March 2023 - Kernel - mailweb.openeuler.org

[PATCH openEuler-1.0-LTS 1/8] fs/proc: task_mmu.c: don't read mapcount for migration entry
by Yongqiang Liu 17 Mar '23

17 Mar '23

From: Yang Shi <shy828301(a)gmail.com> stable inclusion from stable-v5.10.102 commit db3f3636e4aed2cba3e4e7897a053323f7a62249 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6EVUJ Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit 24d7275ce2791829953ed4e72f68277ceb2571c6 upstream. The syzbot reported the below BUG: kernel BUG at include/linux/page-flags.h:785! invalid opcode: 0000 [#1] PREEMPT SMP KASAN CPU: 1 PID: 4392 Comm: syz-executor560 Not tainted 5.16.0-rc6-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:PageDoubleMap include/linux/page-flags.h:785 [inline] RIP: 0010:__page_mapcount+0x2d2/0x350 mm/util.c:744 Call Trace: page_mapcount include/linux/mm.h:837 [inline] smaps_account+0x470/0xb10 fs/proc/task_mmu.c:466 smaps_pte_entry fs/proc/task_mmu.c:538 [inline] smaps_pte_range+0x611/0x1250 fs/proc/task_mmu.c:601 walk_pmd_range mm/pagewalk.c:128 [inline] walk_pud_range mm/pagewalk.c:205 [inline] walk_p4d_range mm/pagewalk.c:240 [inline] walk_pgd_range mm/pagewalk.c:277 [inline] __walk_page_range+0xe23/0x1ea0 mm/pagewalk.c:379 walk_page_vma+0x277/0x350 mm/pagewalk.c:530 smap_gather_stats.part.0+0x148/0x260 fs/proc/task_mmu.c:768 smap_gather_stats fs/proc/task_mmu.c:741 [inline] show_smap+0xc6/0x440 fs/proc/task_mmu.c:822 seq_read_iter+0xbb0/0x1240 fs/seq_file.c:272 seq_read+0x3e0/0x5b0 fs/seq_file.c:162 vfs_read+0x1b5/0x600 fs/read_write.c:479 ksys_read+0x12d/0x250 fs/read_write.c:619 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae The reproducer was trying to read /proc/$PID/smaps when calling MADV_FREE at the mean time. MADV_FREE may split THPs if it is called for partial THP. It may trigger the below race: CPU A CPU B ----- ----- smaps walk: MADV_FREE: page_mapcount() PageCompound() split_huge_page() page = compound_head(page) PageDoubleMap(page) When calling PageDoubleMap() this page is not a tail page of THP anymore so the BUG is triggered. This could be fixed by elevated refcount of the page before calling mapcount, but that would prevent it from counting migration entries, and it seems overkilling because the race just could happen when PMD is split so all PTE entries of tail pages are actually migration entries, and smaps_account() does treat migration entries as mapcount == 1 as Kirill pointed out. Add a new parameter for smaps_account() to tell this entry is migration entry then skip calling page_mapcount(). Don't skip getting mapcount for device private entries since they do track references with mapcount. Pagemap also has the similar issue although it was not reported. Fixed it as well. [shy828301(a)gmail.com: v4] Link: https://lkml.kernel.org/r/20220203182641.824731-1-shy828301@gmail.com [nathan(a)kernel.org: avoid unused variable warning in pagemap_pmd_range()] Link: https://lkml.kernel.org/r/20220207171049.1102239-1-nathan@kernel.org Link: https://lkml.kernel.org/r/20220120202805.3369-1-shy828301@gmail.com Fixes: e9b61f19858a ("thp: reintroduce split_huge_page()") Signed-off-by: Yang Shi <shy828301(a)gmail.com> Signed-off-by: Nathan Chancellor <nathan(a)kernel.org> Reported-by: syzbot+1f52b3a18d5633fa7f82(a)syzkaller.appspotmail.com Acked-by: David Hildenbrand <david(a)redhat.com> Cc: "Kirill A. Shutemov" <kirill.shutemov(a)linux.intel.com> Cc: Jann Horn <jannh(a)google.com> Cc: Matthew Wilcox <willy(a)infradead.org> Cc: Alexey Dobriyan <adobriyan(a)gmail.com> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> conflicts: fs/proc/task_mmu.c Signed-off-by: Zhang Peng <zhangpeng362(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- fs/proc/task_mmu.c | 37 +++++++++++++++++++++++++++---------- 1 file changed, 27 insertions(+), 10 deletions(-) diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 93815b2b8440..0175fd7b3598 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -429,7 +429,8 @@ struct mem_size_stats { }; static void smaps_account(struct mem_size_stats *mss, struct page *page, - bool compound, bool young, bool dirty, bool locked) + bool compound, bool young, bool dirty, bool locked, + bool migration) { int i, nr = compound ? 1 << compound_order(page) : 1; unsigned long size = nr * PAGE_SIZE; @@ -449,8 +450,15 @@ static void smaps_account(struct mem_size_stats *mss, struct page *page, * page_count(page) == 1 guarantees the page is mapped exactly once. * If any subpage of the compound page mapped with PTE it would elevate * page_count(). + * + * The page_mapcount() is called to get a snapshot of the mapcount. + * Without holding the page lock this snapshot can be slightly wrong as + * we cannot always read the mapcount atomically. It is not safe to + * call page_mapcount() even with PTL held if the page is not mapped, + * especially for migration entries. Treat regular migration entries + * as mapcount == 1. */ - if (page_count(page) == 1) { + if ((page_count(page) == 1) || migration) { if (dirty || PageDirty(page)) mss->private_dirty += size; else @@ -505,6 +513,7 @@ static void smaps_pte_entry(pte_t *pte, unsigned long addr, struct vm_area_struct *vma = walk->vma; bool locked = !!(vma->vm_flags & VM_LOCKED); struct page *page = NULL; + bool migration = false; if (pte_present(*pte)) { page = vm_normal_page(vma, addr, *pte); @@ -524,9 +533,10 @@ static void smaps_pte_entry(pte_t *pte, unsigned long addr, } else { mss->swap_pss += (u64)PAGE_SIZE << PSS_SHIFT; } - } else if (is_migration_entry(swpent)) + } else if (is_migration_entry(swpent)) { + migration = true; page = migration_entry_to_page(swpent); - else if (is_device_private_entry(swpent)) + } else if (is_device_private_entry(swpent)) page = device_private_entry_to_page(swpent); } else if (unlikely(IS_ENABLED(CONFIG_SHMEM) && mss->check_shmem_swap && pte_none(*pte))) { @@ -546,7 +556,8 @@ static void smaps_pte_entry(pte_t *pte, unsigned long addr, if (!page) return; - smaps_account(mss, page, false, pte_young(*pte), pte_dirty(*pte), locked); + smaps_account(mss, page, false, pte_young(*pte), pte_dirty(*pte), + locked, migration); } #ifdef CONFIG_TRANSPARENT_HUGEPAGE @@ -570,7 +581,8 @@ static void smaps_pmd_entry(pmd_t *pmd, unsigned long addr, /* pass */; else VM_BUG_ON_PAGE(1, page); - smaps_account(mss, page, true, pmd_young(*pmd), pmd_dirty(*pmd), locked); + smaps_account(mss, page, true, pmd_young(*pmd), pmd_dirty(*pmd), + locked, false); } #else static void smaps_pmd_entry(pmd_t *pmd, unsigned long addr, @@ -1285,6 +1297,7 @@ static pagemap_entry_t pte_to_pagemap_entry(struct pagemapread *pm, { u64 frame = 0, flags = 0; struct page *page = NULL; + bool migration = false; if (pte_present(pte)) { if (pm->show_pfn) @@ -1302,8 +1315,10 @@ static pagemap_entry_t pte_to_pagemap_entry(struct pagemapread *pm, frame = swp_type(entry) | (swp_offset(entry) << MAX_SWAPFILES_SHIFT); flags |= PM_SWAP; - if (is_migration_entry(entry)) + if (is_migration_entry(entry)) { + migration = true; page = migration_entry_to_page(entry); + } if (is_device_private_entry(entry)) page = device_private_entry_to_page(entry); @@ -1311,7 +1326,7 @@ static pagemap_entry_t pte_to_pagemap_entry(struct pagemapread *pm, if (page && !PageAnon(page)) flags |= PM_FILE; - if (page && page_mapcount(page) == 1) + if (page && !migration && page_mapcount(page) == 1) flags |= PM_MMAP_EXCLUSIVE; if (vma->vm_flags & VM_SOFTDIRTY) flags |= PM_SOFT_DIRTY; @@ -1327,8 +1342,9 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end, spinlock_t *ptl; pte_t *pte, *orig_pte; int err = 0; - #ifdef CONFIG_TRANSPARENT_HUGEPAGE + bool migration = false; + ptl = pmd_trans_huge_lock(pmdp, vma); if (ptl) { u64 flags = 0, frame = 0; @@ -1363,11 +1379,12 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end, if (pmd_swp_soft_dirty(pmd)) flags |= PM_SOFT_DIRTY; VM_BUG_ON(!is_pmd_migration_entry(pmd)); + migration = is_migration_entry(entry); page = migration_entry_to_page(entry); } #endif - if (page && page_mapcount(page) == 1) + if (page && !migration && page_mapcount(page) == 1) flags |= PM_MMAP_EXCLUSIVE; for (; addr != end; addr += PAGE_SIZE) { -- 2.25.1

