mm,memory_hotplug: Fix scan_movable_pages for gigantic hugepages
Subsystems (2)
Oscar Salvador
maintainerPATCH 1/1 mm,memory_hotplug: Fix scan_movable_pages for gigantic hugepages
This is the same sort of error we saw in [1].
Gigantic hugepages crosses several memblocks, so it can be that the page we get in scan_movable_pages() is a page-tail belonging to a 1G-hugepage.
If that happens, page_hstate()->size_to_hstate() will return NULL, and we will blow up in hugepage_migration_supported().
The splat is as follows:
kernel: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 kernel: #PF error: [normal kernel read fault] kernel: PGD 0 P4D 0 kernel: Oops: 0000 [#1] SMP PTI kernel: CPU: 1 PID: 1350 Comm: bash Tainted: G E 5.0.0-rc1-mm1-1-default+ #27 kernel: Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.0.0-prebuilt.qemu-project.org 04/01/2014 kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 kernel: Call Trace: kernel: ? klist_next+0x79/0xe0 kernel: memory_subsys_offline+0x42/0x60 kernel: device_offline+0x80/0xa0 kernel: state_store+0xab/0xc0 kernel: kernfs_fop_write+0x102/0x180 kernel: __vfs_write+0x26/0x190 kernel: ? set_close_on_exec+0x49/0x70 kernel: vfs_write+0xad/0x1b0 kernel: ksys_write+0x42/0x90 kernel: do_syscall_64+0x5b/0x180 kernel: entry_SYSCALL_64_after_hwframe+0x44/0xa9 kernel: RIP: 0033:0x7ff1719febe4 kernel: Code: 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb b7 0f 1f 80 00 00 00 00 8b 05 4a fc 2c 00 48 63 ff 85 c0 75 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 f3 c3 66 90 55 53 48 89 d5 48 89 f3 48 83 kernel: RSP: 002b:00007ffd50b7ddc8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 kernel: RAX: ffffffffffffffda RBX: 0000000000000008 RCX: 00007ff1719febe4 kernel: RDX: 0000000000000008 RSI: 00005556e9216b20 RDI: 0000000000000001 kernel: RBP: 00005556e9216b20 R08: 000000000000000a R09: 0000000000000000 kernel: R10: 000000000000000a R11: 0000000000000246 R12: 0000000000000008 kernel: R13: 0000000000000001 R14: 00007ff171cca720 R15: 0000000000000008 kernel: Modules linked in: af_packet(E) xt_tcpudp(E) ipt_REJECT(E) xt_conntrack(E) nf_conntrack(E) nf_defrag_ipv4(E) ip_set(E) nfnetlink(E) ebtable_nat(E) ebtable_broute(E) bridge(E) stp(E) llc(E) iptable_mangle(E) iptable_raw(E) iptable_security(E) ebtable_filter(E) ebtables(E) iptable_filter(E) ip_tables(E) x_tables(E) kvm_intel(E) kvm(E) irqbypass(E) crct10dif_pclmul(E) crc32_pclmul(E) ghash_clmulni_intel(E) bochs_drm(E) ttm(E) aesni_intel(E) drm_kms_helper(E) aes_x86_64(E) crypto_simd(E) cryptd(E) glue_helper(E) drm(E) virtio_net(E) syscopyarea(E) sysfillrect(E) net_failover(E) sysimgblt(E) pcspkr(E) failover(E) i2c_piix4(E) fb_sys_fops(E) parport_pc(E) parport(E) button(E) btrfs(E) libcrc32c(E) xor(E) zstd_decompress(E) zstd_compress(E) xxhash(E) raid6_pq(E) sd_mod(E) ata_generic(E) ata_piix(E) ahci(E) libahci(E) libata(E) crc32c_intel(E) serio_raw(E) virtio_pci(E) virtio_ring(E) virtio(E) sg(E) scsi_mod(E) autofs4(E) kernel: CR2: 0000000000000008 kernel: ---[ end trace bdb71590872849fb ]--- kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Fix this by getting the head page and testing against it.
