[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64 and armhf
This bug was fixed in the package linux - 3.16.0-22.29
---
linux (3.16.0-22.29) utopic; urgency=low
[ Tim Gardner ]
* Release Tracking Bug
- LP: #1379321
[ Andrew Morton ]
* SAUCE: (no-up) mm-introduce-a-general-rcu-get_user_pages_fast-fix
- LP: #1309221
* SAUCE: (no-up) arm64-mm-enable-rcu-fast_gup-checkpatch-fixes
- LP: #1309221
[ Andy Whitcroft ]
* [Config] CONFIG_PATA_MACIO=y
- LP: #1378894
* [Config] enable cloud tools on i386
- LP: #1367399
* SAUCE: scsi: hyper-v storsvc switch up to SPC-3
- LP: #1354397
* SAUCE: powerpc -- fix mm/slice.c switch include to linux/hugetlb.h
[ dann frazier ]
* [Config] CONFIG_HAVE_GENERIC_RCU_GUP=y
- LP: #1309221
[ Feng Kan ]
* SAUCE: (no-up) power: reset: Add generic SYSCON register mapped reset
- LP: #1284433
* SAUCE: (no-up) arm64: dts: Add X-Gene reboot driver dts node
- LP: #1284433
[ Ian Munsie ]
* SAUCE: (no-up) powerpc/cell: Move spu_handle_mm_fault() out of cell platform
* SAUCE: (no-up) powerpc/cell: Move data segment faulting code out of cell
platform
* SAUCE: (no-up) powerpc/cell: Make spu_flush_all_slbs() generic
* SAUCE: (no-up) powerpc/msi: Improve IRQ bitmap allocator
* SAUCE: (no-up) powerpc/mm: Export mmu_kernel_ssize and mmu_linear_psize
* SAUCE: (no-up) powerpc/powernv: Split out set MSI IRQ chip code
* SAUCE: (no-up) cxl: Add new header for call backs and structs
* SAUCE: (no-up) powerpc/powerpc: Add new PCIe functions for allocating cxl
interrupts
* SAUCE: (no-up) powerpc/mm: Add new hash_page_mm()
* SAUCE: (no-up) powerpc/opal: Add PHB to cxl mode call
* SAUCE: (no-up) powerpc/mm: Add hooks for cxl
* SAUCE: (no-up) cxl: Add base builtin support
* SAUCE: (no-up) cxl: Driver code for powernv PCIe based cards for userspace
access
* SAUCE: (no-up) cxl: Add userspace header file
* SAUCE: (no-up) cxl: Add driver to Kbuild and Makefiles
* SAUCE: (no-up) cxl: Add documentation for userspace APIs
* SAUCE: (no-up) cxl: Fix afu_read() not doing finish_wait() on signal or
non-blocking
[ John Johansen ]
* SAUCE: Revert: fix: only allow a single threaded process to ...
