drm/edid: Fix the HDTV hack yet more. #15

andrewshadura · 2012-01-04T12:26:00Z

Checking for just two variants of standard timings for
1366x768 isn't quite correct, let's check for ranges
instead.

Signed-off-by: Andrew Shadura [email protected]

Checking for just two variants of standard timings for 1366x768 isn't quite correct, let's check for ranges instead. Signed-off-by: Andrew Shadura <[email protected]>

If the netdev is already in NETREG_UNREGISTERING/_UNREGISTERED state, do not update the real num tx queues. netdev_queue_update_kobjects() is already called via remove_queue_kobjects() at NETREG_UNREGISTERING time. So, when upper layer driver, e.g., FCoE protocol stack is monitoring the netdev event of NETDEV_UNREGISTER and calls back to LLD ndo_fcoe_disable() to remove extra queues allocated for FCoE, the associated txq sysfs kobjects are already removed, and trying to update the real num queues would cause something like below: ... PID: 25138 TASK: ffff88021e64c440 CPU: 3 COMMAND: "kworker/3:3" #0 [ffff88021f007760] machine_kexec at ffffffff810226d9 #1 [ffff88021f0077d0] crash_kexec at ffffffff81089d2d #2 [ffff88021f0078a0] oops_end at ffffffff813bca78 #3 [ffff88021f0078d0] no_context at ffffffff81029e72 #4 [ffff88021f007920] __bad_area_nosemaphore at ffffffff8102a155 #5 [ffff88021f0079f0] bad_area_nosemaphore at ffffffff8102a23e torvalds#6 [ffff88021f007a00] do_page_fault at ffffffff813bf32e torvalds#7 [ffff88021f007b10] page_fault at ffffffff813bc045 [exception RIP: sysfs_find_dirent+17] RIP: ffffffff81178611 RSP: ffff88021f007bc0 RFLAGS: 00010246 RAX: ffff88021e64c440 RBX: ffffffff8156cc63 RCX: 0000000000000004 RDX: ffffffff8156cc63 RSI: 0000000000000000 RDI: 0000000000000000 RBP: ffff88021f007be0 R8: 0000000000000004 R9: 0000000000000008 R10: ffffffff816fed00 R11: 0000000000000004 R12: 0000000000000000 R13: ffffffff8156cc63 R14: 0000000000000000 R15: ffff8802222a0000 ORIG_RAX: ffffffffffffffff CS: 0010 SS: 0018 torvalds#8 [ffff88021f007be8] sysfs_get_dirent at ffffffff81178c07 torvalds#9 [ffff88021f007c18] sysfs_remove_group at ffffffff8117ac27 torvalds#10 [ffff88021f007c48] netdev_queue_update_kobjects at ffffffff813178f9 torvalds#11 [ffff88021f007c88] netif_set_real_num_tx_queues at ffffffff81303e38 torvalds#12 [ffff88021f007cc8] ixgbe_set_num_queues at ffffffffa0249763 [ixgbe] torvalds#13 [ffff88021f007cf8] ixgbe_init_interrupt_scheme at ffffffffa024ea89 [ixgbe] torvalds#14 [ffff88021f007d48] ixgbe_fcoe_disable at ffffffffa0267113 [ixgbe] torvalds#15 [ffff88021f007d68] vlan_dev_fcoe_disable at ffffffffa014fef5 [8021q] torvalds#16 [ffff88021f007d78] fcoe_interface_cleanup at ffffffffa02b7dfd [fcoe] torvalds#17 [ffff88021f007df8] fcoe_destroy_work at ffffffffa02b7f08 [fcoe] torvalds#18 [ffff88021f007e18] process_one_work at ffffffff8105d7ca torvalds#19 [ffff88021f007e68] worker_thread at ffffffff81060513 torvalds#20 [ffff88021f007ee8] kthread at ffffffff810648b6 torvalds#21 [ffff88021f007f48] kernel_thread_helper at ffffffff813c40f4 Signed-off-by: Yi Zou <[email protected]> Tested-by: Ross Brattain <[email protected]> Tested-by: Stephen Ko <[email protected]> Signed-off-by: Jeff Kirsher <[email protected]>

peterwillcn · 2012-02-13T08:20:21Z

please add LinuxPPS in kernel support ntpd for gps...
http://wiki.enneenne.com/index.php/LinuxPPS_installation

…S block during isolation for migration commit 0bf380b upstream. When isolating for migration, migration starts at the start of a zone which is not necessarily pageblock aligned. Further, it stops isolating when COMPACT_CLUSTER_MAX pages are isolated so migrate_pfn is generally not aligned. This allows isolate_migratepages() to call pfn_to_page() on an invalid PFN which can result in a crash. This was originally reported against a 3.0-based kernel with the following trace in a crash dump. PID: 9902 TASK: d47aecd0 CPU: 0 COMMAND: "memcg_process_s" #0 [d72d3ad0] crash_kexec at c028cfdb #1 [d72d3b24] oops_end at c05c5322 #2 [d72d3b38] __bad_area_nosemaphore at c0227e60 #3 [d72d3bec] bad_area at c0227fb6 #4 [d72d3c00] do_page_fault at c05c72ec #5 [d72d3c80] error_code (via page_fault) at c05c47a4 EAX: 00000000 EBX: 000c0000 ECX: 00000001 EDX: 00000807 EBP: 000c0000 DS: 007b ESI: 00000001 ES: 007b EDI: f3000a80 GS: 6f50 CS: 0060 EIP: c030b15a ERR: ffffffff EFLAGS: 00010002 torvalds#6 [d72d3cb4] isolate_migratepages at c030b15a torvalds#7 [d72d3d1] zone_watermark_ok at c02d26cb torvalds#8 [d72d3d2c] compact_zone at c030b8de torvalds#9 [d72d3d68] compact_zone_order at c030bba1 torvalds#10 [d72d3db4] try_to_compact_pages at c030bc84 torvalds#11 [d72d3ddc] __alloc_pages_direct_compact at c02d61e7 torvalds#12 [d72d3e08] __alloc_pages_slowpath at c02d66c7 torvalds#13 [d72d3e78] __alloc_pages_nodemask at c02d6a97 torvalds#14 [d72d3eb8] alloc_pages_vma at c030a845 torvalds#15 [d72d3ed4] do_huge_pmd_anonymous_page at c03178eb torvalds#16 [d72d3f00] handle_mm_fault at c02f36c6 torvalds#17 [d72d3f30] do_page_fault at c05c70ed torvalds#18 [d72d3fb0] error_code (via page_fault) at c05c47a4 EAX: b71ff000 EBX: 00000001 ECX: 00001600 EDX: 00000431 DS: 007b ESI: 08048950 ES: 007b EDI: bfaa3788 SS: 007b ESP: bfaa36e0 EBP: bfaa3828 GS: 6f50 CS: 0073 EIP: 080487c8 ERR: ffffffff EFLAGS: 00010202 It was also reported by Herbert van den Bergh against 3.1-based kernel with the following snippet from the console log. BUG: unable to handle kernel paging request at 01c00008 IP: [<c0522399>] isolate_migratepages+0x119/0x390 *pdpt = 000000002f7ce001 *pde = 0000000000000000 It is expected that it also affects 3.2.x and current mainline. The problem is that pfn_valid is only called on the first PFN being checked and that PFN is not necessarily aligned. Lets say we have a case like this H = MAX_ORDER_NR_PAGES boundary | = pageblock boundary m = cc->migrate_pfn f = cc->free_pfn o = memory hole H------|------H------|----m-Hoooooo|ooooooH-f----|------H The migrate_pfn is just below a memory hole and the free scanner is beyond the hole. When isolate_migratepages started, it scans from migrate_pfn to migrate_pfn+pageblock_nr_pages which is now in a memory hole. It checks pfn_valid() on the first PFN but then scans into the hole where there are not necessarily valid struct pages. This patch ensures that isolate_migratepages calls pfn_valid when necessary. Reported-by: Herbert van den Bergh <[email protected]> Tested-by: Herbert van den Bergh <[email protected]> Signed-off-by: Mel Gorman <[email protected]> Acked-by: Michal Nazarewicz <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]> Signed-off-by: Greg Kroah-Hartman <[email protected]>

…S block during isolation for migration commit 0bf380b upstream. When isolating for migration, migration starts at the start of a zone which is not necessarily pageblock aligned. Further, it stops isolating when COMPACT_CLUSTER_MAX pages are isolated so migrate_pfn is generally not aligned. This allows isolate_migratepages() to call pfn_to_page() on an invalid PFN which can result in a crash. This was originally reported against a 3.0-based kernel with the following trace in a crash dump. PID: 9902 TASK: d47aecd0 CPU: 0 COMMAND: "memcg_process_s" #0 [d72d3ad0] crash_kexec at c028cfdb #1 [d72d3b24] oops_end at c05c5322 #2 [d72d3b38] __bad_area_nosemaphore at c0227e60 #3 [d72d3bec] bad_area at c0227fb6 #4 [d72d3c00] do_page_fault at c05c72ec #5 [d72d3c80] error_code (via page_fault) at c05c47a4 EAX: 00000000 EBX: 000c0000 ECX: 00000001 EDX: 00000807 EBP: 000c0000 DS: 007b ESI: 00000001 ES: 007b EDI: f3000a80 GS: 6f50 CS: 0060 EIP: c030b15a ERR: ffffffff EFLAGS: 00010002 #6 [d72d3cb4] isolate_migratepages at c030b15a #7 [d72d3d1] zone_watermark_ok at c02d26cb #8 [d72d3d2c] compact_zone at c030b8de #9 [d72d3d68] compact_zone_order at c030bba1 torvalds#10 [d72d3db4] try_to_compact_pages at c030bc84 torvalds#11 [d72d3ddc] __alloc_pages_direct_compact at c02d61e7 torvalds#12 [d72d3e08] __alloc_pages_slowpath at c02d66c7 torvalds#13 [d72d3e78] __alloc_pages_nodemask at c02d6a97 torvalds#14 [d72d3eb8] alloc_pages_vma at c030a845 torvalds#15 [d72d3ed4] do_huge_pmd_anonymous_page at c03178eb torvalds#16 [d72d3f00] handle_mm_fault at c02f36c6 torvalds#17 [d72d3f30] do_page_fault at c05c70ed torvalds#18 [d72d3fb0] error_code (via page_fault) at c05c47a4 EAX: b71ff000 EBX: 00000001 ECX: 00001600 EDX: 00000431 DS: 007b ESI: 08048950 ES: 007b EDI: bfaa3788 SS: 007b ESP: bfaa36e0 EBP: bfaa3828 GS: 6f50 CS: 0073 EIP: 080487c8 ERR: ffffffff EFLAGS: 00010202 It was also reported by Herbert van den Bergh against 3.1-based kernel with the following snippet from the console log. BUG: unable to handle kernel paging request at 01c00008 IP: [<c0522399>] isolate_migratepages+0x119/0x390 *pdpt = 000000002f7ce001 *pde = 0000000000000000 It is expected that it also affects 3.2.x and current mainline. The problem is that pfn_valid is only called on the first PFN being checked and that PFN is not necessarily aligned. Lets say we have a case like this H = MAX_ORDER_NR_PAGES boundary | = pageblock boundary m = cc->migrate_pfn f = cc->free_pfn o = memory hole H------|------H------|----m-Hoooooo|ooooooH-f----|------H The migrate_pfn is just below a memory hole and the free scanner is beyond the hole. When isolate_migratepages started, it scans from migrate_pfn to migrate_pfn+pageblock_nr_pages which is now in a memory hole. It checks pfn_valid() on the first PFN but then scans into the hole where there are not necessarily valid struct pages. This patch ensures that isolate_migratepages calls pfn_valid when necessary. Reported-by: Herbert van den Bergh <[email protected]> Tested-by: Herbert van den Bergh <[email protected]> Signed-off-by: Mel Gorman <[email protected]> Acked-by: Michal Nazarewicz <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]> Signed-off-by: Greg Kroah-Hartman <[email protected]>