[1] https://patchwork.kernel.org/patch/10739963/
--- mm/memory_hotplug.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-)
mm/memory_hotplug.c +10 -5
@@ -1335,12 +1335,17 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)if (__PageMovable (page))return pfn;if (PageHuge (page)) { -if (hugepage_migration_supported (page_hstate (page)) && -page_huge_active (page)) +struct page *head =compound_head (page); + +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head))return pfn; -else - pfn =round_up (pfn +1 , -1 <<compound_order (page)) -1 ; +else { +unsigned long skip; + + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; + } } } }
Signature · 1 line
2.13.7
Quoted 66 lines · Oscar Salvador
This is the same sort of error we saw in [1].
Gigantic hugepages crosses several memblocks, so it can be that the page we get in
scan_movable_pages()is a page-tail belonging to a 1G-hugepage.
If that happens, page_hstate()->size_to_hstate() will return NULL, and we will blow up inhugepage_migration_supported().The splat is as follows:
kernel: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 kernel: #PF error: [normal kernel read fault] kernel: PGD 0 P4D 0 kernel: Oops: 0000 [#1] SMP PTI kernel: CPU: 1 PID: 1350 Comm: bash Tainted: G E 5.0.0-rc1-mm1-1-default+ #27 kernel: Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.0.0-prebuilt.qemu-project.org 04/01/2014 kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 kernel: Call Trace: kernel: ? klist_next+0x79/0xe0 kernel: memory_subsys_offline+0x42/0x60 kernel: device_offline+0x80/0xa0 kernel: state_store+0xab/0xc0 kernel: kernfs_fop_write+0x102/0x180 kernel: __vfs_write+0x26/0x190 kernel: ? set_close_on_exec+0x49/0x70 kernel: vfs_write+0xad/0x1b0 kernel: ksys_write+0x42/0x90 kernel: do_syscall_64+0x5b/0x180 kernel: entry_SYSCALL_64_after_hwframe+0x44/0xa9 kernel: RIP: 0033:0x7ff1719febe4 kernel: Code: 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb b7 0f 1f 80 00 00 00 00 8b 05 4a fc 2c 00 48 63 ff 85 c0 75 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 f3 c3 66 90 55 53 48 89 d5 48 89 f3 48 83 kernel: RSP: 002b:00007ffd50b7ddc8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 kernel: RAX: ffffffffffffffda RBX: 0000000000000008 RCX: 00007ff1719febe4 kernel: RDX: 0000000000000008 RSI: 00005556e9216b20 RDI: 0000000000000001 kernel: RBP: 00005556e9216b20 R08: 000000000000000a R09: 0000000000000000 kernel: R10: 000000000000000a R11: 0000000000000246 R12: 0000000000000008 kernel: R13: 0000000000000001 R14: 00007ff171cca720 R15: 0000000000000008 kernel: Modules linked in: af_packet(E) xt_tcpudp(E) ipt_REJECT(E) xt_conntrack(E) nf_conntrack(E) nf_defrag_ipv4(E) ip_set(E) nfnetlink(E) ebtable_nat(E) ebtable_broute(E) bridge(E) stp(E) llc(E) iptable_mangle(E) iptable_raw(E) iptable_security(E) ebtable_filter(E) ebtables(E) iptable_filter(E) ip_tables(E) x_tables(E) kvm_intel(E) kvm(E) irqbypass(E) crct10dif_pclmul(E) crc32_pclmul(E) ghash_clmulni_intel(E) bochs_drm(E) ttm(E) aesni_intel(E) drm_kms_helper(E) aes_x86_64(E) crypto_simd(E) cryptd(E) glue_helper(E) drm(E) virtio_net(E) syscopyarea(E) sysfillrect(E) net_failover(E) sysimgblt(E) pcspkr(E) failover(E) i2c_piix4(E) fb_sys_fops(E) parport_pc(E) parport(E) button(E) btrfs(E) libcrc32c(E) xor(E) zstd_decompress(E) zstd_compress(E) xxhash(E) raid6_pq(E) sd_mod(E) ata_generic(E) ata_piix(E) ahci(E) libahci(E) libata(E) crc32c_intel(E) serio_raw(E) virtio_pci(E) virtio_ring(E) virtio(E) sg(E) scsi_mod(E) autofs4(E) kernel: CR2: 0000000000000008 kernel: ---[ end trace bdb71590872849fb ]--- kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Quoted 5 lines
Fix this by getting the head page and testing against it.