2 9

[PATCH OLK-5.10 0/8] SAS-related bugfix
by Yihang Li 16 Mar '23

16 Mar '23

From: xiabing <xiabing12(a)h-partners.com> 1. In the NCQ scenario, if multiple I/Os are delivered and one of the I/Os is faulty, a group of slow disks will always occur. 2. During I/O running, a null pointer exception occurs during the pressure test when a disk is removed. 3. When the length of the DMA Setup frame returned by the disk is abnormal, a group of slow disks are triggered. John Garry (2): scsi: libsas: Add sas_ata_device_link_abort() scsi: libsas: Update SATA dev FIS in sas_ata_task_done() Xingui Yang (6): scsi: hisi_sas: Move slot variable definition in hisi_sas_abort_task() scsi: hisi_sas: Add SATA_DISK_ERR bit handling for v3 hw scsi: hisi_sas: Use abort task set to reset SAS disks when discovered scsi: libsas: Grab the ATA port lock in sas_ata_device_link_abort() {topost} scsi: hisi_sas: Handle NCQ error when IPTT is valid {topost} scsi: hisi_sas: Grab sas_dev lock when traversing the members of sas_dev.list drivers/scsi/hisi_sas/hisi_sas.h | 4 +- drivers/scsi/hisi_sas/hisi_sas_main.c | 52 ++++++++++++++++------- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 8 +++- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 8 +++- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 58 ++++++++++++++++++++++++-- drivers/scsi/libsas/sas_ata.c | 22 +++++++++- include/scsi/sas_ata.h | 6 +++ 7 files changed, 132 insertions(+), 26 deletions(-) -- 2.30.0