…S block during isolation for migration BugLink: http://bugs.launchpad.net/bugs/931719 commit 0bf380b upstream. When isolating for migration, migration starts at the start of a zone which is not necessarily pageblock aligned. Further, it stops isolating when COMPACT_CLUSTER_MAX pages are isolated so migrate_pfn is generally not aligned. This allows isolate_migratepages() to call pfn_to_page() on an invalid PFN which can result in a crash. This was originally reported against a 3.0-based kernel with the following trace in a crash dump. PID: 9902 TASK: d47aecd0 CPU: 0 COMMAND: "memcg_process_s" #0 [d72d3ad0] crash_kexec at c028cfdb #1 [d72d3b24] oops_end at c05c5322 #2 [d72d3b38] __bad_area_nosemaphore at c0227e60 #3 [d72d3bec] bad_area at c0227fb6 #4 [d72d3c00] do_page_fault at c05c72ec #5 [d72d3c80] error_code (via page_fault) at c05c47a4 EAX: 00000000 EBX: 000c0000 ECX: 00000001 EDX: 00000807 EBP: 000c0000 DS: 007b ESI: 00000001 ES: 007b EDI: f3000a80 GS: 6f50 CS: 0060 EIP: c030b15a ERR: ffffffff EFLAGS: 00010002 torvalds#6 [d72d3cb4] isolate_migratepages at c030b15a torvalds#7 [d72d3d1] zone_watermark_ok at c02d26cb torvalds#8 [d72d3d2c] compact_zone at c030b8de torvalds#9 [d72d3d68] compact_zone_order at c030bba1 torvalds#10 [d72d3db4] try_to_compact_pages at c030bc84 torvalds#11 [d72d3ddc] __alloc_pages_direct_compact at c02d61e7 torvalds#12 [d72d3e08] __alloc_pages_slowpath at c02d66c7 torvalds#13 [d72d3e78] __alloc_pages_nodemask at c02d6a97 torvalds#14 [d72d3eb8] alloc_pages_vma at c030a845 torvalds#15 [d72d3ed4] do_huge_pmd_anonymous_page at c03178eb torvalds#16 [d72d3f00] handle_mm_fault at c02f36c6 torvalds#17 [d72d3f30] do_page_fault at c05c70ed torvalds#18 [d72d3fb0] error_code (via page_fault) at c05c47a4 EAX: b71ff000 EBX: 00000001 ECX: 00001600 EDX: 00000431 DS: 007b ESI: 08048950 ES: 007b EDI: bfaa3788 SS: 007b ESP: bfaa36e0 EBP: bfaa3828 GS: 6f50 CS: 0073 EIP: 080487c8 ERR: ffffffff EFLAGS: 00010202 It was also reported by Herbert van den Bergh against 3.1-based kernel with the following snippet from the console log. BUG: unable to handle kernel paging request at 01c00008 IP: [<c0522399>] isolate_migratepages+0x119/0x390 *pdpt = 000000002f7ce001 *pde = 0000000000000000 It is expected that it also affects 3.2.x and current mainline. The problem is that pfn_valid is only called on the first PFN being checked and that PFN is not necessarily aligned. Lets say we have a case like this H = MAX_ORDER_NR_PAGES boundary | = pageblock boundary m = cc->migrate_pfn f = cc->free_pfn o = memory hole H------|------H------|----m-Hoooooo|ooooooH-f----|------H The migrate_pfn is just below a memory hole and the free scanner is beyond the hole. When isolate_migratepages started, it scans from migrate_pfn to migrate_pfn+pageblock_nr_pages which is now in a memory hole. It checks pfn_valid() on the first PFN but then scans into the hole where there are not necessarily valid struct pages. This patch ensures that isolate_migratepages calls pfn_valid when necessary. Reported-by: Herbert van den Bergh <[email protected]> Tested-by: Herbert van den Bergh <[email protected]> Signed-off-by: Mel Gorman <[email protected]> Acked-by: Michal Nazarewicz <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]> Signed-off-by: Greg Kroah-Hartman <[email protected]> Signed-off-by: Tim Gardner <[email protected]>

…S block during isolation for migration commit 0bf380b upstream. When isolating for migration, migration starts at the start of a zone which is not necessarily pageblock aligned. Further, it stops isolating when COMPACT_CLUSTER_MAX pages are isolated so migrate_pfn is generally not aligned. This allows isolate_migratepages() to call pfn_to_page() on an invalid PFN which can result in a crash. This was originally reported against a 3.0-based kernel with the following trace in a crash dump. PID: 9902 TASK: d47aecd0 CPU: 0 COMMAND: "memcg_process_s" #0 [d72d3ad0] crash_kexec at c028cfdb #1 [d72d3b24] oops_end at c05c5322 #2 [d72d3b38] __bad_area_nosemaphore at c0227e60 #3 [d72d3bec] bad_area at c0227fb6 #4 [d72d3c00] do_page_fault at c05c72ec #5 [d72d3c80] error_code (via page_fault) at c05c47a4 EAX: 00000000 EBX: 000c0000 ECX: 00000001 EDX: 00000807 EBP: 000c0000 DS: 007b ESI: 00000001 ES: 007b EDI: f3000a80 GS: 6f50 CS: 0060 EIP: c030b15a ERR: ffffffff EFLAGS: 00010002 #6 [d72d3cb4] isolate_migratepages at c030b15a #7 [d72d3d1] zone_watermark_ok at c02d26cb #8 [d72d3d2c] compact_zone at c030b8de #9 [d72d3d68] compact_zone_order at c030bba1 torvalds#10 [d72d3db4] try_to_compact_pages at c030bc84 torvalds#11 [d72d3ddc] __alloc_pages_direct_compact at c02d61e7 torvalds#12 [d72d3e08] __alloc_pages_slowpath at c02d66c7 torvalds#13 [d72d3e78] __alloc_pages_nodemask at c02d6a97 torvalds#14 [d72d3eb8] alloc_pages_vma at c030a845 torvalds#15 [d72d3ed4] do_huge_pmd_anonymous_page at c03178eb torvalds#16 [d72d3f00] handle_mm_fault at c02f36c6 torvalds#17 [d72d3f30] do_page_fault at c05c70ed torvalds#18 [d72d3fb0] error_code (via page_fault) at c05c47a4 EAX: b71ff000 EBX: 00000001 ECX: 00001600 EDX: 00000431 DS: 007b ESI: 08048950 ES: 007b EDI: bfaa3788 SS: 007b ESP: bfaa36e0 EBP: bfaa3828 GS: 6f50 CS: 0073 EIP: 080487c8 ERR: ffffffff EFLAGS: 00010202 It was also reported by Herbert van den Bergh against 3.1-based kernel with the following snippet from the console log. BUG: unable to handle kernel paging request at 01c00008 IP: [<c0522399>] isolate_migratepages+0x119/0x390 *pdpt = 000000002f7ce001 *pde = 0000000000000000 It is expected that it also affects 3.2.x and current mainline. The problem is that pfn_valid is only called on the first PFN being checked and that PFN is not necessarily aligned. Lets say we have a case like this H = MAX_ORDER_NR_PAGES boundary | = pageblock boundary m = cc->migrate_pfn f = cc->free_pfn o = memory hole H------|------H------|----m-Hoooooo|ooooooH-f----|------H The migrate_pfn is just below a memory hole and the free scanner is beyond the hole. When isolate_migratepages started, it scans from migrate_pfn to migrate_pfn+pageblock_nr_pages which is now in a memory hole. It checks pfn_valid() on the first PFN but then scans into the hole where there are not necessarily valid struct pages. This patch ensures that isolate_migratepages calls pfn_valid when necessary. Reported-by: Herbert van den Bergh <[email protected]> Tested-by: Herbert van den Bergh <[email protected]> Signed-off-by: Mel Gorman <[email protected]> Acked-by: Michal Nazarewicz <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]> Signed-off-by: Greg Kroah-Hartman <[email protected]>

