at +1, so that p can never be < map */
int *m = (int *) map + 1;
- signed char *p, *q;
+ char *p, *q;

for (; m < end; m++) {
- if (*m == INT_CONF) { p = (signed char *) m ; goto conf; }
- if (*m == INT_ONFI) { p = (signed char *) m-1; goto conf; }
- if (*m == INT_NFIG) { p = (signed char *) m-2; goto conf; }
- if (*m == INT_FIG_) { p = (signed char *) m-3; goto conf; }
+ if (*m == INT_CONF) { p = (char *) m ; goto conf; }
+ if (*m == INT_ONFI) { p = (char *) m-1; goto conf; }
+ if (*m == INT_NFIG) { p = (char *) m-2; goto conf; }
+ if (*m == INT_FIG_) { p = (char *) m-3; goto conf; }
continue;
conf:
if (p > map + len - 7)
@@ -291,9 +291,9 @@

void parse_dep_file(void *map, size_t len)
{
- signed char *m = map;
- signed char *end = m + len;
- signed char *p;
+ char *m = map;
+ char *end = m + len;
+ char *p;
char s[PATH_MAX];

p = strchr(m, ':');
--- linux-2.6.12-rc5-mm1/scripts/basic/docproc.c.orig 2005-05-28 00:09:23.000000000 +0200
+++ linux-2.6.12-rc5-mm1/scripts/basic/docproc.c 2005-05-28 00:10:12.000000000 +0200
@@ -52,7 +52,7 @@
FILEONLY *externalfunctions;
FILEONLY *symbolsonly;

-typedef void FILELINE(char * file, signed char * line);
+typedef void FILELINE(char * file, char * line);
FILELINE * singlefunctions;
FILELINE * entity_system;

@@ -148,9 +148,9 @@
* Files are separated by tabs.
*/
void adddep(char * file) { printf("\t%s", file); }
-void adddep2(char * file, signed char * line) { line = line; adddep(file); }
+void adddep2(char * file, char * line) { line = line; adddep(file); }
void noaction(char * line) { line = line; }
-void noaction2(char * file, signed char * line) { file = file; line = line; }
+void noaction2(char * file, char * line) { file = file; line = line; }

/* Echo the line without further action */
void printline(char * line) { printf("%s", line); }
@@ -179,8 +179,8 @@
perror(real_filename);
}
while(fgets(line, MAXLINESZ, fp)) {
- signed char *p;
- signed char *e;
+ char *p;
+ char *e;
if (((p = strstr(line, "EXPORT_SYMBOL_GPL")) != 0) ||
((p = strstr(line, "EXPORT_SYMBOL")) != 0)) {
/* Skip EXPORT_SYMBOL{_GPL} */
@@ -253,7 +253,7 @@
* Call kernel-doc with the following parameters:
* kernel-doc -docbook -function function1 [-function function2]
*/
-void singfunc(char * filename, signed char * line)
+void singfunc(char * filename, char * line)
{
char *vec[200]; /* Enough for specific functions */
int i, idx = 0;
@@ -290,7 +290,7 @@
void parse_file(FILE *infile)
{
char line[MAXLINESZ];
- signed char * s;
+ char * s;
while(fgets(line, MAXLINESZ, infile)) {
if (line[0] == '!') {
s = line + 2;
--- linux-2.6.12-rc5-mm1/scripts/basic/split-include.c.orig 2005-05-28 00:08:02.000000000 +0200
+++ linux-2.6.12-rc5-mm1/scripts/basic/split-include.c 2005-05-28 00:08:52.000000000 +0200
@@ -104,7 +104,7 @@
/* Read config lines. */
while (fgets(line, buffer_size, fp_config))
{
- const signed char * str_config;
+ const char * str_config;
int is_same;
int itarget;


--
J.A. Magallon <jamagallon()able!es> \ Software is like sex:
werewolf!able!es \ It's better when it's free
Mandriva Linux release 2006.0 (Cooker) for i586
Linux 2.6.11-jam20 (gcc 4.0.0 (4.0.0-3mdk for Mandriva Linux release 2006.0))