[1] https://patchwork.kernel.org/patch/10739963/
Signed-off-by Oscar Salvador <[email protected]>
Looks good.
Quoted 31 lines
--- mm/memory_hotplug.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-)mm/memory_hotplug.c +10
-5@@ -1335,12 +1335,17 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)if (__PageMovable (page))return pfn;if (PageHuge (page)) { -if (hugepage_migration_supported (page_hstate (page)) && -page_huge_active (page)) +struct page *head =compound_head (page); + +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head))return pfn; -else - pfn =round_up (pfn +1 , -1 <<compound_order (page)) -1 ; +else { +unsigned long skip; + + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; + } } } }
Quoted 66 lines · Oscar Salvador
This is the same sort of error we saw in [1].
Gigantic hugepages crosses several memblocks, so it can be that the page we get in
scan_movable_pages()is a page-tail belonging to a 1G-hugepage.
If that happens, page_hstate()->size_to_hstate() will return NULL, and we will blow up inhugepage_migration_supported().The splat is as follows:
kernel: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 kernel: #PF error: [normal kernel read fault] kernel: PGD 0 P4D 0 kernel: Oops: 0000 [#1] SMP PTI kernel: CPU: 1 PID: 1350 Comm: bash Tainted: G E 5.0.0-rc1-mm1-1-default+ #27 kernel: Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.0.0-prebuilt.qemu-project.org 04/01/2014 kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 kernel: Call Trace: kernel: ? klist_next+0x79/0xe0 kernel: memory_subsys_offline+0x42/0x60 kernel: device_offline+0x80/0xa0 kernel: state_store+0xab/0xc0 kernel: kernfs_fop_write+0x102/0x180 kernel: __vfs_write+0x26/0x190 kernel: ? set_close_on_exec+0x49/0x70 kernel: vfs_write+0xad/0x1b0 kernel: ksys_write+0x42/0x90 kernel: do_syscall_64+0x5b/0x180 kernel: entry_SYSCALL_64_after_hwframe+0x44/0xa9 kernel: RIP: 0033:0x7ff1719febe4 kernel: Code: 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb b7 0f 1f 80 00 00 00 00 8b 05 4a fc 2c 00 48 63 ff 85 c0 75 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 f3 c3 66 90 55 53 48 89 d5 48 89 f3 48 83 kernel: RSP: 002b:00007ffd50b7ddc8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 kernel: RAX: ffffffffffffffda RBX: 0000000000000008 RCX: 00007ff1719febe4 kernel: RDX: 0000000000000008 RSI: 00005556e9216b20 RDI: 0000000000000001 kernel: RBP: 00005556e9216b20 R08: 000000000000000a R09: 0000000000000000 kernel: R10: 000000000000000a R11: 0000000000000246 R12: 0000000000000008 kernel: R13: 0000000000000001 R14: 00007ff171cca720 R15: 0000000000000008 kernel: Modules linked in: af_packet(E) xt_tcpudp(E) ipt_REJECT(E) xt_conntrack(E) nf_conntrack(E) nf_defrag_ipv4(E) ip_set(E) nfnetlink(E) ebtable_nat(E) ebtable_broute(E) bridge(E) stp(E) llc(E) iptable_mangle(E) iptable_raw(E) iptable_security(E) ebtable_filter(E) ebtables(E) iptable_filter(E) ip_tables(E) x_tables(E) kvm_intel(E) kvm(E) irqbypass(E) crct10dif_pclmul(E) crc32_pclmul(E) ghash_clmulni_intel(E) bochs_drm(E) ttm(E) aesni_intel(E) drm_kms_helper(E) aes_x86_64(E) crypto_simd(E) cryptd(E) glue_helper(E) drm(E) virtio_net(E) syscopyarea(E) sysfillrect(E) net_failover(E) sysimgblt(E) pcspkr(E) failover(E) i2c_piix4(E) fb_sys_fops(E) parport_pc(E) parport(E) button(E) btrfs(E) libcrc32c(E) xor(E) zstd_decompress(E) zstd_compress(E) xxhash(E) raid6_pq(E) sd_mod(E) ata_generic(E) ata_piix(E) ahci(E) libahci(E) libata(E) crc32c_intel(E) serio_raw(E) virtio_pci(E) virtio_ring(E) virtio(E) sg(E) scsi_mod(E) autofs4(E) kernel: CR2: 0000000000000008 kernel: ---[ end trace bdb71590872849fb ]--- kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Quoted 3 lines
Fix this by getting the head page and testing against it.