Smsc95xx patches

…S block during isolation for migration commit 0bf380b upstream. When isolating for migration, migration starts at the start of a zone which is not necessarily pageblock aligned. Further, it stops isolating when COMPACT_CLUSTER_MAX pages are isolated so migrate_pfn is generally not aligned. This allows isolate_migratepages() to call pfn_to_page() on an invalid PFN which can result in a crash. This was originally reported against a 3.0-based kernel with the following trace in a crash dump. PID: 9902 TASK: d47aecd0 CPU: 0 COMMAND: "memcg_process_s" #0 [d72d3ad0] crash_kexec at c028cfdb #1 [d72d3b24] oops_end at c05c5322 #2 [d72d3b38] __bad_area_nosemaphore at c0227e60 #3 [d72d3bec] bad_area at c0227fb6 #4 [d72d3c00] do_page_fault at c05c72ec #5 [d72d3c80] error_code (via page_fault) at c05c47a4 EAX: 00000000 EBX: 000c0000 ECX: 00000001 EDX: 00000807 EBP: 000c0000 DS: 007b ESI: 00000001 ES: 007b EDI: f3000a80 GS: 6f50 CS: 0060 EIP: c030b15a ERR: ffffffff EFLAGS: 00010002 #6 [d72d3cb4] isolate_migratepages at c030b15a #7 [d72d3d1] zone_watermark_ok at c02d26cb #8 [d72d3d2c] compact_zone at c030b8de #9 [d72d3d68] compact_zone_order at c030bba1 torvalds#10 [d72d3db4] try_to_compact_pages at c030bc84 torvalds#11 [d72d3ddc] __alloc_pages_direct_compact at c02d61e7 torvalds#12 [d72d3e08] __alloc_pages_slowpath at c02d66c7 torvalds#13 [d72d3e78] __alloc_pages_nodemask at c02d6a97 torvalds#14 [d72d3eb8] alloc_pages_vma at c030a845 torvalds#15 [d72d3ed4] do_huge_pmd_anonymous_page at c03178eb torvalds#16 [d72d3f00] handle_mm_fault at c02f36c6 torvalds#17 [d72d3f30] do_page_fault at c05c70ed torvalds#18 [d72d3fb0] error_code (via page_fault) at c05c47a4 EAX: b71ff000 EBX: 00000001 ECX: 00001600 EDX: 00000431 DS: 007b ESI: 08048950 ES: 007b EDI: bfaa3788 SS: 007b ESP: bfaa36e0 EBP: bfaa3828 GS: 6f50 CS: 0073 EIP: 080487c8 ERR: ffffffff EFLAGS: 00010202 It was also reported by Herbert van den Bergh against 3.1-based kernel with the following snippet from the console log. BUG: unable to handle kernel paging request at 01c00008 IP: [<c0522399>] isolate_migratepages+0x119/0x390 *pdpt = 000000002f7ce001 *pde = 0000000000000000 It is expected that it also affects 3.2.x and current mainline. The problem is that pfn_valid is only called on the first PFN being checked and that PFN is not necessarily aligned. Lets say we have a case like this H = MAX_ORDER_NR_PAGES boundary | = pageblock boundary m = cc->migrate_pfn f = cc->free_pfn o = memory hole H------|------H------|----m-Hoooooo|ooooooH-f----|------H The migrate_pfn is just below a memory hole and the free scanner is beyond the hole. When isolate_migratepages started, it scans from migrate_pfn to migrate_pfn+pageblock_nr_pages which is now in a memory hole. It checks pfn_valid() on the first PFN but then scans into the hole where there are not necessarily valid struct pages. This patch ensures that isolate_migratepages calls pfn_valid when necessary. Reported-by: Herbert van den Bergh <[email protected]> Tested-by: Herbert van den Bergh <[email protected]> Signed-off-by: Mel Gorman <[email protected]> Acked-by: Michal Nazarewicz <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]> Signed-off-by: Greg Kroah-Hartman <[email protected]>

The TX path may release the dmabuf in a context where we cannot wait. This happens when the user unbinds a TX dmabuf while there are still references to its netmems in the TX path. In that case, the netmems will be put_netmem'd from a context where we can't unmap the dmabuf, resulting in a BUG like seen by Stan: [ 1.548495] BUG: sleeping function called from invalid context at drivers/dma-buf/dma-buf.c:1255 [ 1.548741] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 149, name: ncdevmem [ 1.548926] preempt_count: 201, expected: 0 [ 1.549026] RCU nest depth: 0, expected: 0 [ 1.549197] [ 1.549237] ============================= [ 1.549331] [ BUG: Invalid wait context ] [ 1.549425] 6.13.0-rc3-00770-gbc9ef9606dc9-dirty torvalds#15 Tainted: G W [ 1.549609] ----------------------------- [ 1.549704] ncdevmem/149 is trying to lock: [ 1.549801] ffff8880066701c0 (reservation_ww_class_mutex){+.+.}-{4:4}, at: dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.550051] other info that might help us debug this: [ 1.550167] context-{5:5} [ 1.550229] 3 locks held by ncdevmem/149: [ 1.550322] #0: ffff888005730208 (&sb->s_type->i_mutex_key#11){+.+.}-{4:4}, at: sock_close+0x40/0xf0 [ 1.550530] #1: ffff88800b148f98 (sk_lock-AF_INET6){+.+.}-{0:0}, at: tcp_close+0x19/0x80 [ 1.550731] #2: ffff88800b148f18 (slock-AF_INET6){+.-.}-{3:3}, at: __tcp_close+0x185/0x4b0 [ 1.550921] stack backtrace: [ 1.550990] CPU: 0 UID: 0 PID: 149 Comm: ncdevmem Tainted: G W 6.13.0-rc3-00770-gbc9ef9606dc9-dirty torvalds#15 [ 1.551233] Tainted: [W]=WARN [ 1.551304] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Arch Linux 1.16.3-1-1 04/01/2014 [ 1.551518] Call Trace: [ 1.551584] <TASK> [ 1.551636] dump_stack_lvl+0x86/0xc0 [ 1.551723] __lock_acquire+0xb0f/0xc30 [ 1.551814] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.551941] lock_acquire+0xf1/0x2a0 [ 1.552026] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552152] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552281] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552408] __ww_mutex_lock+0x121/0x1060 [ 1.552503] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552648] ww_mutex_lock+0x3d/0xa0 [ 1.552733] dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552857] __net_devmem_dmabuf_binding_free+0x56/0xb0 [ 1.552979] skb_release_data+0x120/0x1f0 [ 1.553074] __kfree_skb+0x29/0xa0 [ 1.553156] tcp_write_queue_purge+0x41/0x310 [ 1.553259] tcp_v4_destroy_sock+0x127/0x320 [ 1.553363] ? __tcp_close+0x169/0x4b0 [ 1.553452] inet_csk_destroy_sock+0x53/0x130 [ 1.553560] __tcp_close+0x421/0x4b0 [ 1.553646] tcp_close+0x24/0x80 [ 1.553724] inet_release+0x5d/0x90 [ 1.553806] sock_close+0x4a/0xf0 [ 1.553886] __fput+0x9c/0x2b0 [ 1.553960] task_work_run+0x89/0xc0 [ 1.554046] do_exit+0x27f/0x980 [ 1.554125] do_group_exit+0xa4/0xb0 [ 1.554211] __x64_sys_exit_group+0x17/0x20 [ 1.554309] x64_sys_call+0x21a0/0x21a0 [ 1.554400] do_syscall_64+0xec/0x1d0 [ 1.554487] ? exc_page_fault+0x8a/0xf0 [ 1.554585] entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 1.554703] RIP: 0033:0x7f2f8a27abcd Resolve this by making __net_devmem_dmabuf_binding_free schedule_work'd. Suggested-by: Stanislav Fomichev <[email protected]> Signed-off-by: Mina Almasry <[email protected]> Acked-by: Stanislav Fomichev <[email protected]> Signed-off-by: NipaLocal <nipa@local>

When a bio with REQ_PREFLUSH is submitted to dm, __send_empty_flush() generates a flush_bio with REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC, which causes the flush_bio to be throttled by wbt_wait(). An example from v5.4, similar problem also exists in upstream: crash> bt 2091206 PID: 2091206 TASK: ffff2050df92a300 CPU: 109 COMMAND: "kworker/u260:0" #0 [ffff800084a2f7f0] __switch_to at ffff80004008aeb8 #1 [ffff800084a2f820] __schedule at ffff800040bfa0c4 #2 [ffff800084a2f880] schedule at ffff800040bfa4b4 #3 [ffff800084a2f8a0] io_schedule at ffff800040bfa9c4 #4 [ffff800084a2f8c0] rq_qos_wait at ffff8000405925bc #5 [ffff800084a2f940] wbt_wait at ffff8000405bb3a0 torvalds#6 [ffff800084a2f9a0] __rq_qos_throttle at ffff800040592254 torvalds#7 [ffff800084a2f9c0] blk_mq_make_request at ffff80004057cf38 torvalds#8 [ffff800084a2fa60] generic_make_request at ffff800040570138 torvalds#9 [ffff800084a2fae0] submit_bio at ffff8000405703b4 torvalds#10 [ffff800084a2fb50] xlog_write_iclog at ffff800001280834 [xfs] torvalds#11 [ffff800084a2fbb0] xlog_sync at ffff800001280c3c [xfs] torvalds#12 [ffff800084a2fbf0] xlog_state_release_iclog at ffff800001280df4 [xfs] torvalds#13 [ffff800084a2fc10] xlog_write at ffff80000128203c [xfs] torvalds#14 [ffff800084a2fcd0] xlog_cil_push at ffff8000012846dc [xfs] torvalds#15 [ffff800084a2fda0] xlog_cil_push_work at ffff800001284a2c [xfs] torvalds#16 [ffff800084a2fdb0] process_one_work at ffff800040111d08 torvalds#17 [ffff800084a2fe00] worker_thread at ffff8000401121cc torvalds#18 [ffff800084a2fe70] kthread at ffff800040118de4 After commit 2def284 ("xfs: don't allow log IO to be throttled"), the metadata submitted by xlog_write_iclog() should not be throttled. But due to the existence of the dm layer, throttling flush_bio indirectly causes the metadata bio to be throttled. Fix this by conditionally adding REQ_IDLE to flush_bio.bi_opf, which makes wbt_should_throttle() return false to avoid wbt_wait(). Signed-off-by: Jinliang Zheng <[email protected]> Reviewed-by: Tianxiang Peng <[email protected]> Reviewed-by: Hao Peng <[email protected]> Signed-off-by: Mikulas Patocka <[email protected]>