2005-05-27 22:33:50

by J.A. Magallon

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1


On 05.27, Chen, Kenneth W wrote:
> Rafael J. Wysocki wrote on Friday, May 27, 2005 3:29 AM
> > > I assume the problem is due to one of the ASLA patches in rc5-mm1, but it's
> > > possible that it lies elsewhere.
> >
> > Well, yes. Apparently, it goes away if you revert the following patch:
> >
> > avoiding-mmap-fragmentation-fix-2.patch
>
>
> avoiding-mmap-fragmentation-fix-2.patch has a major clash using
> vm_private_data where alsa is also using. I just posted a patch,
> please try that out. Thanks.
>
> http://marc.theaimsgroup.com/?l=linux-mm&m=111721191501940&w=2
>

This makes my oopses go away in beep-media-player.
Thanks!

--
J.A. Magallon <jamagallon()able!es> \ Software is like sex:
werewolf!able!es \ It's better when it's free
Mandriva Linux release 2006.0 (Cooker) for i586
Linux 2.6.11-jam20 (gcc 4.0.0 (4.0.0-3mdk for Mandriva Linux release 2006.0))


2005-05-27 23:41:34

by Jesper Juhl

[permalink] [raw]
Subject: Re: Kill signed chars !!! [was Re: 2.6.12-rc5-mm1]

On Fri, 27 May 2005, J.A. Magallon wrote:

> ... and make gcc4 happy.
>
> On 05.25, Andrew Morton wrote:
> >
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.6.12-rc5-mm1/
> >
> >
> > - Again, if there are patches in here which you think should be merged in
> > 2.6.12, please point them out to me.
> >
>
> scripts/ is full of mismatches between char* params an signed char* arguments,
> and viceversa. gcc4 now complaints loud about this. Patch below deletes all
> those 'signed'. Anyways, which was the purpose of declaring 'signed char's
> to store text ?

Not commenting on your specific patch here, just the bit about what the
point is of using explicitly signed chars. C doesn't define if chars are
signed or unsigned by default, that's up to the implementation. So if it
matters to you that chars should be either signed or unsigned you have to
explicitly specify them as signed or unsigned.
I haven't read the code you change, not the patch, so I don't know if it's
an issue, but there are certainly arguments for explicitly specifying
signed char or unsigned char in C code.


--
Jesper Juhl

2005-05-28 07:07:23

by Christoph Hellwig

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

On Fri, May 27, 2005 at 11:13:19PM +0200, Arnd Bergmann wrote:
> On Middeweken 25 Mai 2005 22:49, Andrew Morton wrote:
> >
> > - New Xtensa architecture: Tensilica Xtensa CPU series.
>
> I noticed this has another copy of all the backwards compatibility
> syscalls in its arch/*/kernel/syscall.c file. This doesn't make
> sense for a new architecture added to the tree.
>
> Chris, are there any existing binaries that rely on your implementations
> of old_mmap, sys_fork, sys_vfork, sys_olduname or sys_ipc and need to
> work with future kernels? Otherwise, you should probably drop these.
> For sys_ipc, you would need to add the subcalls directly to the table,
> like parisc does.

We should do that either way. If people have existing binaries relying
on broken thing before submitting ports for review it's their fault.

2005-05-29 14:26:25

by Adrian Bunk

[permalink] [raw]
Subject: 2.6.12-rc5-mm1: fork connector doesn't compile with gcc 2.95

The following compile breakage with gcc 2.95 is caused by
fork-connector-send-status-to-userspace.patch:

<-- snip -->

...
CC drivers/connector/cn_fork.o
In file included from drivers/connector/cn_fork.c:36:
include/linux/cn_fork.h:58: warning: unnamed struct/union that defines
no instances
include/linux/cn_fork.h:60: warning: unnamed struct/union that defines
no instances
drivers/connector/cn_fork.c: In function `fork_connector':
drivers/connector/cn_fork.c:75: structure has no member named `ppid'
drivers/connector/cn_fork.c:76: structure has no member named `ptid'
drivers/connector/cn_fork.c:77: structure has no member named `cpid'
drivers/connector/cn_fork.c:78: structure has no member named `ctid'
drivers/connector/cn_fork.c: In function `cn_fork_send_status':
drivers/connector/cn_fork.c:110: structure has no member named `status'
make[2]: *** [drivers/connector/cn_fork.o] Error 1