So this should be probably folded into the above patch as it is incomplete unless I am missing something.
Quoted 1 line
Signed-off-by Oscar Salvador <[email protected]>
Other than that the change looks good to me.
Quoted 33 lines
--- mm/memory_hotplug.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-)mm/memory_hotplug.c +10
-5@@ -1335,12 +1335,17 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)if (__PageMovable (page))return pfn;if (PageHuge (page)) { -if (hugepage_migration_supported (page_hstate (page)) && -page_huge_active (page)) +struct page *head =compound_head (page); + +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head))return pfn; -else - pfn =round_up (pfn +1 , -1 <<compound_order (page)) -1 ; +else { +unsigned long skip; + + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; + } } } }Signature · 1 line
2.13.7
Signature · 2 lines
Michal Hocko SUSE Labs
Oscar Salvador
maintainerQuoted 2 lines · Michal Hocko
So this should be probably folded into the above patch as it is incomplete unless I am missing something.
Well, they are triggered from different paths.
The former error was triggered in:
removable_show
is_mem_section_removable
is_pageblock_removable_nolock
has_unmovable_pages
while this one is triggered when actually doing the offline operation
__offline_pages
scan_movable_pages
But I do agree that one without the other is not really useful, an incomplete. The truth is that I did not spot this one when fixing [1] because I did not really try to offline the memblock back then, so my fault.
While I agree that the best approach would be to fold this one into [1], I am not sure if it is too late for that as it seems that [1] was already released into mainline, and moreover to stable.
I guess I will have Andrew decide what is the best way to carry on here.
[1] https://patchwork.kernel.org/patch/10739963/
Quoted 6 lines
Quoted 2 lines
Signed-off-by Oscar Salvador <[email protected]>Other than that the change looks good to me.
Acked-by Michal Hocko <[email protected]>
Thanks!
Signature · 2 lines
Oscar Salvador SUSE L3
Quoted 7 lines · Oscar Salvador
Quoted 2 lines · Michal Hocko
So this should be probably folded into the above patch as it is incomplete unless I am missing something.
Well, they are triggered from different paths.
The former error was triggered in:
Quoted 6 lines
removable_show
is_mem_section_removable
is_pageblock_removable_nolock
has_unmovable_pageswhile this one is triggered when actually doing the offline operation
But it would trigger from the offline path as well, no?
Quoted 6 lines
__offline_pages
scan_movable_pagesBut I do agree that one without the other is not really useful, an incomplete. The truth is that I did not spot this one when fixing [1] because I did not really try to offline the memblock back then, so my fault.
I should have noticed that during the review but those paths are really far away from each other so this is hard to spot indeed
Quoted 3 lines
While I agree that the best approach would be to fold this one into [1], I am not sure if it is too late for that as it seems that [1] was already released into mainline, and moreover to stable.
OK, I wasn't aware of that. Then my suggestion is clearly moot.
Quoted 15 lines
I guess I will have Andrew decide what is the best way to carry on here.
[1] https://patchwork.kernel.org/patch/10739963/
Quoted 6 lines
Quoted 2 lines
Signed-off-by Oscar Salvador <[email protected]>Other than that the change looks good to me.
Acked-by Michal Hocko <[email protected]>Thanks!