The TX path may release the dmabuf in a context where we cannot wait. This happens when the user unbinds a TX dmabuf while there are still references to its netmems in the TX path. In that case, the netmems will be put_netmem'd from a context where we can't unmap the dmabuf, resulting in a BUG like seen by Stan: [ 1.548495] BUG: sleeping function called from invalid context at drivers/dma-buf/dma-buf.c:1255 [ 1.548741] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 149, name: ncdevmem [ 1.548926] preempt_count: 201, expected: 0 [ 1.549026] RCU nest depth: 0, expected: 0 [ 1.549197] [ 1.549237] ============================= [ 1.549331] [ BUG: Invalid wait context ] [ 1.549425] 6.13.0-rc3-00770-gbc9ef9606dc9-dirty torvalds#15 Tainted: G W [ 1.549609] ----------------------------- [ 1.549704] ncdevmem/149 is trying to lock: [ 1.549801] ffff8880066701c0 (reservation_ww_class_mutex){+.+.}-{4:4}, at: dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.550051] other info that might help us debug this: [ 1.550167] context-{5:5} [ 1.550229] 3 locks held by ncdevmem/149: [ 1.550322] #0: ffff888005730208 (&sb->s_type->i_mutex_key#11){+.+.}-{4:4}, at: sock_close+0x40/0xf0 [ 1.550530] #1: ffff88800b148f98 (sk_lock-AF_INET6){+.+.}-{0:0}, at: tcp_close+0x19/0x80 [ 1.550731] #2: ffff88800b148f18 (slock-AF_INET6){+.-.}-{3:3}, at: __tcp_close+0x185/0x4b0 [ 1.550921] stack backtrace: [ 1.550990] CPU: 0 UID: 0 PID: 149 Comm: ncdevmem Tainted: G W 6.13.0-rc3-00770-gbc9ef9606dc9-dirty torvalds#15 [ 1.551233] Tainted: [W]=WARN [ 1.551304] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Arch Linux 1.16.3-1-1 04/01/2014 [ 1.551518] Call Trace: [ 1.551584] <TASK> [ 1.551636] dump_stack_lvl+0x86/0xc0 [ 1.551723] __lock_acquire+0xb0f/0xc30 [ 1.551814] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.551941] lock_acquire+0xf1/0x2a0 [ 1.552026] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552152] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552281] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552408] __ww_mutex_lock+0x121/0x1060 [ 1.552503] ? dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552648] ww_mutex_lock+0x3d/0xa0 [ 1.552733] dma_buf_unmap_attachment_unlocked+0x4b/0x90 [ 1.552857] __net_devmem_dmabuf_binding_free+0x56/0xb0 [ 1.552979] skb_release_data+0x120/0x1f0 [ 1.553074] __kfree_skb+0x29/0xa0 [ 1.553156] tcp_write_queue_purge+0x41/0x310 [ 1.553259] tcp_v4_destroy_sock+0x127/0x320 [ 1.553363] ? __tcp_close+0x169/0x4b0 [ 1.553452] inet_csk_destroy_sock+0x53/0x130 [ 1.553560] __tcp_close+0x421/0x4b0 [ 1.553646] tcp_close+0x24/0x80 [ 1.553724] inet_release+0x5d/0x90 [ 1.553806] sock_close+0x4a/0xf0 [ 1.553886] __fput+0x9c/0x2b0 [ 1.553960] task_work_run+0x89/0xc0 [ 1.554046] do_exit+0x27f/0x980 [ 1.554125] do_group_exit+0xa4/0xb0 [ 1.554211] __x64_sys_exit_group+0x17/0x20 [ 1.554309] x64_sys_call+0x21a0/0x21a0 [ 1.554400] do_syscall_64+0xec/0x1d0 [ 1.554487] ? exc_page_fault+0x8a/0xf0 [ 1.554585] entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 1.554703] RIP: 0033:0x7f2f8a27abcd Resolve this by making __net_devmem_dmabuf_binding_free schedule_work'd. Suggested-by: Stanislav Fomichev <[email protected]> Signed-off-by: Mina Almasry <[email protected]> Acked-by: Stanislav Fomichev <[email protected]> Signed-off-by: NipaLocal <nipa@local>

…ge_order() Patch series "mm: MM owner tracking for large folios (!hugetlb) + CONFIG_NO_PAGE_MAPCOUNT", v3. Let's add an "easy" way to decide -- without false positives, without page-mapcounts and without page table/rmap scanning -- whether a large folio is "certainly mapped exclusively" into a single MM, or whether it "maybe mapped shared" into multiple MMs. Use that information to implement Copy-on-Write reuse, to convert folio_likely_mapped_shared() to folio_maybe_mapped_share(), and to introduce a kernel config option that lets us not use+maintain per-page mapcounts in large folios anymore. The bigger picture was presented at LSF/MM [1]. This series is effectively a follow-up on my early work [2], which implemented a more precise, but also more complicated, way to identify whether a large folio is "mapped shared" into multiple MMs or "mapped exclusively" into a single MM. 1 Patch Organization ==================== Patch #1 -> torvalds#6: make more room in order-1 folios, so we have two "unsigned long" available for our purposes Patch torvalds#7 -> torvalds#11: preparations Patch torvalds#12: MM owner tracking for large folios Patch torvalds#13: COW reuse for PTE-mapped anon THP Patch torvalds#14: folio_maybe_mapped_shared() Patch torvalds#15 -> torvalds#20: introduce and implement CONFIG_NO_PAGE_MAPCOUNT 2 MM owner tracking =================== We assign each MM a unique ID ("MM ID"), to be able to squeeze more information in our folios. On 32bit we use 15-bit IDs, on 64bit we use 31-bit IDs. For each large folios, we now store two MM-ID+mapcount ("slot") combinations: * mm0_id + mm0_mapcount * mm1_id + mm1_mapcount On 32bit, we use a 16-bit per-MM mapcount, on 64bit an ordinary 32bit mapcount. This way, we require 2x "unsigned long" on 32bit and 64bit for both slots. Paired with the large mapcount, we can reliably identify whether one of these MMs is the current owner (-> owns all mappings) or even holds all folio references (-> owns all mappings, and all references are from mappings). As long as only two MMs map folio pages at a time, we can reliably and precisely identify whether a large folio is "mapped shared" or "mapped exclusively". Any additional MM that starts mapping the folio while there are no free slots becomes an "untracked MM". If one such "untracked MM" is the last one mapping a folio exclusively, we will not detect the folio as "mapped exclusively" but instead as "maybe mapped shared". (exception: only a single mapping remains) So that's where the approach gets imprecise. For now, we use a bit-spinlock to sync the large mapcount + slots, and make sure we do keep the machinery fast, to not degrade (un)map performance drastically: for example, we make sure to only use a single atomic (when grabbing the bit-spinlock), like we would already perform when updating the large mapcount. 3 CONFIG_NO_PAGE_MAPCOUNT ========================= patch torvalds#15 -> torvalds#20 spell out and document what exactly is affected when not maintaining the per-page mapcounts in large folios anymore. Most importantly, as we cannot maintain folio->_nr_pages_mapped anymore when (un)mapping pages, we'll account a complete folio as mapped if a single page is mapped. In addition, we'll not detect partially mapped anonymous folios as such in all cases yet. Likely less relevant changes include that we might now under-estimate the USS (Unique Set Size) of a process, but never over-estimate it. The goal is to make CONFIG_NO_PAGE_MAPCOUNT the default at some point, to then slowly make it the only option, as we learn about real-life impacts and possible ways to mitigate them. 4 Performance ============= Detailed performance numbers were included in v1 [3], and not that much changed between v1 and v2. I did plenty of measurements on different systems in the meantime, that all revealed slightly different results. The pte-mapped-folio micro-benchmarks [4] are fairly sensitive to code layout changes on some systems. Especially the fork() benchmark started being more-shaky-than-before on recent kernels for some reason. In summary, with my micro-benchmarks: * Small folios are not impacted. * CoW performance seems to be mostly unchanged across all folios sizes. * CoW reuse performance of large folios now matches CoW reuse performance of small folios, because we now actually implement the CoW reuse optimization. On an Intel Xeon Silver 4210R I measured a ~65% reduction in runtime, on an arm64 system I measured ~54% reduction. * munmap() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~30% on an Intel Xeon Silver 4210R and up to ~70% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * munmao() performance very slightly (couple percent) degrades without CONFIG_NO_PAGE_MAPCOUNT for smaller folios. For larger folios, there seems to be no change at all. * fork() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~20% on an Intel Xeon Silver 4210R and up to ~10% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * While fork() performance without CONFIG_NO_PAGE_MAPCOUNT seems to be almost unchanged on some systems, I saw some degradation for smaller folios on the AmpereOne A192-32X. I did not investigate the details yet, but I suspect code layout changes or suboptimal code placement / inlining. I'm not to worried about the fork() micro-benchmarks for smaller folios given how shaky the results are lately and by how much we improved fork() performance recently. I also ran case-anon-cow-rand and case-anon-cow-seq part of vm-scalability, to assess the scalability and the impact of the bit-spinlock. My measurements on a two 2-socket 10-core Intel Xeon Silver 4210R CPU revealed no significant changes. Similarly, running these benchmarks with 2 MiB THPs enabled on the AmpereOne A192-32X with 192 cores, I got < 1% difference with < 1% stdev, which is nice. So far, I did not get my hands on a similarly large system with multiple sockets. I found no other fitting scalability benchmarks that seem to really hammer on concurrent mapping/unmapping of large folio pages like case-anon-cow-seq does. 5 Concerns ========== 5.1 Bit spinlock ---------------- I'm not quite happy about the bit-spinlock, but so far it does not seem to affect scalability in my measurements. If it ever becomes a problem we could either investigate improving the locking, or simply stopping the MM tracking once there are "too many mappings" and simply assume that the folio is "mapped shared" until it was freed. This would be similar (but slightly different) to the "0,1,2,stopped" counting idea Willy had at some point. Adding that logic to "stop tracking" adds more code to the hot path, so I avoided that for now. 5.2 folio_maybe_mapped_shared() ------------------------------- I documented the change from folio_likely_mapped_shared() to folio_maybe_mapped_shared() quite extensively. If we run into surprises, I have some ideas on how to resolve them. For now, I think we should be fine. 5.3 Added code to map/unmap hot path ------------------------------------ So far, it looks like the added code on the rmap hot path does not really seem to matter much in the bigger picture. I'd like to further reduce it (and possibly improve fork() performance further), but I don't easily see how right now. Well, and I am out of puff 🙂 Having that said, alternatives I considered (e.g., per-MM per-folio mapcount) would add a lot more overhead to these hot paths. 6 Future Work ============= 6.1 Large mapcount ------------------ It would be very handy if the large mapcount would count how often folio pages are actually mapped into page tables: a PMD on x86-64 would count 512 times. Calculating the average per-page mapcount will be easy, and remapping (PMD->PTE) folios would get even faster. That would also remove the need for the entire mapcount (except for PMD-sized folios for memory statistics reasons ...), and allow for mapping folios larger than PMDs (e.g., 4 MiB) easily. We likely would also have to take the same number of folio references to make our folio_mapcount() == folio_ref_count() work, and we'd want to be able to avoid mapcount+refcount overflows: this could already become an issue with pte-mapped PUD-sized folios (fsdax). One approach we discussed in the THP cabal meeting is (1) extending the mapcount for large folios to 64bit (at least on 64bit systems) and (2) keeping the refcount at 32bit, but (3) having exactly one reference if the the mapcount != 0. It should be doable, but there are some corner cases to consider on the unmap path; it is something that I will be looking into next. 6.2 hugetlb ----------- I'd love to make use of the same tracking also for hugetlb. The real problem is PMD table sharing: getting a page mapped by MM X and unmapped by MM Y will not work. With mshare, that problem should not exist (all mapping/unmapping will be routed through the mshare MM). [1] https://lwn.net/Articles/974223/ [2] https://lore.kernel.org/linux-mm/[email protected]/T/ [3] https://lkml.kernel.org/r/[email protected] [4] https://gitlab.com/davidhildenbrand/scratchspace/-/raw/main/pte-mapped-folio-benchmarks.c This patch (of 20): Let's factor it out into a simple helper function. This helper will also come in handy when working with code where we know that our folio is large. Maybe in the future we'll have the order readily available for small and large folios; in that case, folio_large_order() would simply translate to folio_order(). Link: https://lkml.kernel.org/r/[email protected] Link: https://lkml.kernel.org/r/[email protected] Signed-off-by: David Hildenbrand <[email protected]> Reviewed-by: Lance Yang <[email protected]> Reviewed-by: Kirill A. Shutemov <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Andy Lutomirks^H^Hski <[email protected]> Cc: Borislav Betkov <[email protected]> Cc: Dave Hansen <[email protected]> Cc: David Hildenbrand <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Jann Horn <[email protected]> Cc: Johannes Weiner <[email protected]> Cc: Jonathan Corbet <[email protected]> Cc: Liam Howlett <[email protected]> Cc: Lorenzo Stoakes <[email protected]> Cc: Matthew Wilcow (Oracle) <[email protected]> Cc: Michal Koutn <[email protected]> Cc: Muchun Song <[email protected]> Cc: tejun heo <[email protected]> Cc: Vlastimil Babka <[email protected]> Cc: Zefan Li <[email protected]> Signed-off-by: Andrew Morton <[email protected]>