<-- snip -->

cu
Adrian

--

"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

2005-05-29 14:38:33

by Adrian Bunk

[permalink] [raw]
Subject: 2.6.12-rc5-mm1: drivers/char/tpm/ compile errors with gcc 2.95

It seems it's fix-tpm-driver-sysfs-owernship-changes.patch that causes
the following compile errors with gcc 2.95:

<-- snip -->

...
CC drivers/char/tpm/tpm_nsc.o
drivers/char/tpm/tpm_nsc.c:231: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_nsc.c:232: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_nsc.c:233: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_nsc.c:234: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_nsc.c:254: unknown field `fops' specified in initializer
drivers/char/tpm/tpm_nsc.c:254: warning: missing braces around initializer
drivers/char/tpm/tpm_nsc.c:254: warning: (near initialization for `tpm_nsc.miscdev')
drivers/char/tpm/tpm_nsc.c:254: warning: initialization makes integer from pointer without a cast
make[3]: *** [drivers/char/tpm/tpm_nsc.o] Error 1

<-- snip -->

...
CC drivers/char/tpm/tpm_atmel.o
drivers/char/tpm/tpm_atmel.c:130: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_atmel.c:131: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_atmel.c:132: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_atmel.c:133: warning: initialization from incompatible pointer type
drivers/char/tpm/tpm_atmel.c:153: unknown field `fops' specified in initializer
drivers/char/tpm/tpm_atmel.c:153: warning: missing braces around initializer
drivers/char/tpm/tpm_atmel.c:153: warning: (near initialization for `tpm_atmel.miscdev')
drivers/char/tpm/tpm_atmel.c:153: warning: initialization makes integer from pointer without a cast
make[3]: *** [drivers/char/tpm/tpm_atmel.o] Error 1

<-- snip -->

cu
Adrian

--

"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

2005-05-29 14:38:38

by Adrian Bunk

[permalink] [raw]
Subject: 2.6.12-rc5-mm1: drivers/dlm/: compile error with gcc 2.95

<-- snip -->

...
CC drivers/dlm/lock.o
drivers/dlm/lock.c: In function `receive_grant':
drivers/dlm/lock.c:2658: parse error before `;'
drivers/dlm/lock.c: In function `receive_bast':
drivers/dlm/lock.c:2685: parse error before `;'
drivers/dlm/lock.c: In function `receive_request_reply':
drivers/dlm/lock.c:2757: parse error before `;'
drivers/dlm/lock.c:2765: parse error before `;'
drivers/dlm/lock.c: In function `receive_convert_reply':
drivers/dlm/lock.c:2883: parse error before `;'
drivers/dlm/lock.c:2890: parse error before `;'
drivers/dlm/lock.c: In function `receive_unlock_reply':
drivers/dlm/lock.c:2930: parse error before `;'
drivers/dlm/lock.c:2937: parse error before `;'
drivers/dlm/lock.c: In function `receive_cancel_reply':
drivers/dlm/lock.c:2977: parse error before `;'
drivers/dlm/lock.c:2984: parse error before `;'
drivers/dlm/lock.c: In function `receive_lookup_reply':
drivers/dlm/lock.c:3001: parse error before `;'
drivers/dlm/lock.c:3007: parse error before `;'
drivers/dlm/lock.c: In function `purge_queue':
drivers/dlm/lock.c:3325: parse error before `;'
make[2]: *** [drivers/dlm/lock.o] Error 1

<-- snip -->

cu
Adrian

--

"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

2005-05-29 14:45:57

by Adrian Bunk

[permalink] [raw]
Subject: 2.6.12-rc5-mm1: drivers/media/dvb/dvb-usb/a800.c compile error

On Wed, May 25, 2005 at 01:49:33PM -0700, Andrew Morton wrote:
>...
> Changes since 2.6.12-rc4-mm2:
>...
> +add-generalized-dvb-usb-driver.patch
>...
> DVB updates
>...