- LP: #1371310
[ Steve Capper ]
* SAUCE: (no-up) mm: introduce a general RCU get_user_pages_fast()
- LP: #1309221
* SAUCE: (no-up) arm: mm: introduce special ptes for LPAE
- LP: #1309221
* SAUCE: (no-up) arm: mm: enable HAVE_RCU_TABLE_FREE logic
- LP: #1309221
* SAUCE: (no-up) arm: mm: enable RCU fast_gup
- LP: #1309221
* SAUCE: (no-up) arm64: mm: enable HAVE_RCU_TABLE_FREE logic
- LP: #1309221
* SAUCE: (no-up) arm64: mm: enable RCU fast_gup
- LP: #1309221
[ Tim Gardner ]
* SAUCE: Added bnx2x/bnx2x-e1-7.8.19.0.fw
- LP: #1378491
* [Config] CONFIG_CXL=m
* [Config] CONFIG_POWER_RESET_SYSCON=y for arm64
* SAUCE: (no-up) Restrict CONFIG_POWER_RESET_SYSCON to arm64 only
[ Upstream Kernel Changes ]
* powerpc: implement vmemmap_list_free()
- LP: #1378413
* powerpc: implement vmemmap_remove_mapping() for BOOK3S
- LP: #1378413
* powerpc: implement vmemmap_free()
- LP: #1378413
* powerpc: start loop at section start of start in vmemmap_populated()
- LP: #1378413
* ARM: 8108/1: mm: Introduce {pte,pmd}_isset and {pte,pmd}_isclear
* ARM: 8109/1: mm: Modify pte_write and pmd_write logic for LPAE
* seccomp: create internal mode-setting function
- LP: #1379020
* seccomp: extract check/assign mode helpers
- LP: #1379020
* seccomp: split mode setting routines
- LP: #1379020
* seccomp: add seccomp syscall
- LP: #1379020
* ARM: add seccomp syscall
- LP: #1379020
* MIPS: add seccomp syscall
- LP: #1379020
* sched: move no_new_privs into new atomic flags
- LP: #1379020
* seccomp: split filter prep from check and apply
- LP: #1379020
* seccomp: introduce writer locking
- LP: #1379020
* seccomp: allow mode setting across threads
- LP: #1379020
* seccomp: implement SECCOMP_FILTER_FLAG_TSYNC
- LP: #1379020
* seccomp: Replace BUG(!spin_is_locked()) with assert_spin_lock
- LP: #1379020
-- Tim Gardner tim.gard...@canonical.com Wed, 08 Oct 2014 07:35:30 -0400
** Changed in: linux (Ubuntu Utopic)
Status: In Progress = Fix Released
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1309221
Title:
transparent hugepages flaky on arm64 and armhf
Status in “linux” package in Ubuntu:
Fix Released
Status in “linux” source package in Trusty:
Fix Released
Status in “linux” source package in Utopic:
Fix Released
Bug description:
linux 3.13.0-24.46
transparent hugepages, which are configured on by default in Ubuntu,
are flaky on arm64. The most evident system is that go processes
reliably hang or crash. This has been observed with both building
juju-core with gccgo, and running the resulting juju binaries.
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64 and armhf
** Summary changed: - transparent hugepages flaky on arm64 + transparent hugepages flaky on arm64 and armhf -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 and armhf Status in “linux” package in Ubuntu: Confirmed Status in “linux” source package in Trusty: Fix Released Status in “linux” source package in Utopic: Confirmed Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64 and armhf
** Changed in: linux (Ubuntu Utopic) Status: Confirmed = In Progress -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 and armhf Status in “linux” package in Ubuntu: In Progress Status in “linux” source package in Trusty: Fix Released Status in “linux” source package in Utopic: In Progress Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
** Changed in: linux (Ubuntu Utopic) Status: Fix Committed = Confirmed -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Confirmed Status in “linux” source package in Trusty: Fix Released Status in “linux” source package in Utopic: Confirmed Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
This bug was fixed in the package linux - 3.13.0-27.50 --- linux (3.13.0-27.50) trusty; urgency=low [ Brad Figg ] * Revert rtlwifi: Set the link state linux (3.13.0-27.49) trusty; urgency=low [ Brad Figg ] * Revert SAUCE: (no-up) HID: rmi: do not stop the device at the end of probe * Revert SAUCE: (no-up) HID: rmi: introduce RMI driver for Synaptics touchpads * Revert [Config] CONFIG_HID_RMI=m linux (3.13.0-26.48) trusty; urgency=low [ Benjamin Tissoires ] * SAUCE: (no-up) HID: rmi: introduce RMI driver for