Nikolay has reported a hang when a memcg reclaim got stuck with the following backtrace: PID: 18308 TASK: ffff883d7c9b0a30 CPU: 1 COMMAND: "rsync" #0 __schedule at ffffffff815ab152 #1 schedule at ffffffff815ab76e #2 schedule_timeout at ffffffff815ae5e5 #3 io_schedule_timeout at ffffffff815aad6a #4 bit_wait_io at ffffffff815abfc6 #5 __wait_on_bit at ffffffff815abda5 torvalds#6 wait_on_page_bit at ffffffff8111fd4f torvalds#7 shrink_page_list at ffffffff81135445 torvalds#8 shrink_inactive_list at ffffffff81135845 torvalds#9 shrink_lruvec at ffffffff81135ead torvalds#10 shrink_zone at ffffffff811360c3 torvalds#11 shrink_zones at ffffffff81136eff torvalds#12 do_try_to_free_pages at ffffffff8113712f torvalds#13 try_to_free_mem_cgroup_pages at ffffffff811372be torvalds#14 try_charge at ffffffff81189423 torvalds#15 mem_cgroup_try_charge at ffffffff8118c6f5 torvalds#16 __add_to_page_cache_locked at ffffffff8112137d torvalds#17 add_to_page_cache_lru at ffffffff81121618 torvalds#18 pagecache_get_page at ffffffff8112170b torvalds#19 grow_dev_page at ffffffff811c8297 torvalds#20 __getblk_slow at ffffffff811c91d6 torvalds#21 __getblk_gfp at ffffffff811c92c1 torvalds#22 ext4_ext_grow_indepth at ffffffff8124565c torvalds#23 ext4_ext_create_new_leaf at ffffffff81246ca8 torvalds#24 ext4_ext_insert_extent at ffffffff81246f09 torvalds#25 ext4_ext_map_blocks at ffffffff8124a848 torvalds#26 ext4_map_blocks at ffffffff8121a5b7 torvalds#27 mpage_map_one_extent at ffffffff8121b1fa torvalds#28 mpage_map_and_submit_extent at ffffffff8121f07b torvalds#29 ext4_writepages at ffffffff8121f6d5 torvalds#30 do_writepages at ffffffff8112c490 torvalds#31 __filemap_fdatawrite_range at ffffffff81120199 torvalds#32 filemap_flush at ffffffff8112041c torvalds#33 ext4_alloc_da_blocks at ffffffff81219da1 torvalds#34 ext4_rename at ffffffff81229b91 torvalds#35 ext4_rename2 at ffffffff81229e32 torvalds#36 vfs_rename at ffffffff811a08a5 torvalds#37 SYSC_renameat2 at ffffffff811a3ffc torvalds#38 sys_renameat2 at ffffffff811a408e torvalds#39 sys_rename at ffffffff8119e51e torvalds#40 system_call_fastpath at ffffffff815afa89 Dave Chinner has properly pointed out that this is a deadlock in the reclaim code because ext4 doesn't submit pages which are marked by PG_writeback right away. The heuristic was introduced by commit e62e384 ("memcg: prevent OOM with too many dirty pages") and it was applied only when may_enter_fs was specified. The code has been changed by c3b94f4 ("memcg: further prevent OOM with too many dirty pages") which has removed the __GFP_FS restriction with a reasoning that we do not get into the fs code. But this is not sufficient apparently because the fs doesn't necessarily submit pages marked PG_writeback for IO right away. ext4_bio_write_page calls io_submit_add_bh but that doesn't necessarily submit the bio. Instead it tries to map more pages into the bio and mpage_map_one_extent might trigger memcg charge which might end up waiting on a page which is marked PG_writeback but hasn't been submitted yet so we would end up waiting for something that never finishes. Fix this issue by replacing __GFP_IO by may_enter_fs check (for case 2) before we go to wait on the writeback. The page fault path, which is the only path that triggers memcg oom killer since 3.12, shouldn't require GFP_NOFS and so we shouldn't reintroduce the premature OOM killer issue which was originally addressed by the heuristic. As per David Chinner the xfs is doing similar thing since 2.6.15 already so ext4 is not the only affected filesystem. Moreover he notes: : For example: IO completion might require unwritten extent conversion : which executes filesystem transactions and GFP_NOFS allocations. The : writeback flag on the pages can not be cleared until unwritten : extent conversion completes. Hence memory reclaim cannot wait on : page writeback to complete in GFP_NOFS context because it is not : safe to do so, memcg reclaim or otherwise. Cc: [email protected] # 3.9+ [[email protected]: corrected the control flow] Fixes: c3b94f4 ("memcg: further prevent OOM with too many dirty pages") Reported-by: Nikolay Borisov <[email protected]> Signed-off-by: Michal Hocko <[email protected]> Signed-off-by: Hugh Dickins <[email protected]> Signed-off-by: Linus Torvalds <[email protected]> (cherry picked from commit ecf5fc6) Signed-off-by: Alex Shi <[email protected]>

Ian told me that there are many memory leaks in the hierarchy mode. I can easily reproduce it with the follwing command. $ make DEBUG=1 EXTRA_CFLAGS=-fsanitize=leak $ perf record --latency -g -- ./perf test -w thloop $ perf report -H --stdio ... Indirect leak of 168 byte(s) in 21 object(s) allocated from: #0 0x7f3414c16c65 in malloc ../../../../src/libsanitizer/lsan/lsan_interceptors.cpp:75 #1 0x55ed3602346e in map__get util/map.h:189 #2 0x55ed36024cc4 in hist_entry__init util/hist.c:476 #3 0x55ed36025208 in hist_entry__new util/hist.c:588 #4 0x55ed36027c05 in hierarchy_insert_entry util/hist.c:1587 #5 0x55ed36027e2e in hists__hierarchy_insert_entry util/hist.c:1638 torvalds#6 0x55ed36027fa4 in hists__collapse_insert_entry util/hist.c:1685 torvalds#7 0x55ed360283e8 in hists__collapse_resort util/hist.c:1776 torvalds#8 0x55ed35de0323 in report__collapse_hists /home/namhyung/project/linux/tools/perf/builtin-report.c:735 torvalds#9 0x55ed35de15b4 in __cmd_report /home/namhyung/project/linux/tools/perf/builtin-report.c:1119 torvalds#10 0x55ed35de43dc in cmd_report /home/namhyung/project/linux/tools/perf/builtin-report.c:1867 torvalds#11 0x55ed35e66767 in run_builtin /home/namhyung/project/linux/tools/perf/perf.c:351 torvalds#12 0x55ed35e66a0e in handle_internal_command /home/namhyung/project/linux/tools/perf/perf.c:404 torvalds#13 0x55ed35e66b67 in run_argv /home/namhyung/project/linux/tools/perf/perf.c:448 torvalds#14 0x55ed35e66eb0 in main /home/namhyung/project/linux/tools/perf/perf.c:556 torvalds#15 0x7f340ac33d67 in __libc_start_call_main ../sysdeps/nptl/libc_start_call_main.h:58 ... $ perf report -H --stdio 2>&1 | grep -c '^Indirect leak' 93 I found that hist_entry__delete() missed to release child entries in the hierarchy tree (hroot_{in,out}). It needs to iterate the child entries and call hist_entry__delete() recursively. After this change: $ perf report -H --stdio 2>&1 | grep -c '^Indirect leak' 0 Reported-by: Ian Rogers <[email protected]> Signed-off-by: Namhyung Kim <[email protected]>