1 8

订阅
by 乄語 16 Mar '23

16 Mar '23

subscribe

1 0

[PATCH OLK-5.10 0/8] SAS-related bugfix
by Yihang Li 16 Mar '23

16 Mar '23

From: xiabing <xiabing12(a)h-partners.com> 1. In the NCQ scenario, if multiple I/Os are delivered and one of the I/Os is faulty, a group of slow disks will always occur. 2. During I/O running, a null pointer exception occurs during the pressure test when a disk is removed. 3. When the length of the DMA Setup frame returned by the disk is abnormal, a group of slow disks are triggered. John Garry (2): scsi: libsas: Add sas_ata_device_link_abort() scsi: libsas: Update SATA dev FIS in sas_ata_task_done() Xingui Yang (6): scsi: hisi_sas: Move slot variable definition in hisi_sas_abort_task() scsi: hisi_sas: Add SATA_DISK_ERR bit handling for v3 hw scsi: hisi_sas: Use abort task set to reset SAS disks when discovered scsi: libsas: Grab the ATA port lock in sas_ata_device_link_abort() {topost} scsi: hisi_sas: Handle NCQ error when IPTT is valid {topost} scsi: hisi_sas: Grab sas_dev lock when traversing the members of sas_dev.list drivers/scsi/hisi_sas/hisi_sas.h | 4 +- drivers/scsi/hisi_sas/hisi_sas_main.c | 52 ++++++++++++++++------- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 8 +++- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 8 +++- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 58 ++++++++++++++++++++++++-- drivers/scsi/libsas/sas_ata.c | 22 +++++++++- include/scsi/sas_ata.h | 6 +++ 7 files changed, 132 insertions(+), 26 deletions(-) -- 2.30.0

1 8

[PATCH PR openEuler-22.03-LTS-SP1] mm/vmalloc: huge vmalloc backing pages should be split rather than compound
by Jialin Zhang 16 Mar '23

16 Mar '23

From: Nicholas Piggin <npiggin(a)gmail.com> mainline inclusion from mainline-v5.18-rc4 commit 3b8000ae185cb068adbda5f966a3835053c85fd4 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6LD0S CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Huge vmalloc higher-order backing pages were allocated with __GFP_COMP in order to allow the sub-pages to be refcounted by callers such as "remap_vmalloc_page [sic]" (remap_vmalloc_range). However a similar problem exists for other struct page fields callers use, for example fb_deferred_io_fault() takes a vmalloc'ed page and not only refcounts it but uses ->lru, ->mapping, ->index. This is not compatible with compound sub-pages, and can cause bad page state issues like BUG: Bad page state in process swapper/0 pfn:00743 page:(____ptrval____) refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x743 flags: 0x7ffff000000000(node=0|zone=0|lastcpupid=0x7ffff) raw: 007ffff000000000 c00c00000001d0c8 c00c00000001d0c8 0000000000000000 raw: 0000000000000000 0000000000000000 00000000ffffffff 0000000000000000 page dumped because: corrupted mapping in tail page Modules linked in: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.18.0-rc3-00082-gfc6fff4a7ce1-dirty #2810 Call Trace: dump_stack_lvl+0x74/0xa8 (unreliable) bad_page+0x12c/0x170 free_tail_pages_check+0xe8/0x190 free_pcp_prepare+0x31c/0x4e0 free_unref_page+0x40/0x1b0 __vunmap+0x1d8/0x420 ... The correct approach is to use split high-order pages for the huge vmalloc backing. These allow callers to treat them in exactly the same way as individually-allocated order-0 pages. Link: https://lore.kernel.org/all/14444103-d51b-0fb3-ee63-c3f182f0b546@molgen.mpg… Signed-off-by: Nicholas Piggin <npiggin(a)gmail.com> Cc: Paul Menzel <pmenzel(a)molgen.mpg.de> Cc: Song Liu <songliubraving(a)fb.com> Cc: Rick Edgecombe <rick.p.edgecombe(a)intel.com> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> conflicts: mm/vmalloc.c Signed-off-by: ZhangPeng <zhangpeng362(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- mm/vmalloc.c | 22 +++++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/mm/vmalloc.c b/mm/vmalloc.c index e27cd716ca95..2ca2c1bc0db9 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -2641,14 +2641,17 @@ static void __vunmap(const void *addr, int deallocate_pages) vm_remove_mappings(area, deallocate_pages); if (deallocate_pages) { - unsigned int page_order = vm_area_page_order(area); int i; - for (i = 0; i < area->nr_pages; i += 1U << page_order) { + for (i = 0; i < area->nr_pages; i++) { struct page *page = area->pages[i]; BUG_ON(!page); - __free_pages(page, page_order); + /* + * High-order allocs for huge vmallocs are split, so + * can be freed as an array of order-0 allocations + */ + __free_pages(page, 0); } atomic_long_sub(area->nr_pages, &nr_vmalloc_pages); @@ -2930,8 +2933,7 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask, struct page *page; int p; - /* Compound pages required for remap_vmalloc_page */ - page = alloc_pages_node(node, gfp_mask | __GFP_COMP, page_order); + page = alloc_pages_node(node, gfp_mask, page_order); if (unlikely(!page)) { /* Successfully allocated i pages, free them in __vfree() */ area->nr_pages = i; @@ -2943,6 +2945,16 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask, goto fail; } + /* + * Higher order allocations must be able to be treated as + * indepdenent small pages by callers (as they can with + * small-page vmallocs). Some drivers do their own refcounting + * on vmalloc_to_page() pages, some use page->mapping, + * page->lru, etc. + */ + if (page_order) + split_page(page, page_order); + for (p = 0; p < (1U << page_order); p++) area->pages[i + p] = page + p; -- 2.25.1

1 0

[PATCH PR OLK-5.10] mm/vmalloc: huge vmalloc backing pages should be split rather than compound
by Jialin Zhang 16 Mar '23