Synaptics touchpads - LP: #1305522 * SAUCE: (no-up) HID: rmi: do not stop the device at the end of probe - LP: #1305522 [ Kamal Mostafa ] * Merged back Ubuntu-3.13.0-24.47 security release * Revert n_tty: Fix n_tty_write crash when echoing in raw mode - LP: #1314762 * Release Tracking Bug - LP: #1316835 [ Tim Gardner ] * [Config] CONFIG_HID_RMI=m - LP: #1305522 * [Config] CONFIG_CRYPTO_DEV_NX=n for ppc64el - LP: #1314625 * [Config] CONFIG_ZSWAP=y - LP: #1315203 * Add rpcsec_gss_krb5 to generic inclusion list - LP: #769527 [ Upstream Kernel Changes ] * HID: hidraw: make comment more accurate and nicer - LP: #1305522 * HID: remove hid_get_raw_report in struct hid_device - LP: #1305522 * HID: i2c-hid: implement ll_driver transport-layer callbacks - LP: #1305522 * HID: add inliners for ll_driver transport-layer callbacks - LP: #1305522 * HID: Add transport-driver callbacks to the hid_ll_driver struct - LP: #1305522 * drm/nouveau: fail runtime pm properly. - LP: #1313986 * drm/nouveau: don't suspend/resume display on runtime s/r - LP: #1313986 * n_tty: Fix n_tty_write crash when echoing in raw mode - LP: #1314762 - CVE-2014-0196 * floppy: ignore kernel-only members in FDRAWCMD ioctl input - LP: #1316729 - CVE-2014-1737 * floppy: don't write kernel-only members to FDRAWCMD ioctl output - LP: #1316735 - CVE-2014-1738 linux (3.13.0-25.47) trusty; urgency=low [ Joseph Salisbury ] * Release Tracking Bug - LP: #1313868 [ Adam Lee ] * [Config] CONFIG_RTL8723BE=m, CONFIG_RTL8723_COMMON=m - LP: #1240940 [ Alex Hung ] * SAUCE: (no-up) dell-led: add mic mute led interface - LP: #1308297 [ Andy Whitcroft ] * SAUCE: (no-up) powerpc: Increase COMMAND_LINE_SIZE to 2048 from 512. - LP: #1306677 [ Ben Collins ] * [Config] Disable PAMU on Freescale kernels - LP: #1311738 [ Tim Gardner ] * Revert SAUCE: x86, hyperv: bypass the timer_irq_works() check - LP: #1311683 * SAUCE: (no-up) ALSA: usb-audio: Suppress repetitive debug messages from retire_playback_urb() - LP: #1305133 * SAUCE: (no-up) 'BUG:' message unnecessarily triggers kerneloops - LP: #1305480 * [Config] CONFIG_POWERNV_CPUFREQ=m - LP: #1309576 * [Config] CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND=y for ppc64el - LP: #1309576 * [Config] CONFIG_TRANSPARENT_HUGEPAGE=n for arm64 - LP: #1309221 * [Config] CONFIG_MEMCG_KMEM=y - LP: #1309586 * [Config] CONFIG_CRASH_DUMP=y for ppc64el - LP: #1312783 [ Upstream Kernel Changes ] * Revert rtlwifi: rtl8188ee: enable MSI interrupts mode - LP: #1310512 * mac80211: add length check in ieee80211_is_robust_mgmt_frame() - LP: #1240940 * rtlwifi: rtl8723ae: rtl8723-common: Create new driver for common code - LP: #1240940 * rtlwifi: rtl8723ae: rtl8723-common: Copy common firmware code - LP: #1240940 * rtlwifi: rtl8723ae: rtl8723-common: Copy common dynamic power management code - LP: #1240940 * rtlwifi: rtl8723be: Add new driver - LP: #1240940 * selinux: correctly label /proc inodes in use before the policy is loaded - LP: #1309007 * net: sctp: fix skb leakage in COOKIE ECHO path of chunk-auth_chunk - LP: #1309007 * bridge: multicast: add sanity check for query source addresses - LP: #1309007 * tipc: allow connection shutdown callback to be invoked in advance - LP: #1309007 * tipc: fix connection refcount leak - LP: #1309007 * tipc: drop subscriber connection id invalidation - LP: #1309007 * tipc: fix memory leak during module removal - LP: #1309007 * tipc: don't log disabled tasklet handler errors - LP: #1309007 * inet: frag: make sure forced eviction removes all frags - LP: #1309007 * net: unix: non blocking recvmsg() should not return -EINTR - LP: #1309007 * ipv6: Fix exthdrs offload registration. - LP: #1309007 * bnx2: Fix shutdown sequence - LP: #1309007 * pkt_sched: fq: do not hold qdisc lock while allocating memory - LP: #1309007 * Xen-netback: Fix issue caused by using gso_type wrongly - LP: #1309007 * vlan: Set correct source MAC address with TX VLAN offload enabled - LP: #1309007 * tcp: tcp_release_cb() should release socket ownership - LP: #1309007 * bridge: multicast: add sanity check for general query