It seems this patch is responsible for the following compile error with
gcc 2.95:

<-- snip -->

...
CC drivers/media/dvb/dvb-usb/a800.o
In file included from drivers/media/dvb/dvb-usb/dibusb.h:15,
from drivers/media/dvb/dvb-usb/a800.c:16:
drivers/media/dvb/dvb-usb/dvb-usb.h:196: field `devices' has incomplete type
...
make[4]: *** [drivers/media/dvb/dvb-usb/a800.o] Error 1

<-- snip -->


cu
Adrian

--

"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

2005-05-29 14:47:42

by Matthias-Christian Ott

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/dlm/: compile error with gcc 2.95

Adrian Bunk wrote:
> <-- snip -->
>
> ...
> CC drivers/dlm/lock.o
> [ .. ]
> make[2]: *** [drivers/dlm/lock.o] Error 1
>
> <-- snip -->
>
> cu
> Adrian
>
The gcc 2.95 is deprecated, so I think this is not a real problem.

Matthias-Christian Ott

2005-05-29 15:01:24

by Adrian Bunk

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/dlm/: compile error with gcc 2.95

On Sun, May 29, 2005 at 04:43:55PM +0200, Matthias-Christian Ott wrote:
> Adrian Bunk wrote:
> ><-- snip -->
> >
> >...
> > CC drivers/dlm/lock.o
> >[ .. ]
> >make[2]: *** [drivers/dlm/lock.o] Error 1
> >
> ><-- snip -->
> >
> >cu
> >Adrian
> >
> The gcc 2.95 is deprecated, so I think this is not a real problem.

Who said gcc 2.95 was deprecated? The documentation in the kernel even
lists gcc 2.95 as the recommended compiler.

gcc 2.95 is definitely still supported, and (at least on i386) code that
doesn't compile with gcc 2.95 is not suitable for Linus' tree.

> Matthias-Christian Ott

cu
Adrian

--

"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

2005-05-29 15:12:35

by Adrian Bunk

[permalink] [raw]
Subject: 2.6.12-rc5-mm1: drivers/usb/atm/speedtch.c: gcc 2.95 compile error

The following compile error with gcc 2.95 seems to be caused by
broken-out/gregkh-usb-usb-usbatm-{1,2}.patch:

<-- snip -->

...
CC drivers/usb/atm/speedtch.o
drivers/usb/atm/speedtch.c: In function `speedtch_check_status':
drivers/usb/atm/speedtch.c:447: parse error before `;'
drivers/usb/atm/speedtch.c:456: parse error before `;'
drivers/usb/atm/speedtch.c:463: parse error before `;'
drivers/usb/atm/speedtch.c: In function `speedtch_status_poll':
drivers/usb/atm/speedtch.c:507: parse error before `;'
drivers/usb/atm/speedtch.c: In function `speedtch_handle_int':
drivers/usb/atm/speedtch.c:550: parse error before `;'
drivers/usb/atm/speedtch.c:552: parse error before `;'
make[3]: *** [drivers/usb/atm/speedtch.o] Error 1

<-- snip -->

cu
Adrian

--

"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

2005-05-30 07:45:22

by Duncan Sands

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/usb/atm/speedtch.c: gcc 2.95 compile error

On Sun, 2005-05-29 at 17:12 +0200, Adrian Bunk wrote:
> The following compile error with gcc 2.95 seems to be caused by
> broken-out/gregkh-usb-usb-usbatm-{1,2}.patch:
>
> <-- snip -->
>
> ...
> CC drivers/usb/atm/speedtch.o
> drivers/usb/atm/speedtch.c: In function `speedtch_check_status':
> drivers/usb/atm/speedtch.c:447: parse error before `;'
> drivers/usb/atm/speedtch.c:456: parse error before `;'
> drivers/usb/atm/speedtch.c:463: parse error before `;'
> drivers/usb/atm/speedtch.c: In function `speedtch_status_poll':
> drivers/usb/atm/speedtch.c:507: parse error before `;'
> drivers/usb/atm/speedtch.c: In function `speedtch_handle_int':
> drivers/usb/atm/speedtch.c:550: parse error before `;'
> drivers/usb/atm/speedtch.c:552: parse error before `;'
> make[3]: *** [drivers/usb/atm/speedtch.o] Error 1
>
> <-- snip -->