16 Mar '23

From: Nicholas Piggin <npiggin(a)gmail.com> mainline inclusion from mainline-v5.18-rc4 commit 3b8000ae185cb068adbda5f966a3835053c85fd4 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6LD0S CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Huge vmalloc higher-order backing pages were allocated with __GFP_COMP in order to allow the sub-pages to be refcounted by callers such as "remap_vmalloc_page [sic]" (remap_vmalloc_range). However a similar problem exists for other struct page fields callers use, for example fb_deferred_io_fault() takes a vmalloc'ed page and not only refcounts it but uses ->lru, ->mapping, ->index. This is not compatible with compound sub-pages, and can cause bad page state issues like BUG: Bad page state in process swapper/0 pfn:00743 page:(____ptrval____) refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x743 flags: 0x7ffff000000000(node=0|zone=0|lastcpupid=0x7ffff) raw: 007ffff000000000 c00c00000001d0c8 c00c00000001d0c8 0000000000000000 raw: 0000000000000000 0000000000000000 00000000ffffffff 0000000000000000 page dumped because: corrupted mapping in tail page Modules linked in: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.18.0-rc3-00082-gfc6fff4a7ce1-dirty #2810 Call Trace: dump_stack_lvl+0x74/0xa8 (unreliable) bad_page+0x12c/0x170 free_tail_pages_check+0xe8/0x190 free_pcp_prepare+0x31c/0x4e0 free_unref_page+0x40/0x1b0 __vunmap+0x1d8/0x420 ... The correct approach is to use split high-order pages for the huge vmalloc backing. These allow callers to treat them in exactly the same way as individually-allocated order-0 pages. Link: https://lore.kernel.org/all/14444103-d51b-0fb3-ee63-c3f182f0b546@molgen.mpg… Signed-off-by: Nicholas Piggin <npiggin(a)gmail.com> Cc: Paul Menzel <pmenzel(a)molgen.mpg.de> Cc: Song Liu <songliubraving(a)fb.com> Cc: Rick Edgecombe <rick.p.edgecombe(a)intel.com> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> conflicts: mm/vmalloc.c Signed-off-by: ZhangPeng <zhangpeng362(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- mm/vmalloc.c | 22 +++++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/mm/vmalloc.c b/mm/vmalloc.c index e27cd716ca95..2ca2c1bc0db9 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -2641,14 +2641,17 @@ static void __vunmap(const void *addr, int deallocate_pages) vm_remove_mappings(area, deallocate_pages); if (deallocate_pages) { - unsigned int page_order = vm_area_page_order(area); int i; - for (i = 0; i < area->nr_pages; i += 1U << page_order) { + for (i = 0; i < area->nr_pages; i++) { struct page *page = area->pages[i]; BUG_ON(!page); - __free_pages(page, page_order); + /* + * High-order allocs for huge vmallocs are split, so + * can be freed as an array of order-0 allocations + */ + __free_pages(page, 0); } atomic_long_sub(area->nr_pages, &nr_vmalloc_pages); @@ -2930,8 +2933,7 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask, struct page *page; int p; - /* Compound pages required for remap_vmalloc_page */ - page = alloc_pages_node(node, gfp_mask | __GFP_COMP, page_order); + page = alloc_pages_node(node, gfp_mask, page_order); if (unlikely(!page)) { /* Successfully allocated i pages, free them in __vfree() */ area->nr_pages = i; @@ -2943,6 +2945,16 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask, goto fail; } + /* + * Higher order allocations must be able to be treated as + * indepdenent small pages by callers (as they can with + * small-page vmallocs). Some drivers do their own refcounting + * on vmalloc_to_page() pages, some use page->mapping, + * page->lru, etc. + */ + if (page_order) + split_page(page, page_order); + for (p = 0; p < (1U << page_order); p++) area->pages[i + p] = page + p; -- 2.25.1

1 0

[PATCH openEuler-5.10-LTS 01/13] ext4: fix incorrect options show of original mount_opt and extend mount_opt2
by Jialin Zhang 15 Mar '23

15 Mar '23

From: Zhang Yi <yi.zhang(a)huawei.com> maillist inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6D5XF Reference: https://lore.kernel.org/linux-ext4/20230130111138.76tp6pij3yhh4brh@quack3/T… -------------------------------- Current _ext4_show_options() do not distinguish MOPT_2 flag, so it mixed extend sbi->s_mount_opt2 options with sbi->s_mount_opt, it could lead to show incorrect options, e.g. show fc_debug_force if we mount with errors=continue mode and miss it if we set. $ mkfs.ext4 /dev/pmem0 $ mount -o errors=remount-ro /dev/pmem0 /mnt $ cat /proc/fs/ext4/pmem0/options | grep fc_debug_force #empty $ mount -o remount,errors=continue /mnt $ cat /proc/fs/ext4/pmem0/options | grep fc_debug_force fc_debug_force $ mount -o remount,errors=remount-ro,fc_debug_force /mnt $ cat /proc/fs/ext4/pmem0/options | grep fc_debug_force #empty Fixes: 995a3ed67fc8 ("ext4: add fast_commit feature and handling for extended mount options") Signed-off-by: Zhang Yi <yi.zhang(a)huawei.com> Conflict: fs/ext4/super.c Reviewed-by: Zhihao Cheng <chengzhihao1(a)huawei.com> Reviewed-by: Zhang Xiaoxu <zhangxiaoxu5(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- fs/ext4/ext4.h | 1 + fs/ext4/super.c | 28 +++++++++++++++++++++------- 2 files changed, 22 insertions(+), 7 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index c470a5fb2f20..7e5abaa31fea 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1459,6 +1459,7 @@ struct ext4_sb_info { unsigned int s_mount_opt2; unsigned long s_mount_flags; unsigned int s_def_mount_opt; + unsigned int s_def_mount_opt2; ext4_fsblk_t s_sb_block; atomic64_t s_resv_clusters; kuid_t s_resuid; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index dd3b72ba67e8..da8bd8031119 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -2585,7 +2585,7 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb, { struct ext4_sb_info *sbi = EXT4_SB(sb); struct ext4_super_block *es = sbi->s_es; - int def_errors, def_mount_opt = sbi->s_def_mount_opt; + int def_errors; const struct mount_opts *m; char sep = nodefs ? '\n' : ','; @@ -2597,15 +2597,28 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb, for (m = ext4_mount_opts; m->token != Opt_err; m++) { int want_set = m->flags & MOPT_SET; + int opt_2 = m->flags & MOPT_2; + unsigned int mount_opt, def_mount_opt; + if (((m->flags & (MOPT_SET|MOPT_CLEAR)) == 0) || (m->flags & MOPT_CLEAR_ERR) || m->flags & MOPT_SKIP) continue; - if (!nodefs && !(m->mount_opt & (sbi->s_mount_opt ^ def_mount_opt))) - continue; /* skip if same as the default */ + + if (opt_2) { + mount_opt = sbi->s_mount_opt2; + def_mount_opt = sbi->s_def_mount_opt2; + } else { + mount_opt = sbi->s_mount_opt; + def_mount_opt = sbi->s_def_mount_opt; + } + /* skip if same as the default */ + if (!nodefs && !(m->mount_opt & (mount_opt ^ def_mount_opt))) + continue; + /* select Opt_noFoo vs Opt_Foo */ if ((want_set && - (sbi->s_mount_opt & m->mount_opt) != m->mount_opt) || - (!want_set && (sbi->s_mount_opt & m->mount_opt))) - continue; /* select Opt_noFoo vs Opt_Foo */ + (mount_opt & m->mount_opt) != m->mount_opt) || + (!want_set && (mount_opt & m->mount_opt))) + continue; SEQ_OPTS_PRINT("%s", token2str(m->token)); } @@ -2635,7 +2648,7 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb, if (nodefs || sbi->s_stripe) SEQ_OPTS_PRINT("stripe=%lu", sbi->s_stripe); if (nodefs || EXT4_MOUNT_DATA_FLAGS & - (sbi->s_mount_opt ^ def_mount_opt)) { + (sbi->s_mount_opt ^ sbi->s_def_mount_opt)) { if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) SEQ_OPTS_PUTS("data=journal"); else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA) @@ -4340,6 +4353,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) kfree(s_mount_opts); } sbi->s_def_mount_opt = sbi->s_mount_opt; + sbi->s_def_mount_opt2 = sbi->s_mount_opt2; if (!parse_options((char *) data, sb, &journal_devnum, &journal_ioprio, 0)) goto failed_mount; -- 2.25.1