Signature · 2 lines
Oscar Salvador SUSE L3
Signature · 2 lines
Michal Hocko SUSE Labs
Quoted 84 lines · Oscar Salvador
This is the same sort of error we saw in [1].
Gigantic hugepages crosses several memblocks, so it can be that the page we get in
scan_movable_pages()is a page-tail belonging to a 1G-hugepage.
If that happens, page_hstate()->size_to_hstate() will return NULL, and we will blow up inhugepage_migration_supported().The splat is as follows:
kernel: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 kernel: #PF error: [normal kernel read fault] kernel: PGD 0 P4D 0 kernel: Oops: 0000 [#1] SMP PTI kernel: CPU: 1 PID: 1350 Comm: bash Tainted: G E 5.0.0-rc1-mm1-1-default+ #27 kernel: Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.0.0-prebuilt.qemu-project.org 04/01/2014 kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 kernel: Call Trace: kernel: ? klist_next+0x79/0xe0 kernel: memory_subsys_offline+0x42/0x60 kernel: device_offline+0x80/0xa0 kernel: state_store+0xab/0xc0 kernel: kernfs_fop_write+0x102/0x180 kernel: __vfs_write+0x26/0x190 kernel: ? set_close_on_exec+0x49/0x70 kernel: vfs_write+0xad/0x1b0 kernel: ksys_write+0x42/0x90 kernel: do_syscall_64+0x5b/0x180 kernel: entry_SYSCALL_64_after_hwframe+0x44/0xa9 kernel: RIP: 0033:0x7ff1719febe4 kernel: Code: 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb b7 0f 1f 80 00 00 00 00 8b 05 4a fc 2c 00 48 63 ff 85 c0 75 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 f3 c3 66 90 55 53 48 89 d5 48 89 f3 48 83 kernel: RSP: 002b:00007ffd50b7ddc8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 kernel: RAX: ffffffffffffffda RBX: 0000000000000008 RCX: 00007ff1719febe4 kernel: RDX: 0000000000000008 RSI: 00005556e9216b20 RDI: 0000000000000001 kernel: RBP: 00005556e9216b20 R08: 000000000000000a R09: 0000000000000000 kernel: R10: 000000000000000a R11: 0000000000000246 R12: 0000000000000008 kernel: R13: 0000000000000001 R14: 00007ff171cca720 R15: 0000000000000008 kernel: Modules linked in: af_packet(E) xt_tcpudp(E) ipt_REJECT(E) xt_conntrack(E) nf_conntrack(E) nf_defrag_ipv4(E) ip_set(E) nfnetlink(E) ebtable_nat(E) ebtable_broute(E) bridge(E) stp(E) llc(E) iptable_mangle(E) iptable_raw(E) iptable_security(E) ebtable_filter(E) ebtables(E) iptable_filter(E) ip_tables(E) x_tables(E) kvm_intel(E) kvm(E) irqbypass(E) crct10dif_pclmul(E) crc32_pclmul(E) ghash_clmulni_intel(E) bochs_drm(E) ttm(E) aesni_intel(E) drm_kms_helper(E) aes_x86_64(E) crypto_simd(E) cryptd(E) glue_helper(E) drm(E) virtio_net(E) syscopyarea(E) sysfillrect(E) net_failover(E) sysimgblt(E) pcspkr(E) failover(E) i2c_piix4(E) fb_sys_fops(E) parport_pc(E) parport(E) button(E) btrfs(E) libcrc32c(E) xor(E) zstd_decompress(E) zstd_compress(E) xxhash(E) raid6_pq(E) sd_mod(E) ata_generic(E) ata_piix(E) ahci(E) libahci(E) libata(E) crc32c_intel(E) serio_raw(E) virtio_pci(E) virtio_ring(E) virtio(E) sg(E) scsi_mod(E) autofs4(E) kernel: CR2: 0000000000000008 kernel: ---[ end trace bdb71590872849fb ]--- kernel: RIP: 0010:__offline_pages+0x6ae/0x900 kernel: Code: 48 c7 c6 d0 3e a4 81 e8 44 c8 ad ff 49 8b 04 24 bf 00 10 00 00 a9 00 00 01 00 74 09 41 0f b6 4c 24 51 48 d3 e7 e8 42 2a c1 ff <8b> 40 08 83 f8 09 0f 84 b0 fc ff ff 83 