Hi Adrian, it looks like gcc 2.95 doesn't like this kind of macro

#define atm_info(instance, format, arg...) \
atm_printk(KERN_INFO, instance , format , ## arg)

being called with only two arguments. I don't know what
the best fix is, but this does the trick:

Signed-off-by: Duncan Sands <[email protected]>

--- Linux/drivers/usb/atm/speedtch.c.orig 3 May 2005 07:30:42 -0000 1.58
+++ Linux/drivers/usb/atm/speedtch.c 30 May 2005 07:37:45 -0000
@@ -444,7 +444,7 @@ static void speedtch_check_status(struct
case 0:
if (atm_dev->signal != ATM_PHY_SIG_LOST) {
atm_dev->signal = ATM_PHY_SIG_LOST;
- atm_info(usbatm, "ADSL line is down\n");
+ atm_info(usbatm, "%s\n", "ADSL line is down");
/* It'll never resync again unless we ask it to... */
ret = speedtch_start_synchro(instance);
}
@@ -453,14 +453,14 @@ static void speedtch_check_status(struct
case 0x08:
if (atm_dev->signal != ATM_PHY_SIG_UNKNOWN) {
atm_dev->signal = ATM_PHY_SIG_UNKNOWN;
- atm_info(usbatm, "ADSL line is blocked?\n");
+ atm_info(usbatm, "%s\n", "ADSL line is blocked?");
}
break;

case 0x10:
if (atm_dev->signal != ATM_PHY_SIG_LOST) {
atm_dev->signal = ATM_PHY_SIG_LOST;
- atm_info(usbatm, "ADSL line is synchronising\n");
+ atm_info(usbatm, "%s\n", "ADSL line is synchronising");
}
break;

@@ -504,7 +504,7 @@ static void speedtch_status_poll(unsigne
if (instance->poll_delay < MAX_POLL_DELAY)
mod_timer(&instance->status_checker.timer, jiffies + msecs_to_jiffies(instance->poll_delay));
else
- atm_warn(instance->usbatm, "Too many failures - disabling line status polling\n");
+ atm_warn(instance->usbatm, "%s\n", "Too many failures - disabling line status polling");
}

static void speedtch_resubmit_int(unsigned long data)
@@ -547,9 +547,9 @@ static void speedtch_handle_int(struct u

if ((count == 6) && !memcmp(up_int, instance->int_data, 6)) {
del_timer(&instance->status_checker.timer);
- atm_info(usbatm, "DSL line goes up\n");
+ atm_info(usbatm, "%s\n", "DSL line goes up");
} else if ((count == 6) && !memcmp(down_int, instance->int_data, 6)) {
- atm_info(usbatm, "DSL line goes down\n");
+ atm_info(usbatm, "%s\n", "DSL line goes down");
} else {
int i;



2005-05-30 08:09:11

by Andrew Morton

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/usb/atm/speedtch.c: gcc 2.95 compile error

Duncan Sands <[email protected]> wrote:
>
> Hi Adrian, it looks like gcc 2.95 doesn't like this kind of macro
>
> #define atm_info(instance, format, arg...) \
> atm_printk(KERN_INFO, instance , format , ## arg)
>
> being called with only two arguments. I don't know what
> the best fix is, but this does the trick:

Nope. There's a bug in gcc-2.95.x macro expansion, and here it bites us in
atm_printk():

printk(level "ATM dev %d: " format , (instance)->atm_dev->number, ## arg)

the workaround is to add a space before that final comma:

diff -puN drivers/usb/atm/usbatm.h~a drivers/usb/atm/usbatm.h
--- 25/drivers/usb/atm/usbatm.h~a 2005-05-30 01:02:51.000000000 -0700
+++ 25-akpm/drivers/usb/atm/usbatm.h 2005-05-30 01:03:08.000000000 -0700
@@ -62,7 +62,8 @@

/* FIXME: move to dev_* once ATM is driver model aware */
#define atm_printk(level, instance, format, arg...) \
- printk(level "ATM dev %d: " format , (instance)->atm_dev->number, ## arg)
+ printk(level "ATM dev %d: " format , \
+ (instance)->atm_dev->number , ## arg)