…ge_order() Patch series "mm: MM owner tracking for large folios (!hugetlb) + CONFIG_NO_PAGE_MAPCOUNT", v3. Let's add an "easy" way to decide -- without false positives, without page-mapcounts and without page table/rmap scanning -- whether a large folio is "certainly mapped exclusively" into a single MM, or whether it "maybe mapped shared" into multiple MMs. Use that information to implement Copy-on-Write reuse, to convert folio_likely_mapped_shared() to folio_maybe_mapped_share(), and to introduce a kernel config option that lets us not use+maintain per-page mapcounts in large folios anymore. The bigger picture was presented at LSF/MM [1]. This series is effectively a follow-up on my early work [2], which implemented a more precise, but also more complicated, way to identify whether a large folio is "mapped shared" into multiple MMs or "mapped exclusively" into a single MM. 1 Patch Organization ==================== Patch #1 -> torvalds#6: make more room in order-1 folios, so we have two "unsigned long" available for our purposes Patch torvalds#7 -> torvalds#11: preparations Patch torvalds#12: MM owner tracking for large folios Patch torvalds#13: COW reuse for PTE-mapped anon THP Patch torvalds#14: folio_maybe_mapped_shared() Patch torvalds#15 -> torvalds#20: introduce and implement CONFIG_NO_PAGE_MAPCOUNT 2 MM owner tracking =================== We assign each MM a unique ID ("MM ID"), to be able to squeeze more information in our folios. On 32bit we use 15-bit IDs, on 64bit we use 31-bit IDs. For each large folios, we now store two MM-ID+mapcount ("slot") combinations: * mm0_id + mm0_mapcount * mm1_id + mm1_mapcount On 32bit, we use a 16-bit per-MM mapcount, on 64bit an ordinary 32bit mapcount. This way, we require 2x "unsigned long" on 32bit and 64bit for both slots. Paired with the large mapcount, we can reliably identify whether one of these MMs is the current owner (-> owns all mappings) or even holds all folio references (-> owns all mappings, and all references are from mappings). As long as only two MMs map folio pages at a time, we can reliably and precisely identify whether a large folio is "mapped shared" or "mapped exclusively". Any additional MM that starts mapping the folio while there are no free slots becomes an "untracked MM". If one such "untracked MM" is the last one mapping a folio exclusively, we will not detect the folio as "mapped exclusively" but instead as "maybe mapped shared". (exception: only a single mapping remains) So that's where the approach gets imprecise. For now, we use a bit-spinlock to sync the large mapcount + slots, and make sure we do keep the machinery fast, to not degrade (un)map performance drastically: for example, we make sure to only use a single atomic (when grabbing the bit-spinlock), like we would already perform when updating the large mapcount. 3 CONFIG_NO_PAGE_MAPCOUNT ========================= patch torvalds#15 -> torvalds#20 spell out and document what exactly is affected when not maintaining the per-page mapcounts in large folios anymore. Most importantly, as we cannot maintain folio->_nr_pages_mapped anymore when (un)mapping pages, we'll account a complete folio as mapped if a single page is mapped. In addition, we'll not detect partially mapped anonymous folios as such in all cases yet. Likely less relevant changes include that we might now under-estimate the USS (Unique Set Size) of a process, but never over-estimate it. The goal is to make CONFIG_NO_PAGE_MAPCOUNT the default at some point, to then slowly make it the only option, as we learn about real-life impacts and possible ways to mitigate them. 4 Performance ============= Detailed performance numbers were included in v1 [3], and not that much changed between v1 and v2. I did plenty of measurements on different systems in the meantime, that all revealed slightly different results. The pte-mapped-folio micro-benchmarks [4] are fairly sensitive to code layout changes on some systems. Especially the fork() benchmark started being more-shaky-than-before on recent kernels for some reason. In summary, with my micro-benchmarks: * Small folios are not impacted. * CoW performance seems to be mostly unchanged across all folios sizes. * CoW reuse performance of large folios now matches CoW reuse performance of small folios, because we now actually implement the CoW reuse optimization. On an Intel Xeon Silver 4210R I measured a ~65% reduction in runtime, on an arm64 system I measured ~54% reduction. * munmap() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~30% on an Intel Xeon Silver 4210R and up to ~70% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * munmao() performance very slightly (couple percent) degrades without CONFIG_NO_PAGE_MAPCOUNT for smaller folios. For larger folios, there seems to be no change at all. * fork() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~20% on an Intel Xeon Silver 4210R and up to ~10% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * While fork() performance without CONFIG_NO_PAGE_MAPCOUNT seems to be almost unchanged on some systems, I saw some degradation for smaller folios on the AmpereOne A192-32X. I did not investigate the details yet, but I suspect code layout changes or suboptimal code placement / inlining. I'm not to worried about the fork() micro-benchmarks for smaller folios given how shaky the results are lately and by how much we improved fork() performance recently. I also ran case-anon-cow-rand and case-anon-cow-seq part of vm-scalability, to assess the scalability and the impact of the bit-spinlock. My measurements on a two 2-socket 10-core Intel Xeon Silver 4210R CPU revealed no significant changes. Similarly, running these benchmarks with 2 MiB THPs enabled on the AmpereOne A192-32X with 192 cores, I got < 1% difference with < 1% stdev, which is nice. So far, I did not get my hands on a similarly large system with multiple sockets. I found no other fitting scalability benchmarks that seem to really hammer on concurrent mapping/unmapping of large folio pages like case-anon-cow-seq does. 5 Concerns ========== 5.1 Bit spinlock ---------------- I'm not quite happy about the bit-spinlock, but so far it does not seem to affect scalability in my measurements. If it ever becomes a problem we could either investigate improving the locking, or simply stopping the MM tracking once there are "too many mappings" and simply assume that the folio is "mapped shared" until it was freed. This would be similar (but slightly different) to the "0,1,2,stopped" counting idea Willy had at some point. Adding that logic to "stop tracking" adds more code to the hot path, so I avoided that for now. 5.2 folio_maybe_mapped_shared() ------------------------------- I documented the change from folio_likely_mapped_shared() to folio_maybe_mapped_shared() quite extensively. If we run into surprises, I have some ideas on how to resolve them. For now, I think we should be fine. 5.3 Added code to map/unmap hot path ------------------------------------ So far, it looks like the added code on the rmap hot path does not really seem to matter much in the bigger picture. I'd like to further reduce it (and possibly improve fork() performance further), but I don't easily see how right now. Well, and I am out of puff 🙂 Having that said, alternatives I considered (e.g., per-MM per-folio mapcount) would add a lot more overhead to these hot paths. 6 Future Work ============= 6.1 Large mapcount ------------------ It would be very handy if the large mapcount would count how often folio pages are actually mapped into page tables: a PMD on x86-64 would count 512 times. Calculating the average per-page mapcount will be easy, and remapping (PMD->PTE) folios would get even faster. That would also remove the need for the entire mapcount (except for PMD-sized folios for memory statistics reasons ...), and allow for mapping folios larger than PMDs (e.g., 4 MiB) easily. We likely would also have to take the same number of folio references to make our folio_mapcount() == folio_ref_count() work, and we'd want to be able to avoid mapcount+refcount overflows: this could already become an issue with pte-mapped PUD-sized folios (fsdax). One approach we discussed in the THP cabal meeting is (1) extending the mapcount for large folios to 64bit (at least on 64bit systems) and (2) keeping the refcount at 32bit, but (3) having exactly one reference if the the mapcount != 0. It should be doable, but there are some corner cases to consider on the unmap path; it is something that I will be looking into next. 6.2 hugetlb ----------- I'd love to make use of the same tracking also for hugetlb. The real problem is PMD table sharing: getting a page mapped by MM X and unmapped by MM Y will not work. With mshare, that problem should not exist (all mapping/unmapping will be routed through the mshare MM). [1] https://lwn.net/Articles/974223/ [2] https://lore.kernel.org/linux-mm/[email protected]/T/ [3] https://lkml.kernel.org/r/[email protected] [4] https://gitlab.com/davidhildenbrand/scratchspace/-/raw/main/pte-mapped-folio-benchmarks.c This patch (of 20): Let's factor it out into a simple helper function. This helper will also come in handy when working with code where we know that our folio is large. Maybe in the future we'll have the order readily available for small and large folios; in that case, folio_large_order() would simply translate to folio_order(). Link: https://lkml.kernel.org/r/[email protected] Link: https://lkml.kernel.org/r/[email protected] Signed-off-by: David Hildenbrand <[email protected]> Reviewed-by: Lance Yang <[email protected]> Reviewed-by: Kirill A. Shutemov <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Andy Lutomirks^H^Hski <[email protected]> Cc: Borislav Betkov <[email protected]> Cc: Dave Hansen <[email protected]> Cc: David Hildenbrand <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Jann Horn <[email protected]> Cc: Johannes Weiner <[email protected]> Cc: Jonathan Corbet <[email protected]> Cc: Liam Howlett <[email protected]> Cc: Lorenzo Stoakes <[email protected]> Cc: Matthew Wilcow (Oracle) <[email protected]> Cc: Michal Koutn <[email protected]> Cc: Muchun Song <[email protected]> Cc: tejun heo <[email protected]> Cc: Vlastimil Babka <[email protected]> Cc: Zefan Li <[email protected]> Signed-off-by: Andrew Morton <[email protected]>

Ian told me that there are many memory leaks in the hierarchy mode. I can easily reproduce it with the follwing command. $ make DEBUG=1 EXTRA_CFLAGS=-fsanitize=leak $ perf record --latency -g -- ./perf test -w thloop $ perf report -H --stdio ... Indirect leak of 168 byte(s) in 21 object(s) allocated from: #0 0x7f3414c16c65 in malloc ../../../../src/libsanitizer/lsan/lsan_interceptors.cpp:75 #1 0x55ed3602346e in map__get util/map.h:189 #2 0x55ed36024cc4 in hist_entry__init util/hist.c:476 #3 0x55ed36025208 in hist_entry__new util/hist.c:588 #4 0x55ed36027c05 in hierarchy_insert_entry util/hist.c:1587 #5 0x55ed36027e2e in hists__hierarchy_insert_entry util/hist.c:1638 torvalds#6 0x55ed36027fa4 in hists__collapse_insert_entry util/hist.c:1685 torvalds#7 0x55ed360283e8 in hists__collapse_resort util/hist.c:1776 torvalds#8 0x55ed35de0323 in report__collapse_hists /home/namhyung/project/linux/tools/perf/builtin-report.c:735 torvalds#9 0x55ed35de15b4 in __cmd_report /home/namhyung/project/linux/tools/perf/builtin-report.c:1119 torvalds#10 0x55ed35de43dc in cmd_report /home/namhyung/project/linux/tools/perf/builtin-report.c:1867 torvalds#11 0x55ed35e66767 in run_builtin /home/namhyung/project/linux/tools/perf/perf.c:351 torvalds#12 0x55ed35e66a0e in handle_internal_command /home/namhyung/project/linux/tools/perf/perf.c:404 torvalds#13 0x55ed35e66b67 in run_argv /home/namhyung/project/linux/tools/perf/perf.c:448 torvalds#14 0x55ed35e66eb0 in main /home/namhyung/project/linux/tools/perf/perf.c:556 torvalds#15 0x7f340ac33d67 in __libc_start_call_main ../sysdeps/nptl/libc_start_call_main.h:58 ... $ perf report -H --stdio 2>&1 | grep -c '^Indirect leak' 93 I found that hist_entry__delete() missed to release child entries in the hierarchy tree (hroot_{in,out}). It needs to iterate the child entries and call hist_entry__delete() recursively. After this change: $ perf report -H --stdio 2>&1 | grep -c '^Indirect leak' 0 Reported-by: Ian Rogers <[email protected]> Signed-off-by: Namhyung Kim <[email protected]>