f8 12 0f 84 a7 fc ff ff 83 kernel: RSP: 0018:ffffc900008e3d20 EFLAGS: 00010246 kernel: RAX: 0000000000000000 RBX: ffffea0000000000 RCX: 0000000000000009 kernel: RDX: ffffffff825c64f0 RSI: 0000000000001000 RDI: 0000000000001000 kernel: RBP: ffffc900008e3d68 R08: 0000000000200000 R09: 00000000000001e4 kernel: R10: 0000000000000058 R11: ffffffff8254a854 R12: ffffea0004200000 kernel: R13: 0000000000108000 R14: 0000000000110000 R15: 0000000000000000 kernel: FS: 00007ff172339b80(0000) GS:ffff88803eb00000(0000) knlGS:0000000000000000 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: CR2: 0000000000000008 CR3: 0000000038d78006 CR4: 00000000003606a0 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400Fix this by getting the head page and testing against it.
[1] https://patchwork.kernel.org/patch/10739963/
Signed-off-by Oscar Salvador <[email protected]>--- mm/memory_hotplug.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-)mm/memory_hotplug.c +0
-1@@ -1335,12 +1335,17 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)if (__PageMovable (page))return pfn;if (PageHuge (page)) { - if (hugepage_migration_supported(page_hstate(page)) &&
mm/memory_hotplug.c static unsigned long scan_movable_pages(unsigned long start, unsigned long end) · line 1343
Quoted 10 lines
mm/memory_hotplug.c +5
-4- page_huge_active(page)) +struct page *head =compound_head (page); + +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head))return pfn; - else - pfn =round_up (pfn +1 , -1 <<compound_order (page)) -1 ; +else {
If you use {} for the else case, please also do so for the if case.
Apart from that this looks good to me
Quoted 9 lines
mm/memory_hotplug.c +5
-0+unsigned long skip; + + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; + } } } }
Signature · 3 lines
Thanks, David / dhildenb
Oscar Salvador
maintainerQuoted 1 line · David Hildenbrand
If you use {} for the else case, please also do so for the if case.
Diff on top:
mm/memory_hotplug.c +2 -2
@@ -1338,9 +1338,9 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)struct page *head =compound_head (page);if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) +page_huge_active (head)) {return pfn; -else { + }else {unsigned long skip; skip = (1 <<compound_order (head)) - (page - head);
Quoted 3 lines
Apart from that this looks good to me
Reviewed-by David Hildenbrand <[email protected]>
Thanks David ;-)
Signature · 2 lines
Oscar Salvador SUSE L3
Andrew Morton
maintainerQuoted 12 lines · On Fri, 25 Jan 2019 08:58:33 +0100 Oscar Salvador <[email protected]> wrote:
Quoted 1 line · David Hildenbrand
If you use {} for the else case, please also do so for the if case.
Diff on top:
mm/memory_hotplug.c +0
-0@@ -1338,9 +1338,9 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)struct page *head =compound_head (page);
mm/memory_hotplug.c static unsigned long scan_movable_pages(unsigned long start, unsigned long end) · line 1346
Quoted 10 lines
mm/memory_hotplug.c +2
-2if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) +page_huge_active (head)) {return pfn; -else { + }else {unsigned long skip; skip = (1 <<compound_order (head)) - (page - head);
The indenting is getting a bit deep also, so how about this?