#define atm_err(instance, format, arg...) \
atm_printk(KERN_ERR, instance , format , ## arg)
_

2005-05-30 08:16:08

by Duncan Sands

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/usb/atm/speedtch.c: gcc 2.95 compile error

> > Hi Adrian, it looks like gcc 2.95 doesn't like this kind of macro
> >
> > #define atm_info(instance, format, arg...) \
> > atm_printk(KERN_INFO, instance , format , ## arg)
> >
> > being called with only two arguments. I don't know what
> > the best fix is, but this does the trick:
>
> Nope. There's a bug in gcc-2.95.x macro expansion, and here it bites us in
> atm_printk():
>
> printk(level "ATM dev %d: " format , (instance)->atm_dev->number, ## arg)
>
> the workaround is to add a space before that final comma:

Yes, this fixes the problem.

Thanks a lot!

Duncan.

2005-05-30 08:38:05

by Patrick Boettcher

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/media/dvb/dvb-usb/a800.c compile error

Hi,

On Sun, 29 May 2005, Adrian Bunk wrote:
> It seems this patch is responsible for the following compile error with
> gcc 2.95:
> <-- snip -->
>
> ...
> CC drivers/media/dvb/dvb-usb/a800.o
> In file included from drivers/media/dvb/dvb-usb/dibusb.h:15,
> from drivers/media/dvb/dvb-usb/a800.c:16:
> drivers/media/dvb/dvb-usb/dvb-usb.h:196: field `devices' has incomplete type
> ...
> make[4]: *** [drivers/media/dvb/dvb-usb/a800.o] Error 1
>
> <-- snip -->

The attached patch solves the problem by adding a '0' to the array
declaration. I don't know if this is ethically correct, but I saw it in
some v4l-code, so I assume it is.

Signed-off-by: Patrick Boettcher <[email protected]>


Thanks for reporting,
Patrick.

--
Mail: [email protected]
WWW: http://www.wi-bw.tfh-wildau.de/~pboettch/


Attachments:
dvb-usb-incomplete-type-gcc-2.95-2.6.12-rc5-mm1.patch (569.00 B)

2005-05-30 09:14:36

by Johannes Stezenbach

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/media/dvb/dvb-usb/a800.c compile error

On Mon, May 30, 2005 at 10:29:26AM +0200, Patrick Boettcher wrote:
> On Sun, 29 May 2005, Adrian Bunk wrote:
> >It seems this patch is responsible for the following compile error with
> >gcc 2.95:
...
> The attached patch solves the problem by adding a '0' to the array
> declaration. I don't know if this is ethically correct, but I saw it in
> some v4l-code, so I assume it is.
...
> - struct dvb_usb_device_description devices[];
> + struct dvb_usb_device_description devices[0];

That can't work in this context. Did you even try to compile?

.../dibusb-mb.c:178: warning: excess elements in array initializer
.../dibusb-mb.c:178: warning: (near initialization for `dibusb1_1_properties.devices')

The solution below wastes a few 100 bytes of space, but I think we can
live with that. The same thing was done in dvb-pll.h.
---
gcc-2.95 compile fix.

Signed-off-by: Patrick Boettcher <[email protected]>
Signed-off-by: Johannes Stezenbach <[email protected]>

Index: linux/drivers/media/dvb/dvb-usb/dvb-usb.h
===================================================================
RCS file: /cvs/linuxtv/dvb-kernel/linux/drivers/media/dvb/dvb-usb/dvb-usb.h,v
retrieving revision 1.16
diff -u -p -r1.16 dvb-usb.h
--- linux/drivers/media/dvb/dvb-usb/dvb-usb.h 2 May 2005 12:48:01 -0000 1.16
+++ linux/drivers/media/dvb/dvb-usb/dvb-usb.h 30 May 2005 09:09:30 -0000
@@ -193,7 +193,7 @@ struct dvb_usb_properties {
} urb;

int num_device_descs;
- struct dvb_usb_device_description devices[];
+ struct dvb_usb_device_description devices[8];
};


2005-05-30 09:42:04

by Patrick Boettcher

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1: drivers/media/dvb/dvb-usb/a800.c compile error

Hi,

On Mon, 30 May 2005, Johannes Stezenbach wrote:
>> - struct dvb_usb_device_description devices[];
>> + struct dvb_usb_device_description devices[0];
>
> That can't work in this context. Did you even try to compile?

