Greetings, 0day kernel testing robot got the below dmesg and the first bad commit is https://github.com/0day-ci/linux/commits/Timofey-Titovets/ksm-replace-jha... commit 038adb295b0c04721a12b276f8b521249fd2db9d Author: Timofey Titovets <nefelim4ag(a)gmail.com&gt; AuthorDate: Fri Dec 29 12:52:41 2017 +0300 Commit: 0day robot <fengguang.wu(a)intel.com&gt; CommitDate: Sun Dec 31 15:54:27 2017 +0800 ksm: replace jhash2 with faster hash Pickup, Sioh Lee crc32 patch, after some long conversation and hassles, merge with my work on xxhash, add choice fastest hash helper. Base idea are same, replace jhash2 with something faster. Perf numbers: Intel(R) Xeon(R) CPU E5-2420 v2 @ 2.20GHz ksm: crc32c hash() 12081 MB/s ksm: jhash2 hash() 1569 MB/s ksm: xxh64 hash() 8770 MB/s ksm: xxh32 hash() 4529 MB/s As jhash2 always will be slower, just drop it from choice. Add function to autoselect hash algo on boot, based on speed, like raid6 code does. Move init of zero_hash from init, to start of ksm thread, as ksm init run on early kernel init, run perf testing stuff on main kernel thread looks bad to me. One problem exists with that patch, ksm init run too early, and crc32c module, even compiled in can't be found, so i see: - ksm: alloc crc32c shash error 2 in dmesg. I give up on that, so ideas welcomed. Only idea that i have, are to avoid early init by moving zero_checksum to sysfs_store parm, i.e. that's default to false, and that will work, i think. Thanks. Changes: v1 -> v2: - Merge xxhash/crc32 patches - Replace crc32 with crc32c (crc32 have same as jhash2 speed) - Add auto speed test and auto choice of fastest hash function Signed-off-by: Timofey Titovets <nefelim4ag(a)gmail.com&gt; Signed-off-by: leesioh <solee(a)os.korea.ac.kr&gt; CC: Andrea Arcangeli <aarcange(a)redhat.com&gt; CC: linux-mm(a)kvack.org CC: kvm(a)vger.kernel.org a4f20e3ed1 pci: test for unexpectedly disabled bridges 038adb295b ksm: replace jhash2 with faster hash +----------------------------------------------------------------+------------+------------+ | | a4f20e3ed1 | 038adb295b | +----------------------------------------------------------------+------------+------------+ | boot_successes | 35 | 0 | | boot_failures | 0 | 34 | | BUG:sleeping_function_called_from_invalid_context_at_mm/slab.h | 0 | 34 | +----------------------------------------------------------------+------------+------------+ [ 0.361000] ACPI: bus type PCI registered [ 0.373807] PCI: PCI BIOS area is rw and x. Use pci=nobios if you want it NX. [ 0.375445] PCI: PCI BIOS revision 2.10 entry at 0xfd501, last bus=0 [ 0.376802] PCI: Using configuration type 1 for base access [ 0.420610] HugeTLB registered 2.00 MiB page size, pre-allocated 0 pages [ 0.425175] BUG: sleeping function called from invalid context at mm/slab.h:420 [ 0.425386] in_atomic(): 1, irqs_disabled(): 0, pid: 26, name: ksmd [ 0.425386] no locks held by ksmd/26. [ 0.425386] CPU: 1 PID: 26 Comm: ksmd Not tainted 4.15.0-rc4-mm1-00235-g038adb2 #1 [ 0.426091] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 [ 0.428008] Call Trace: [ 0.428664] dump_stack+0x92/0xcc [ 0.429474] ___might_sleep+0x137/0x150 [ 0.430000] __might_sleep+0x75/0x80 [ 0.430000] ? fs_reclaim_release+0x11/0x30 [ 0.430000] kmem_cache_alloc_trace+0x56/0x240 [ 0.440092] cryptd: max_cpu_qlen set to 1000 [ 0.440598] ? fasthash+0x5e/0x2f0 [ 0.440598] fasthash+0x5e/0x2f0 [ 0.440598] ? kmap_atomic_prot+0x48/0x170 [ 0.440598] ksm_scan_thread+0x62/0x1900 [ 0.440598] ? __schedule+0x706/0x770 [ 0.440598] ? complete+0x3f/0x50 [ 0.440598] kthread+0x101/0x110 [ 0.440598] ? __stable_node_chain+0x70/0x70 [ 0.440598] ? __kthread_cancel_work+0x70/0x70 [ 0.440598] ret_from_fork+0x19/0x30 [ 0.450055] ksm: alloc crc32c shash error 2 [ 0.563943] ksm: xxh32 hash() 3432 MB/s [ 0.610019] raid6: mmxx1 gen() 3104 MB/s [ 0.683945] ksm: xxh64 hash() 1061 MB/s [ 0.684948] ksm: choise xxh32 as hash function # HH:MM RESULT GOOD BAD GOOD_BUT_DIRTY DIRTY_NOT_BAD git bisect start 038adb295b0c04721a12b276f8b521249fd2db9d a4f20e3ed193cd4b2f742ce37f88112c7441146f -- # first bad commit: [038adb295b0c04721a12b276f8b521249fd2db9d] ksm: replace jhash2 with faster hash git bisect good a4f20e3ed193cd4b2f742ce37f88112c7441146f # 17:31 G 31 0 0 0 pci: test for unexpectedly disabled bridges # extra tests with debug options git bisect bad 038adb295b0c04721a12b276f8b521249fd2db9d # 17:44 B 0 11 32 8 ksm: replace jhash2 with faster hash # extra tests on HEAD of linux-review/Timofey-Titovets/ksm-replace-jhash2-with-faster-hash/20171231-155425 git bisect bad 038adb295b0c04721a12b276f8b521249fd2db9d # 17:44 B 0 34 50 0 ksm: replace jhash2 with faster hash # extra tests on tree/branch linux-review/Timofey-Titovets/ksm-replace-jhash2-with-faster-hash/20171231-155425 git bisect bad 038adb295b0c04721a12b276f8b521249fd2db9d # 17:45 B 0 34 50 0 ksm: replace jhash2 with faster hash # extra tests with first bad commit reverted git bisect good b245f96d362d0b1b9430e7fb1067312a52b8748c # 17:56 G 10 0 0 0 Revert "ksm: replace jhash2 with faster hash" --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/pipermail/lkp Intel Corporation