1 12

[PATCH openEuler-5.10-LTS-SP1 01/13] ext4: fix incorrect options show of original mount_opt and extend mount_opt2
by Jialin Zhang 15 Mar '23

15 Mar '23

From: Zhang Yi <yi.zhang(a)huawei.com> maillist inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6D5XF Reference: https://lore.kernel.org/linux-ext4/20230130111138.76tp6pij3yhh4brh@quack3/T… -------------------------------- Current _ext4_show_options() do not distinguish MOPT_2 flag, so it mixed extend sbi->s_mount_opt2 options with sbi->s_mount_opt, it could lead to show incorrect options, e.g. show fc_debug_force if we mount with errors=continue mode and miss it if we set. $ mkfs.ext4 /dev/pmem0 $ mount -o errors=remount-ro /dev/pmem0 /mnt $ cat /proc/fs/ext4/pmem0/options | grep fc_debug_force #empty $ mount -o remount,errors=continue /mnt $ cat /proc/fs/ext4/pmem0/options | grep fc_debug_force fc_debug_force $ mount -o remount,errors=remount-ro,fc_debug_force /mnt $ cat /proc/fs/ext4/pmem0/options | grep fc_debug_force #empty Fixes: 995a3ed67fc8 ("ext4: add fast_commit feature and handling for extended mount options") Signed-off-by: Zhang Yi <yi.zhang(a)huawei.com> Conflict: fs/ext4/super.c Reviewed-by: Zhihao Cheng <chengzhihao1(a)huawei.com> Reviewed-by: Zhang Xiaoxu <zhangxiaoxu5(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- fs/ext4/ext4.h | 1 + fs/ext4/super.c | 28 +++++++++++++++++++++------- 2 files changed, 22 insertions(+), 7 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 8245b94d8fc6..f1d36671bc2b 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1462,6 +1462,7 @@ struct ext4_sb_info { unsigned int s_mount_opt2; unsigned long s_mount_flags; unsigned int s_def_mount_opt; + unsigned int s_def_mount_opt2; ext4_fsblk_t s_sb_block; atomic64_t s_resv_clusters; kuid_t s_resuid; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index c94ea845ea57..4fd680507948 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -2586,7 +2586,7 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb, { struct ext4_sb_info *sbi = EXT4_SB(sb); struct ext4_super_block *es = sbi->s_es; - int def_errors, def_mount_opt = sbi->s_def_mount_opt; + int def_errors; const struct mount_opts *m; char sep = nodefs ? '\n' : ','; @@ -2598,15 +2598,28 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb, for (m = ext4_mount_opts; m->token != Opt_err; m++) { int want_set = m->flags & MOPT_SET; + int opt_2 = m->flags & MOPT_2; + unsigned int mount_opt, def_mount_opt; + if (((m->flags & (MOPT_SET|MOPT_CLEAR)) == 0) || (m->flags & MOPT_CLEAR_ERR) || m->flags & MOPT_SKIP) continue; - if (!nodefs && !(m->mount_opt & (sbi->s_mount_opt ^ def_mount_opt))) - continue; /* skip if same as the default */ + + if (opt_2) { + mount_opt = sbi->s_mount_opt2; + def_mount_opt = sbi->s_def_mount_opt2; + } else { + mount_opt = sbi->s_mount_opt; + def_mount_opt = sbi->s_def_mount_opt; + } + /* skip if same as the default */ + if (!nodefs && !(m->mount_opt & (mount_opt ^ def_mount_opt))) + continue; + /* select Opt_noFoo vs Opt_Foo */ if ((want_set && - (sbi->s_mount_opt & m->mount_opt) != m->mount_opt) || - (!want_set && (sbi->s_mount_opt & m->mount_opt))) - continue; /* select Opt_noFoo vs Opt_Foo */ + (mount_opt & m->mount_opt) != m->mount_opt) || + (!want_set && (mount_opt & m->mount_opt))) + continue; SEQ_OPTS_PRINT("%s", token2str(m->token)); } @@ -2636,7 +2649,7 @@ static int _ext4_show_options(struct seq_file *seq, struct super_block *sb, if (nodefs || sbi->s_stripe) SEQ_OPTS_PRINT("stripe=%lu", sbi->s_stripe); if (nodefs || EXT4_MOUNT_DATA_FLAGS & - (sbi->s_mount_opt ^ def_mount_opt)) { + (sbi->s_mount_opt ^ sbi->s_def_mount_opt)) { if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA) SEQ_OPTS_PUTS("data=journal"); else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA) @@ -4341,6 +4354,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) kfree(s_mount_opts); } sbi->s_def_mount_opt = sbi->s_mount_opt; + sbi->s_def_mount_opt2 = sbi->s_mount_opt2; if (!parse_options((char *) data, sb, &journal_devnum, &journal_ioprio, 0)) goto failed_mount; -- 2.25.1