Only with gcc-2.95 and it worked...

I guess I should have tried it with gcc-3.x before sending, sorry.

regards,
Patrick.

2005-05-30 13:57:11

by Stefano Rivoir

[permalink] [raw]
Subject: Re: 2.6.12-rc5-mm1

Alle 22:49, mercoled? 25 maggio 2005, Andrew Morton ha scritto:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.12-rc5/2.

Hi all,

I have 2 problems with 2.6.12-rc5-mm1:

1. glxinfo causes a bad error:

May 30 15:02:41 nbsteu kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000000
May 30 15:02:41 nbsteu kernel: printing eip:
May 30 15:02:41 nbsteu kernel: c013e16d
May 30 15:02:41 nbsteu kernel: *pde = 00000000
May 30 15:02:41 nbsteu kernel: Oops: 0000 [#1]
May 30 15:02:41 nbsteu kernel: PREEMPT
May 30 15:02:41 nbsteu kernel: Modules linked in: radeon drm sd_mod scsi_mod lp hotkey fan button ide_cd 8250_pnp parport_pc parport floppy pcspkr sk98lin yenta_socket rsrc_nonstatic pcmcia_core ehci_hcd usbhid ohci_hcd snd_intel8x0m 8250_pci 8250 serial_core ohci1394 ieee1394 sis_agp agpgart vfat fat video thermal processor ac battery evdev snd_intel8x0 snd_ac97_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore snd_page_alloc skge
May 30 15:02:41 nbsteu kernel: CPU: 0
May 30 15:02:41 nbsteu kernel: EIP: 0060:[kfree+45/96] Not tainted VLI
May 30 15:02:41 nbsteu kernel: EFLAGS: 00210002 (2.6.12-rc5-mm1)
May 30 15:02:41 nbsteu kernel: EIP is at kfree+0x2d/0x60
May 30 15:02:41 nbsteu kernel: eax: 00000000 ebx: db7ae720 ecx: db7ae400 edx: c1000000
May 30 15:02:41 nbsteu kernel: esi: b7f4f000 edi: 00200286 ebp: b7f4f000 esp: d203bf30
May 30 15:02:41 nbsteu kernel: ds: 007b es: 007b ss: 0068
May 30 15:02:41 nbsteu kernel: Process glxinfo (pid: 4809, threadinfo=d203a000 task=d4ea1030)
May 30 15:02:41 nbsteu kernel: Stack: db7ae720 00000000 de567800 e1a2a736 c0142ee2 b7f3d000 b7f4f000 d27a3ee0
May 30 15:02:41 nbsteu kernel: db7ae680 b7f4f000 de567800 d203a000 d350b180 d27a3438 d27a3438 c0145f36
May 30 15:02:41 nbsteu kernel: 00000000 de780300 de780300 c01476ea b7f4f000 b7f4d000 c0147a7a b7f4d000
May 30 15:02:41 nbsteu kernel: Call Trace:
May 30 15:02:41 nbsteu kernel: [pg0+560449334/1069876224] drm_vm_shm_close+0x216/0x2c0 [drm]
May 30 15:02:41 nbsteu kernel: [free_pgtables+98/128] free_pgtables+0x62/0x80
May 30 15:02:41 nbsteu kernel: [remove_vm_struct+118/128] remove_vm_struct+0x76/0x80
May 30 15:02:41 nbsteu kernel: [unmap_vma_list+26/48] unmap_vma_list+0x1a/0x30
May 30 15:02:41 nbsteu kernel: [do_munmap+250/304] do_munmap+0xfa/0x130
May 30 15:02:41 nbsteu kernel: [sys_munmap+64/112] sys_munmap+0x40/0x70
May 30 15:02:41 nbsteu kernel: [sysenter_past_esp+84/117] sysenter_past_esp+0x54/0x75
May 30 15:02:41 nbsteu kernel: Code: 0c 85 c0 89 74 24 04 89 c6 89 1c 24 89 7c 24 08 74 2a 9c 5f fa 8b 15 50 db 39 c0 8d 80 00 00 00 40 c1 e8 0c c1 e0 05 8b 44 02 18 <8b> 18 8b 13 3b 53 04 73 17 89 74 93 10 ff 03 57 9d 8b 1c 24 8b