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed- trusty' to 'verification-done-trusty'. If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you! ** Tags added: verification-needed-trusty -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Fix Committed Status in “linux” source package in Trusty: Fix Committed Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
dannf@mustang:~$ cat /proc/version Linux version 3.13.0-26-generic (buildd@magic) (gcc version 4.8.2 (Ubuntu/Linaro 4.8.2-19ubuntu1) ) #48-Ubuntu SMP Wed May 7 23:35:13 UTC 2014 dannf@mustang:~$ juju -h Juju -- devops distilled https://juju.ubuntu.com/ Juju provides easy, intelligent service orchestration on top of environments such as Amazon EC2, HP Cloud, OpenStack, MaaS, or your own local machine. Basic commands: juju init generate boilerplate configuration for juju environments juju bootstrapstart up an environment from scratch juju deploy deploy a new service juju add-relation add a relation between two services juju expose expose a service juju help bootstrap more help on e.g. bootstrap command juju help commandslist all commands juju help glossaryglossary of terms juju help topics list all help topics Provider information: juju help azure-provider use on Windows Azure juju help ec2-provider use on Amazon EC2 juju help hpcloud-provider use on HP Cloud juju help local-provider use on this computer juju help openstack-provider use on OpenStack ** Tags removed: verification-needed-trusty ** Tags added: verification-done-trusty -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Fix Committed Status in “linux” source package in Trusty: Fix Committed Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
** Also affects: linux (Ubuntu Utopic) Importance: High Assignee: Tim Gardner (timg-tpi) Status: Fix Committed -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Fix Committed Status in “linux” source package in Trusty: Fix Committed Status in “linux” source package in Utopic: Fix Committed Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
** Branch linked: lp:~ce-hyperscale/ubuntu/trusty/flash- kernel/arm0429demo -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Fix Committed Status in “linux” source package in Trusty: Fix Committed Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
UBUNTU: [Config] CONFIG_TRANSPARENT_HUGEPAGE=n for arm64 ** Also affects: linux (Ubuntu Trusty) Importance: High Status: Triaged ** Changed in: linux (Ubuntu Trusty) Status: Triaged = Fix Committed ** Changed in: linux (Ubuntu Trusty) Assignee: (unassigned) = Tim Gardner (timg-tpi) -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Fix Committed Status in “linux” source package in Trusty: Fix Committed Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
** Changed in: linux (Ubuntu) Importance: Undecided = High ** Tags added: kernel-da-key trusty -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Incomplete Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1309221] Re: transparent hugepages flaky on arm64
Being discussed on the kernel team mailing list: https://lists.ubuntu.com/archives/kernel-team/2014-April/041621.html ** Changed in: linux (Ubuntu) Status: Incomplete = Triaged -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1309221 Title: transparent hugepages flaky on arm64 Status in “linux” package in Ubuntu: Triaged Bug description: linux 3.13.0-24.46 transparent hugepages, which are configured on by default in Ubuntu, are flaky on arm64. The most evident system is that go processes reliably hang or crash. This has been observed with both building juju-core with gccgo, and running the resulting juju binaries. The root cause of this is that get_user_pages_fast is not yet implemented on arm64 (and arm64). Patches are floating around, but haven't landed. http://comments.gmane.org/gmane.linux.ports.arm.kernel/299268 We've verified that either applying those patches, or disabling THP via sysfs, prevents the above issues. At minimum we should change the default boottime sysfs setting for THP to never until we can fix this feature properly. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1309221/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