…ge_order() Patch series "mm: MM owner tracking for large folios (!hugetlb) + CONFIG_NO_PAGE_MAPCOUNT", v3. Let's add an "easy" way to decide -- without false positives, without page-mapcounts and without page table/rmap scanning -- whether a large folio is "certainly mapped exclusively" into a single MM, or whether it "maybe mapped shared" into multiple MMs. Use that information to implement Copy-on-Write reuse, to convert folio_likely_mapped_shared() to folio_maybe_mapped_share(), and to introduce a kernel config option that lets us not use+maintain per-page mapcounts in large folios anymore. The bigger picture was presented at LSF/MM [1]. This series is effectively a follow-up on my early work [2], which implemented a more precise, but also more complicated, way to identify whether a large folio is "mapped shared" into multiple MMs or "mapped exclusively" into a single MM. 1 Patch Organization ==================== Patch #1 -> torvalds#6: make more room in order-1 folios, so we have two "unsigned long" available for our purposes Patch torvalds#7 -> torvalds#11: preparations Patch torvalds#12: MM owner tracking for large folios Patch torvalds#13: COW reuse for PTE-mapped anon THP Patch torvalds#14: folio_maybe_mapped_shared() Patch torvalds#15 -> torvalds#20: introduce and implement CONFIG_NO_PAGE_MAPCOUNT 2 MM owner tracking =================== We assign each MM a unique ID ("MM ID"), to be able to squeeze more information in our folios. On 32bit we use 15-bit IDs, on 64bit we use 31-bit IDs. For each large folios, we now store two MM-ID+mapcount ("slot") combinations: * mm0_id + mm0_mapcount * mm1_id + mm1_mapcount On 32bit, we use a 16-bit per-MM mapcount, on 64bit an ordinary 32bit mapcount. This way, we require 2x "unsigned long" on 32bit and 64bit for both slots. Paired with the large mapcount, we can reliably identify whether one of these MMs is the current owner (-> owns all mappings) or even holds all folio references (-> owns all mappings, and all references are from mappings). As long as only two MMs map folio pages at a time, we can reliably and precisely identify whether a large folio is "mapped shared" or "mapped exclusively". Any additional MM that starts mapping the folio while there are no free slots becomes an "untracked MM". If one such "untracked MM" is the last one mapping a folio exclusively, we will not detect the folio as "mapped exclusively" but instead as "maybe mapped shared". (exception: only a single mapping remains) So that's where the approach gets imprecise. For now, we use a bit-spinlock to sync the large mapcount + slots, and make sure we do keep the machinery fast, to not degrade (un)map performance drastically: for example, we make sure to only use a single atomic (when grabbing the bit-spinlock), like we would already perform when updating the large mapcount. 3 CONFIG_NO_PAGE_MAPCOUNT ========================= patch torvalds#15 -> torvalds#20 spell out and document what exactly is affected when not maintaining the per-page mapcounts in large folios anymore. Most importantly, as we cannot maintain folio->_nr_pages_mapped anymore when (un)mapping pages, we'll account a complete folio as mapped if a single page is mapped. In addition, we'll not detect partially mapped anonymous folios as such in all cases yet. Likely less relevant changes include that we might now under-estimate the USS (Unique Set Size) of a process, but never over-estimate it. The goal is to make CONFIG_NO_PAGE_MAPCOUNT the default at some point, to then slowly make it the only option, as we learn about real-life impacts and possible ways to mitigate them. 4 Performance ============= Detailed performance numbers were included in v1 [3], and not that much changed between v1 and v2. I did plenty of measurements on different systems in the meantime, that all revealed slightly different results. The pte-mapped-folio micro-benchmarks [4] are fairly sensitive to code layout changes on some systems. Especially the fork() benchmark started being more-shaky-than-before on recent kernels for some reason. In summary, with my micro-benchmarks: * Small folios are not impacted. * CoW performance seems to be mostly unchanged across all folios sizes. * CoW reuse performance of large folios now matches CoW reuse performance of small folios, because we now actually implement the CoW reuse optimization. On an Intel Xeon Silver 4210R I measured a ~65% reduction in runtime, on an arm64 system I measured ~54% reduction. * munmap() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~30% on an Intel Xeon Silver 4210R and up to ~70% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * munmao() performance very slightly (couple percent) degrades without CONFIG_NO_PAGE_MAPCOUNT for smaller folios. For larger folios, there seems to be no change at all. * fork() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~20% on an Intel Xeon Silver 4210R and up to ~10% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * While fork() performance without CONFIG_NO_PAGE_MAPCOUNT seems to be almost unchanged on some systems, I saw some degradation for smaller folios on the AmpereOne A192-32X. I did not investigate the details yet, but I suspect code layout changes or suboptimal code placement / inlining. I'm not to worried about the fork() micro-benchmarks for smaller folios given how shaky the results are lately and by how much we improved fork() performance recently. I also ran case-anon-cow-rand and case-anon-cow-seq part of vm-scalability, to assess the scalability and the impact of the bit-spinlock. My measurements on a two 2-socket 10-core Intel Xeon Silver 4210R CPU revealed no significant changes. Similarly, running these benchmarks with 2 MiB THPs enabled on the AmpereOne A192-32X with 192 cores, I got < 1% difference with < 1% stdev, which is nice. So far, I did not get my hands on a similarly large system with multiple sockets. I found no other fitting scalability benchmarks that seem to really hammer on concurrent mapping/unmapping of large folio pages like case-anon-cow-seq does. 5 Concerns ========== 5.1 Bit spinlock ---------------- I'm not quite happy about the bit-spinlock, but so far it does not seem to affect scalability in my measurements. If it ever becomes a problem we could either investigate improving the locking, or simply stopping the MM tracking once there are "too many mappings" and simply assume that the folio is "mapped shared" until it was freed. This would be similar (but slightly different) to the "0,1,2,stopped" counting idea Willy had at some point. Adding that logic to "stop tracking" adds more code to the hot path, so I avoided that for now. 5.2 folio_maybe_mapped_shared() ------------------------------- I documented the change from folio_likely_mapped_shared() to folio_maybe_mapped_shared() quite extensively. If we run into surprises, I have some ideas on how to resolve them. For now, I think we should be fine. 5.3 Added code to map/unmap hot path ------------------------------------ So far, it looks like the added code on the rmap hot path does not really seem to matter much in the bigger picture. I'd like to further reduce it (and possibly improve fork() performance further), but I don't easily see how right now. Well, and I am out of puff 🙂 Having that said, alternatives I considered (e.g., per-MM per-folio mapcount) would add a lot more overhead to these hot paths. 6 Future Work ============= 6.1 Large mapcount ------------------ It would be very handy if the large mapcount would count how often folio pages are actually mapped into page tables: a PMD on x86-64 would count 512 times. Calculating the average per-page mapcount will be easy, and remapping (PMD->PTE) folios would get even faster. That would also remove the need for the entire mapcount (except for PMD-sized folios for memory statistics reasons ...), and allow for mapping folios larger than PMDs (e.g., 4 MiB) easily. We likely would also have to take the same number of folio references to make our folio_mapcount() == folio_ref_count() work, and we'd want to be able to avoid mapcount+refcount overflows: this could already become an issue with pte-mapped PUD-sized folios (fsdax). One approach we discussed in the THP cabal meeting is (1) extending the mapcount for large folios to 64bit (at least on 64bit systems) and (2) keeping the refcount at 32bit, but (3) having exactly one reference if the the mapcount != 0. It should be doable, but there are some corner cases to consider on the unmap path; it is something that I will be looking into next. 6.2 hugetlb ----------- I'd love to make use of the same tracking also for hugetlb. The real problem is PMD table sharing: getting a page mapped by MM X and unmapped by MM Y will not work. With mshare, that problem should not exist (all mapping/unmapping will be routed through the mshare MM). [1] https://lwn.net/Articles/974223/ [2] https://lore.kernel.org/linux-mm/[email protected]/T/ [3] https://lkml.kernel.org/r/[email protected] [4] https://gitlab.com/davidhildenbrand/scratchspace/-/raw/main/pte-mapped-folio-benchmarks.c This patch (of 20): Let's factor it out into a simple helper function. This helper will also come in handy when working with code where we know that our folio is large. Maybe in the future we'll have the order readily available for small and large folios; in that case, folio_large_order() would simply translate to folio_order(). Link: https://lkml.kernel.org/r/[email protected] Link: https://lkml.kernel.org/r/[email protected] Signed-off-by: David Hildenbrand <[email protected]> Reviewed-by: Lance Yang <[email protected]> Reviewed-by: Kirill A. Shutemov <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Andy Lutomirks^H^Hski <[email protected]> Cc: Borislav Betkov <[email protected]> Cc: Dave Hansen <[email protected]> Cc: David Hildenbrand <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Jann Horn <[email protected]> Cc: Johannes Weiner <[email protected]> Cc: Jonathan Corbet <[email protected]> Cc: Liam Howlett <[email protected]> Cc: Lorenzo Stoakes <[email protected]> Cc: Matthew Wilcow (Oracle) <[email protected]> Cc: Michal Koutn <[email protected]> Cc: Muchun Song <[email protected]> Cc: tejun heo <[email protected]> Cc: Vlastimil Babka <[email protected]> Cc: Zefan Li <[email protected]> Signed-off-by: Andrew Morton <[email protected]>

Ian told me that there are many memory leaks in the hierarchy mode. I can easily reproduce it with the follwing command. $ make DEBUG=1 EXTRA_CFLAGS=-fsanitize=leak $ perf record --latency -g -- ./perf test -w thloop $ perf report -H --stdio ... Indirect leak of 168 byte(s) in 21 object(s) allocated from: #0 0x7f3414c16c65 in malloc ../../../../src/libsanitizer/lsan/lsan_interceptors.cpp:75 #1 0x55ed3602346e in map__get util/map.h:189 #2 0x55ed36024cc4 in hist_entry__init util/hist.c:476 #3 0x55ed36025208 in hist_entry__new util/hist.c:588 #4 0x55ed36027c05 in hierarchy_insert_entry util/hist.c:1587 #5 0x55ed36027e2e in hists__hierarchy_insert_entry util/hist.c:1638 torvalds#6 0x55ed36027fa4 in hists__collapse_insert_entry util/hist.c:1685 torvalds#7 0x55ed360283e8 in hists__collapse_resort util/hist.c:1776 torvalds#8 0x55ed35de0323 in report__collapse_hists /home/namhyung/project/linux/tools/perf/builtin-report.c:735 torvalds#9 0x55ed35de15b4 in __cmd_report /home/namhyung/project/linux/tools/perf/builtin-report.c:1119 torvalds#10 0x55ed35de43dc in cmd_report /home/namhyung/project/linux/tools/perf/builtin-report.c:1867 torvalds#11 0x55ed35e66767 in run_builtin /home/namhyung/project/linux/tools/perf/perf.c:351 torvalds#12 0x55ed35e66a0e in handle_internal_command /home/namhyung/project/linux/tools/perf/perf.c:404 torvalds#13 0x55ed35e66b67 in run_argv /home/namhyung/project/linux/tools/perf/perf.c:448 torvalds#14 0x55ed35e66eb0 in main /home/namhyung/project/linux/tools/perf/perf.c:556 torvalds#15 0x7f340ac33d67 in __libc_start_call_main ../sysdeps/nptl/libc_start_call_main.h:58 ... $ perf report -H --stdio 2>&1 | grep -c '^Indirect leak' 93 I found that hist_entry__delete() missed to release child entries in the hierarchy tree (hroot_{in,out}). It needs to iterate the child entries and call hist_entry__delete() recursively. After this change: $ perf report -H --stdio 2>&1 | grep -c '^Indirect leak' 0 Reported-by: Ian Rogers <[email protected]> Tested-by Thomas Falcon <[email protected]> Reviewed-by: Ian Rogers <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Namhyung Kim <[email protected]>