1 12

[RFC PATCH openEuler-1.0-LTS v2] sched: optimise the way to get util
by Wang ShaoBo 15 Mar '23

15 Mar '23

From: tanghui <tanghui20(a)huawei.com> optimise the way to get util. Signed-off-by: tanghui <tanghui20(a)huawei.com> Signed-off-by: Wang ShaoBo <bobo.shaobowang(a)huawei.com> v2: refresh low_pct change way through sysctl. --- fs/proc/stat.c | 4 ++ include/linux/sched/cputime.h | 3 ++ include/linux/sched/sysctl.h | 2 + kernel/sched/fair.c | 83 +++++++++++++++++++++++++++++++++++ kernel/sched/sched.h | 1 + kernel/sysctl.c | 9 ++++ 6 files changed, 102 insertions(+) diff --git a/fs/proc/stat.c b/fs/proc/stat.c index 7e832b24847dd..3fe60a77b0b4d 100644 --- a/fs/proc/stat.c +++ b/fs/proc/stat.c @@ -63,7 +63,11 @@ u64 get_idle_time(int cpu) return idle; } +#ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY +u64 get_iowait_time(int cpu) +#else static u64 get_iowait_time(int cpu) +#endif { u64 iowait, iowait_usecs = -1ULL; diff --git a/include/linux/sched/cputime.h b/include/linux/sched/cputime.h index 6b1793606fc95..4a092e006f5b2 100644 --- a/include/linux/sched/cputime.h +++ b/include/linux/sched/cputime.h @@ -189,6 +189,9 @@ task_sched_runtime(struct task_struct *task); extern int use_sched_idle_time; extern int sched_idle_time_adjust(int cpu, u64 *utime, u64 *stime); extern unsigned long long sched_get_idle_time(int cpu); +#ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY +extern u64 get_iowait_time(int cpu); +#endif #ifdef CONFIG_PROC_FS extern u64 get_idle_time(int cpu); diff --git a/include/linux/sched/sysctl.h b/include/linux/sched/sysctl.h index 04eb5b127867b..8223a1fce176c 100644 --- a/include/linux/sched/sysctl.h +++ b/include/linux/sched/sysctl.h @@ -31,6 +31,8 @@ extern unsigned int sysctl_sched_latency; extern unsigned int sysctl_sched_min_granularity; extern unsigned int sysctl_sched_wakeup_granularity; extern unsigned int sysctl_sched_child_runs_first; +extern int sysctl_sched_util_update_interval; +extern unsigned long sysctl_sched_util_update_interval_max; #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY extern int sysctl_sched_util_low_pct; diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index ad6a7923c9edb..af55a26d11fcb 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -6682,6 +6682,73 @@ static int wake_cap(struct task_struct *p, int cpu, int prev_cpu) */ int sysctl_sched_util_low_pct = 85; +struct cpu_timeinfo { + u64 systime; + u64 idletime; + unsigned long next_update; + int vutil; +}; + +/* + * The time interval to update CPU utilization + * (default 1ms, max 10min) + */ +int sysctl_sched_util_update_interval = 1; +unsigned long sysctl_sched_util_update_interval_max = 600000; + +static DEFINE_PER_CPU(struct cpu_timeinfo, qos_cputime); + +static inline u64 cpu_systime(int cpu) +{ + u64 user, nice, system, idle, iowait, irq, softirq, steal; + + user = kcpustat_cpu(cpu).cpustat[CPUTIME_USER]; + system = kcpustat_cpu(cpu).cpustat[CPUTIME_SYSTEM]; + iowait = get_iowait_time(cpu); + irq = kcpustat_cpu(cpu).cpustat[CPUTIME_IRQ]; + softirq = kcpustat_cpu(cpu).cpustat[CPUTIME_SOFTIRQ]; + nice = kcpustat_cpu(cpu).cpustat[CPUTIME_NICE]; + steal = kcpustat_cpu(cpu).cpustat[CPUTIME_STEAL]; + idle = get_idle_time(cpu); + + return user + system + iowait + irq + softirq + nice + idle + steal; +} + +static inline u64 cpu_idletime(int cpu) +{ + return get_idle_time(cpu) + get_iowait_time(cpu); +} + +static inline void update_cpu_vutil(void) +{ + struct cpu_timeinfo *cputime = per_cpu_ptr(&qos_cputime, smp_processor_id()); + u64 delta_systime, delta_idle, systime, idletime; + int cpu = smp_processor_id(); + unsigned long interval; + + if (time_after(jiffies, cputime->next_update)) { + interval = msecs_to_jiffies(sysctl_sched_util_update_interval); + cputime->next_update = jiffies + interval; + systime = cpu_systime(cpu); + idletime = cpu_idletime(cpu); + delta_systime = systime - cputime->systime; + delta_idle = idletime - cputime->idletime; + if (!delta_systime) + return; + + cputime->systime = systime; + cputime->idletime = idletime; + cputime->vutil = (delta_systime - delta_idle) * 100 / delta_systime; + } +} + +static inline int cpu_vutil_of(int cpu) +{ + struct cpu_timeinfo *cputime = per_cpu_ptr(&qos_cputime, cpu); + + return cputime->vutil; +} + static inline bool prefer_cpus_valid(struct task_struct *p) { return p->prefer_cpus && @@ -6741,17 +6808,29 @@ static void set_task_select_cpus(struct task_struct *p, int *idlest_cpu, return; } +#if 0 util_avg_sum += tg->se[cpu]->avg.util_avg; tg_capacity += capacity_of(cpu); +#endif + util_avg_sum += cpu_vutil_of(cpu); } rcu_read_unlock(); +#if 0 if (tg_capacity > cpumask_weight(p->prefer_cpus) && util_avg_sum * 100 <= tg_capacity * sysctl_sched_util_low_pct) { p->select_cpus = p->prefer_cpus; if (sd_flag & SD_BALANCE_WAKE) schedstat_inc(p->se.dyn_affi_stats->nr_wakeups_preferred_cpus); } +#endif + + if (util_avg_sum < sysctl_sched_util_low_pct * + cpumask_weight(p->prefer_cpus)) { + p->select_cpus = p->prefer_cpus; + if (sd_flag & SD_BALANCE_WAKE) + schedstat_inc(p->se.dyn_affi_stats->nr_wakeups_preferred_cpus); + } } #endif @@ -10610,6 +10689,10 @@ static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued) if (static_branch_unlikely(&sched_numa_balancing)) task_tick_numa(rq, curr); + +#ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY + update_cpu_vutil(); +#endif } /* diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ae30681530938..045fbb3871bbe 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -2354,3 +2354,4 @@ static inline void membarrier_switch_mm(struct rq *rq, { } #endif + diff --git a/kernel/sysctl.c b/kernel/sysctl.c index ad62ea156afd9..685f9881b8e23 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -352,6 +352,15 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = proc_dointvec, }, + { + .procname = "sched_util_update_interval_ms", + .data = &sysctl_sched_util_update_interval, + .maxlen = sizeof(sysctl_sched_util_update_interval), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = &one, + .extra2 = &sysctl_sched_util_update_interval_max, + }, #ifdef CONFIG_SCHED_DEBUG { .procname = "sched_min_granularity_ns", -- 2.25.1