2. ASUS ACPI is not working anymore (this since some -mm ago), modprobe asus_acpi
fails to load saying a generic "No such device"

Note that both 1. and 2. work on plain 2.6.12-rc5, and 1. was fine in
2.6.12-rc4-mm*

Attached lspci-v and .config

Bye.

--
Stefano Rivoir


Attachments:
(No filename) (2.96 kB)
.config.kernel.2.6.12-rc5-mm1 (30.66 kB)
lspciv (5.47 kB)
Download all attachments

2005-05-30 19:51:09

by Adrian Bunk

[permalink] [raw]
Subject: [-mm patch] drivers/message/i2o/device.c: i2o_parm_issue has to be global

On Wed, May 25, 2005 at 01:49:33PM -0700, Andrew Morton wrote:
>...
> Changes since 2.6.12-rc4-mm2:
>...
> +i2o-first-code-cleanup-of-spare-warnings-and-unused.patch
>...
> i2o driver updates
>...


This patch makes i2o_parm_issue static resulting in the following
compile error in the static case (the modular case doesn't give a
compile error, but EXPORT_SYMBOL'ing a static functions is equally
wrong):

<-- snip -->

...
LD .tmp_vmlinux1
drivers/built-in.o(.text+0x691d55): In function `i2o_cfg_parms':
: undefined reference to `i2o_parm_issue'
make: *** [.tmp_vmlinux1] Error 1

<-- snip -->


This patch turns i2o_parm_issue back into a global function.

Signed-off-by: Adrian Bunk <[email protected]>

--- linux-2.6.12-rc5-mm1-full/drivers/message/i2o/device.c.old 2005-05-30 21:17:10.000000000 +0200
+++ linux-2.6.12-rc5-mm1-full/drivers/message/i2o/device.c 2005-05-30 21:17:21.000000000 +0200
@@ -443,8 +443,8 @@
* Note that the minimum sized reslist is 8 bytes and contains
* ResultCount, ErrorInfoSize, BlockStatus and BlockSize.
*/
-static int i2o_parm_issue(struct i2o_device *i2o_dev, int cmd, void *oplist,
- int oplen, void *reslist, int reslen)
+int i2o_parm_issue(struct i2o_device *i2o_dev, int cmd, void *oplist,
+ int oplen, void *reslist, int reslen)
{
struct i2o_message __iomem *msg;
u32 m;

2005-05-31 12:00:23

by Hirokazu Takata

[permalink] [raw]
Subject: [PATCH 2.6.12-rc5-mm1] m32r: Insert set_tsk_need_resched() to cpu_idle() (was Re: 2.6.12-rc5-mm1)

From: Andrew Morton <[email protected]>
Subject: 2.6.12-rc5-mm1
Date: Wed, 25 May 2005 13:49:33 -0700
> - The CPU scheduler is probably busted on the less-common architectures.
> For now, those architectures will need to emulate
> sched-remove-set_tsk_need_resched-from-init_idle-v2-ia64-fix.patch

Here is a patch for m32r.
Please apply.

Thanks,

Signed-off-by: Hirokazu Takata <[email protected]>
---

diff -ruNp a/arch/m32r/kernel/process.c b/arch/m32r/kernel/process.c
--- a/arch/m32r/kernel/process.c 2005-03-07 14:10:21.000000000 +0900
+++ b/arch/m32r/kernel/process.c 2005-05-31 10:49:03.000000000 +0900
@@ -94,6 +94,8 @@ static void poll_idle (void)
*/
void cpu_idle (void)
{
+ set_tsk_need_resched(current);
+
/* endless idle loop with no priority at all */
while (1) {
while (!need_resched()) {


--
Hirokazu Takata <[email protected]>
Linux/M32R Project: http://www.linux-m32r.org/