…ge_order() Patch series "mm: MM owner tracking for large folios (!hugetlb) + CONFIG_NO_PAGE_MAPCOUNT", v3. Let's add an "easy" way to decide -- without false positives, without page-mapcounts and without page table/rmap scanning -- whether a large folio is "certainly mapped exclusively" into a single MM, or whether it "maybe mapped shared" into multiple MMs. Use that information to implement Copy-on-Write reuse, to convert folio_likely_mapped_shared() to folio_maybe_mapped_share(), and to introduce a kernel config option that lets us not use+maintain per-page mapcounts in large folios anymore. The bigger picture was presented at LSF/MM [1]. This series is effectively a follow-up on my early work [2], which implemented a more precise, but also more complicated, way to identify whether a large folio is "mapped shared" into multiple MMs or "mapped exclusively" into a single MM. 1 Patch Organization ==================== Patch #1 -> torvalds#6: make more room in order-1 folios, so we have two "unsigned long" available for our purposes Patch torvalds#7 -> torvalds#11: preparations Patch torvalds#12: MM owner tracking for large folios Patch torvalds#13: COW reuse for PTE-mapped anon THP Patch torvalds#14: folio_maybe_mapped_shared() Patch torvalds#15 -> torvalds#20: introduce and implement CONFIG_NO_PAGE_MAPCOUNT 2 MM owner tracking =================== We assign each MM a unique ID ("MM ID"), to be able to squeeze more information in our folios. On 32bit we use 15-bit IDs, on 64bit we use 31-bit IDs. For each large folios, we now store two MM-ID+mapcount ("slot") combinations: * mm0_id + mm0_mapcount * mm1_id + mm1_mapcount On 32bit, we use a 16-bit per-MM mapcount, on 64bit an ordinary 32bit mapcount. This way, we require 2x "unsigned long" on 32bit and 64bit for both slots. Paired with the large mapcount, we can reliably identify whether one of these MMs is the current owner (-> owns all mappings) or even holds all folio references (-> owns all mappings, and all references are from mappings). As long as only two MMs map folio pages at a time, we can reliably and precisely identify whether a large folio is "mapped shared" or "mapped exclusively". Any additional MM that starts mapping the folio while there are no free slots becomes an "untracked MM". If one such "untracked MM" is the last one mapping a folio exclusively, we will not detect the folio as "mapped exclusively" but instead as "maybe mapped shared". (exception: only a single mapping remains) So that's where the approach gets imprecise. For now, we use a bit-spinlock to sync the large mapcount + slots, and make sure we do keep the machinery fast, to not degrade (un)map performance drastically: for example, we make sure to only use a single atomic (when grabbing the bit-spinlock), like we would already perform when updating the large mapcount. 3 CONFIG_NO_PAGE_MAPCOUNT ========================= patch torvalds#15 -> torvalds#20 spell out and document what exactly is affected when not maintaining the per-page mapcounts in large folios anymore. Most importantly, as we cannot maintain folio->_nr_pages_mapped anymore when (un)mapping pages, we'll account a complete folio as mapped if a single page is mapped. In addition, we'll not detect partially mapped anonymous folios as such in all cases yet. Likely less relevant changes include that we might now under-estimate the USS (Unique Set Size) of a process, but never over-estimate it. The goal is to make CONFIG_NO_PAGE_MAPCOUNT the default at some point, to then slowly make it the only option, as we learn about real-life impacts and possible ways to mitigate them. 4 Performance ============= Detailed performance numbers were included in v1 [3], and not that much changed between v1 and v2. I did plenty of measurements on different systems in the meantime, that all revealed slightly different results. The pte-mapped-folio micro-benchmarks [4] are fairly sensitive to code layout changes on some systems. Especially the fork() benchmark started being more-shaky-than-before on recent kernels for some reason. In summary, with my micro-benchmarks: * Small folios are not impacted. * CoW performance seems to be mostly unchanged across all folios sizes. * CoW reuse performance of large folios now matches CoW reuse performance of small folios, because we now actually implement the CoW reuse optimization. On an Intel Xeon Silver 4210R I measured a ~65% reduction in runtime, on an arm64 system I measured ~54% reduction. * munmap() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~30% on an Intel Xeon Silver 4210R and up to ~70% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * munmao() performance very slightly (couple percent) degrades without CONFIG_NO_PAGE_MAPCOUNT for smaller folios. For larger folios, there seems to be no change at all. * fork() performance improves with CONFIG_NO_PAGE_MAPCOUNT. I saw double-digit % reduction (up to ~20% on an Intel Xeon Silver 4210R and up to ~10% on an AmpereOne A192-32X) with larger folios. The larger the folios, the larger the performance improvement. * While fork() performance without CONFIG_NO_PAGE_MAPCOUNT seems to be almost unchanged on some systems, I saw some degradation for smaller folios on the AmpereOne A192-32X. I did not investigate the details yet, but I suspect code layout changes or suboptimal code placement / inlining. I'm not to worried about the fork() micro-benchmarks for smaller folios given how shaky the results are lately and by how much we improved fork() performance recently. I also ran case-anon-cow-rand and case-anon-cow-seq part of vm-scalability, to assess the scalability and the impact of the bit-spinlock. My measurements on a two 2-socket 10-core Intel Xeon Silver 4210R CPU revealed no significant changes. Similarly, running these benchmarks with 2 MiB THPs enabled on the AmpereOne A192-32X with 192 cores, I got < 1% difference with < 1% stdev, which is nice. So far, I did not get my hands on a similarly large system with multiple sockets. I found no other fitting scalability benchmarks that seem to really hammer on concurrent mapping/unmapping of large folio pages like case-anon-cow-seq does. 5 Concerns ========== 5.1 Bit spinlock ---------------- I'm not quite happy about the bit-spinlock, but so far it does not seem to affect scalability in my measurements. If it ever becomes a problem we could either investigate improving the locking, or simply stopping the MM tracking once there are "too many mappings" and simply assume that the folio is "mapped shared" until it was freed. This would be similar (but slightly different) to the "0,1,2,stopped" counting idea Willy had at some point. Adding that logic to "stop tracking" adds more code to the hot path, so I avoided that for now. 5.2 folio_maybe_mapped_shared() ------------------------------- I documented the change from folio_likely_mapped_shared() to folio_maybe_mapped_shared() quite extensively. If we run into surprises, I have some ideas on how to resolve them. For now, I think we should be fine. 5.3 Added code to map/unmap hot path ------------------------------------ So far, it looks like the added code on the rmap hot path does not really seem to matter much in the bigger picture. I'd like to further reduce it (and possibly improve fork() performance further), but I don't easily see how right now. Well, and I am out of puff 🙂 Having that said, alternatives I considered (e.g., per-MM per-folio mapcount) would add a lot more overhead to these hot paths. 6 Future Work ============= 6.1 Large mapcount ------------------ It would be very handy if the large mapcount would count how often folio pages are actually mapped into page tables: a PMD on x86-64 would count 512 times. Calculating the average per-page mapcount will be easy, and remapping (PMD->PTE) folios would get even faster. That would also remove the need for the entire mapcount (except for PMD-sized folios for memory statistics reasons ...), and allow for mapping folios larger than PMDs (e.g., 4 MiB) easily. We likely would also have to take the same number of folio references to make our folio_mapcount() == folio_ref_count() work, and we'd want to be able to avoid mapcount+refcount overflows: this could already become an issue with pte-mapped PUD-sized folios (fsdax). One approach we discussed in the THP cabal meeting is (1) extending the mapcount for large folios to 64bit (at least on 64bit systems) and (2) keeping the refcount at 32bit, but (3) having exactly one reference if the the mapcount != 0. It should be doable, but there are some corner cases to consider on the unmap path; it is something that I will be looking into next. 6.2 hugetlb ----------- I'd love to make use of the same tracking also for hugetlb. The real problem is PMD table sharing: getting a page mapped by MM X and unmapped by MM Y will not work. With mshare, that problem should not exist (all mapping/unmapping will be routed through the mshare MM). [1] https://lwn.net/Articles/974223/ [2] https://lore.kernel.org/linux-mm/[email protected]/T/ [3] https://lkml.kernel.org/r/[email protected] [4] https://gitlab.com/davidhildenbrand/scratchspace/-/raw/main/pte-mapped-folio-benchmarks.c This patch (of 20): Let's factor it out into a simple helper function. This helper will also come in handy when working with code where we know that our folio is large. Maybe in the future we'll have the order readily available for small and large folios; in that case, folio_large_order() would simply translate to folio_order(). Link: https://lkml.kernel.org/r/[email protected] Link: https://lkml.kernel.org/r/[email protected] Signed-off-by: David Hildenbrand <[email protected]> Reviewed-by: Lance Yang <[email protected]> Reviewed-by: Kirill A. Shutemov <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Andy Lutomirks^H^Hski <[email protected]> Cc: Borislav Betkov <[email protected]> Cc: Dave Hansen <[email protected]> Cc: David Hildenbrand <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Jann Horn <[email protected]> Cc: Johannes Weiner <[email protected]> Cc: Jonathan Corbet <[email protected]> Cc: Liam Howlett <[email protected]> Cc: Lorenzo Stoakes <[email protected]> Cc: Matthew Wilcow (Oracle) <[email protected]> Cc: Michal Koutn <[email protected]> Cc: Muchun Song <[email protected]> Cc: tejun heo <[email protected]> Cc: Vlastimil Babka <[email protected]> Cc: Zefan Li <[email protected]> Signed-off-by: Andrew Morton <[email protected]>

drm/edid: Fix the HDTV hack yet more.

bbaabb8

Checking for just two variants of standard timings for 1366x768 isn't quite correct, let's check for ranges instead. Signed-off-by: Andrew Shadura <[email protected]>

nomis pushed a commit to nomis/linux that referenced this pull request May 15, 2012

Merge pull request torvalds#15 from steveglen/smsc95xx-patches

3a277d3

Smsc95xx patches

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

drm/edid: Fix the HDTV hack yet more. #15

drm/edid: Fix the HDTV hack yet more. #15

andrewshadura commented Jan 4, 2012

peterwillcn commented Feb 13, 2012

drm/edid: Fix the HDTV hack yet more. #15

drm/edid: Fix the HDTV hack yet more. #15

Conversation

andrewshadura commented Jan 4, 2012

peterwillcn commented Feb 13, 2012