1 0

[PATCH] sched: optimise the way to get util
by Wang ShaoBo 15 Mar '23

15 Mar '23

From: tanghui <tanghui20(a)huawei.com> sched: optimise the way to get util, which only canbe used when CONFIG_QOS_SCHED_DYNAMIC_AFFINITY is set. Signed-off-by: tanghui <tanghui20(a)huawei.com> --- fs/proc/stat.c | 4 ++ include/linux/sched/cputime.h | 3 ++ include/linux/sched/sysctl.h | 2 + kernel/sched/fair.c | 97 ++++++++++++++++++++++++++++++----- kernel/sched/sched.h | 1 + kernel/sysctl.c | 9 ++++ 6 files changed, 104 insertions(+), 12 deletions(-) diff --git a/fs/proc/stat.c b/fs/proc/stat.c index c83a10e895f4..2eaba2b78f47 100644 --- a/fs/proc/stat.c +++ b/fs/proc/stat.c @@ -63,7 +63,11 @@ u64 get_idle_time(int cpu) return idle; } +#ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY +u64 get_iowait_time(int cpu) +#else static u64 get_iowait_time(int cpu) +#endif { u64 iowait, iowait_usecs = -1ULL; diff --git a/include/linux/sched/cputime.h b/include/linux/sched/cputime.h index 1ebbeec02051..f6244d48f357 100644 --- a/include/linux/sched/cputime.h +++ b/include/linux/sched/cputime.h @@ -190,5 +190,8 @@ extern int use_sched_idle_time; extern int sched_idle_time_adjust(int cpu, u64 *utime, u64 *stime); extern unsigned long long sched_get_idle_time(int cpu); extern u64 get_idle_time(int cpu); +#ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY +extern u64 get_iowait_time(int cpu); +#endif #endif /* _LINUX_SCHED_CPUTIME_H */ diff --git a/include/linux/sched/sysctl.h b/include/linux/sched/sysctl.h index f5031a607df8..386ef53017ca 100644 --- a/include/linux/sched/sysctl.h +++ b/include/linux/sched/sysctl.h @@ -28,6 +28,8 @@ extern unsigned int sysctl_sched_child_runs_first; extern int sysctl_sched_util_low_pct; extern int sysctl_sched_util_higher_pct; extern int sysctl_sched_load_higher_pct; +extern int sysctl_sched_util_update_interval; +extern unsigned long sysctl_sched_util_update_interval_max; #endif enum sched_tunable_scaling { diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 5344a68a463e..3217e3998fdf 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -6458,6 +6458,15 @@ static int wake_cap(struct task_struct *p, int cpu, int prev_cpu) } #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY +#define UTIL_PCT_HIGH 85 + +struct cpu_timeinfo { + u64 systime; + u64 idletime; + unsigned long next_update; + int vutil; +}; + /* * Light load threshold for CPU: just use cpu Utilization to measure * @@ -6484,15 +6493,75 @@ int sysctl_sched_util_higher_pct = 100; */ int sysctl_sched_load_higher_pct = 10; +/* + * The time interval to update CPU utilization + * (default 1ms, max 10min) + */ +int sysctl_sched_util_update_interval = 1; +unsigned long sysctl_sched_util_update_interval_max = 600000; + +static DEFINE_PER_CPU(struct cpu_timeinfo, qos_cputime); + +static inline u64 cpu_systime(int cpu) +{ + u64 user, nice, system, idle, iowait, irq, softirq, steal; + + user = kcpustat_cpu(cpu).cpustat[CPUTIME_USER]; + system = kcpustat_cpu(cpu).cpustat[CPUTIME_SYSTEM]; + iowait = get_iowait_time(cpu); + irq = kcpustat_cpu(cpu).cpustat[CPUTIME_IRQ]; + softirq = kcpustat_cpu(cpu).cpustat[CPUTIME_SOFTIRQ]; + nice = kcpustat_cpu(cpu).cpustat[CPUTIME_NICE]; + steal = kcpustat_cpu(cpu).cpustat[CPUTIME_STEAL]; + idle = get_idle_time(cpu); + + return user + system + iowait + irq + softirq + nice + idle + steal; +} + +static inline u64 cpu_idletime(int cpu) +{ + return get_idle_time(cpu) + get_iowait_time(cpu); +} + +static inline void update_cpu_vutil(void) +{ + struct cpu_timeinfo *cputime = per_cpu_ptr(&qos_cputime, smp_processor_id()); + u64 delta_systime, delta_idle, systime, idletime; + int cpu = smp_processor_id(); + unsigned long interval; + + if (time_after(jiffies, cputime->next_update)) { + interval = msecs_to_jiffies(sysctl_sched_util_update_interval); + cputime->next_update = jiffies + interval; + systime = cpu_systime(cpu); + idletime = cpu_idletime(cpu); + delta_systime = systime - cputime->systime; + delta_idle = idletime - cputime->idletime; + if (!delta_systime) + return; + + cputime->systime = systime; + cputime->idletime = idletime; + cputime->vutil = (delta_systime - delta_idle) * 100 / delta_systime; + } +} + +static inline int cpu_vutil_of(int cpu) +{ + struct cpu_timeinfo *cputime = per_cpu_ptr(&qos_cputime, cpu); + + return cputime->vutil; +} + static inline bool prefer_cpu_util_low(int cpu) { unsigned long capacity = capacity_of(cpu); - unsigned long util = cpu_util(cpu); + unsigned long util_pct = cpu_vutil_of(cpu); - if (util >= capacity || capacity <= 1) + if (util_pct >= 100 || capacity <= 1) return sysctl_sched_util_low_pct == 100; - return util * 100 <= capacity * sysctl_sched_util_low_pct; + return util_pct <= sysctl_sched_util_low_pct; } /* @@ -6504,8 +6573,8 @@ static inline int compare_cpu_util(int preferred_cpu, int external_cpu) { unsigned long capacity_cpux = capacity_of(preferred_cpu); unsigned long capacity_cpuy = capacity_of(external_cpu); - unsigned long cpu_util_x = cpu_util(preferred_cpu); - unsigned long cpu_util_y = cpu_util(external_cpu); + unsigned long cpu_util_x = cpu_vutil_of(preferred_cpu); + unsigned long cpu_util_y = cpu_vutil_of(external_cpu); int ratio; /* @@ -6519,7 +6588,7 @@ static inline int compare_cpu_util(int preferred_cpu, int external_cpu) if (capacity_cpux <= 1) return 1; - if (cpu_util_x >= capacity_cpux && available_idle_cpu(external_cpu)) + if (cpu_util_x >= UTIL_PCT_HIGH && available_idle_cpu(external_cpu)) return 1; if (!cpu_util_x) @@ -6529,12 +6598,12 @@ static inline int compare_cpu_util(int preferred_cpu, int external_cpu) * The lower the CPU utilization, the larger the ratio of * CPU utilization gap. */ - ratio = cpu_util_x >= capacity_cpux ? 1 : capacity_cpux / cpu_util_x; + ratio = cpu_util_x >= 100 ? 0 : 100 / cpu_util_x; if (ratio > 10) ratio = 10; - return (sysctl_sched_util_higher_pct * ratio + 100) * cpu_util_y * - capacity_cpux < 100 * cpu_util_x * capacity_cpuy; + return (sysctl_sched_util_higher_pct * ratio + 100) * + cpu_util_y * capacity_cpux < 100 * cpu_util_x * capacity_cpuy; } static inline bool prefer_cpus_valid(struct task_struct *p) @@ -6549,7 +6618,7 @@ static inline bool prefer_cpus_valid(struct task_struct *p) * @p: the task whose available cpu range will to set * @idlest_cpu: the cpu which is the idlest in prefer cpus * - * x: the cpu of min util_avg from preferred set + * x: the cpu of min util from preferred set * y: the cpu from allowed set but exclude preferred set * * If x's utilization is low, select preferred cpu range for task @@ -6578,7 +6647,7 @@ static void set_task_select_cpus(struct task_struct *p, int *idlest_cpu, goto cpus_allowed; cpu = cpumask_first(p->se.prefer_cpus); - min_util = cpu_util(cpu); + min_util = cpu_vutil_of(cpu); for_each_cpu(i, p->se.prefer_cpus) { if (prefer_cpu_util_low(i) || available_idle_cpu(i)) { cpu = i; @@ -6589,7 +6658,7 @@ static void set_task_select_cpus(struct task_struct *p, int *idlest_cpu, if (capacity_of(i) <= 1) continue; - c_util = cpu_util(i); + c_util = cpu_vutil_of(i); if (min_util * capacity_of(i) > c_util * capacity_of(cpu)) { min_util = c_util; cpu = i; @@ -10278,6 +10347,10 @@ static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued) if (static_branch_unlikely(&sched_numa_balancing)) task_tick_numa(rq, curr); + +#ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY + update_cpu_vutil(); +#endif } /* diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 55216a03d327..4de9c966da01 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -2348,3 +2348,4 @@ static inline void membarrier_switch_mm(struct rq *rq, { } #endif + diff --git a/kernel/sysctl.c b/kernel/sysctl.c index b309aa206697..d52e003f3f56 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1274,6 +1274,15 @@ static struct ctl_table kern_table[] = { .extra1 = &zero, .extra2 = &one_hundred, }, + { + .procname = "sched_util_update_interval_ms", + .data = &sysctl_sched_util_update_interval, + .maxlen = sizeof(sysctl_sched_util_update_interval), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = &one, + .extra2 = &sysctl_sched_util_update_interval_max, + }, #endif #endif { } -- 2.23.0

1 0