static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
{
unsigned long pfn;
for (pfn = start; pfn < end; pfn++) {
struct page *page, *head;
if (!pfn_valid(pfn))
continue;
page = pfn_to_page(pfn);
if (PageLRU(page))
return pfn;
if (__PageMovable(page))
return pfn;
if (!PageHuge(page))
continue;
head = compound_head(page);
if (hugepage_migration_supported(page_hstate(head)) &&
page_huge_active(head)) {
return pfn;
} else {
unsigned long skip;
skip = (1 << compound_order(head)) - (page - head);
pfn += skip - 1;
}
}
return 0;
}From: Andrew Morton <[email protected]>
Subject: mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix
fix brace layout, per David. Also reduce indentation
---
mm/memory_hotplug.c | 38 ++++++++++++++++++++------------------ 1 file changed, 20 insertions(+), 18 deletions(-)
mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +20 -18
@@ -1305,27 +1305,29 @@ int test_pages_in_a_zone(unsigned long sstatic unsigned long scan_movable_pages (unsigned long start,unsigned long end) {unsigned long pfn; -struct page *page; +for (pfn = start; pfn < end; pfn++) { -if (pfn_valid (pfn)) { - page =pfn_to_page (pfn); -if (PageLRU (page)) -return pfn; -if (__PageMovable (page)) -return pfn; -if (PageHuge (page)) { -struct page *head =compound_head (page); +struct page *page, *head; + +if (!pfn_valid (pfn)) +continue ; + page =pfn_to_page (pfn); +if (PageLRU (page)) +return pfn; +if (__PageMovable (page)) +return pfn; -if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) -return pfn; -else { -unsigned long skip; +if (!PageHuge (page)) +continue ; + head =compound_head (page); +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head)) { +return pfn; + }else { +unsigned long skip; - skip = (1 <<compound_order (head)) - (page - head); - pfn += skip -1 ; - } - } + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; } }return 0 ;
_
Andrew Morton
maintainerQuoted 40 lines · On Mon, 28 Jan 2019 14:53:09 -0800 Andrew Morton <[email protected]>…
Quoted 12 lines · On Fri, 25 Jan 2019 08:58:33 +0100 Oscar Salvador <[email protected]> wrote:
Quoted 1 line · David Hildenbrand
If you use {} for the else case, please also do so for the if case.
Diff on top:
mm/memory_hotplug.c +0
-0@@ -1338,9 +1338,9 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)struct page *head =compound_head (page);mm/memory_hotplug.c static unsigned long scan_movable_pages(unsigned long start, unsigned long end) · line 1346
Quoted 10 lines
mm/memory_hotplug.c +2
-2if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) +page_huge_active (head)) {return pfn; -else { + }else {unsigned long skip; skip = (1 <<compound_order (head)) - (page - head);The indenting is getting a bit deep also, so how about this?
static unsigned long
scan_movable_pages(unsigned long start, unsigned long end){ unsigned long pfn; for (pfn = start; pfn < end; pfn++) { struct page *page, *head; if (!pfn_valid(pfn)) continue; page = pfn_to_page(pfn); if (PageLRU(page)) return pfn; if (__PageMovable(page))
Quoted 8 lines
return pfn;
if (!PageHuge(page))
continue;
head = compound_head(page);
if (hugepage_migration_supported(page_hstate(head)) &&
page_huge_active(head)) {
return pfn;checkpatch pointed out that else-after-return isn't needed so we can do
static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
{
unsigned long pfn;
for (pfn = start; pfn < end; pfn++) {
struct page *page, *head;
unsigned long skip;
if (!pfn_valid(pfn))
continue;
page = pfn_to_page(pfn);
if (PageLRU(page))
return pfn;
if (__PageMovable(page))
return pfn;
if (!PageHuge(page))
continue;
head = compound_head(page);
if (hugepage_migration_supported(page_hstate(head)) &&
page_huge_active(head))
return pfn;
skip = (1 << compound_order(head)) - (page - head);
pfn += skip - 1;
}
return 0;
}mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +18 -19
@@ -1305,28 +1305,27 @@ int test_pages_in_a_zone(unsigned long sstatic unsigned long scan_movable_pages (unsigned long start,unsigned long end) {unsigned long pfn; -struct page *page; +for (pfn = start; pfn < end; pfn++) { -if (pfn_valid (pfn)) { - page =pfn_to_page (pfn); -if (PageLRU (page)) -return pfn; -if (__PageMovable (page)) -return pfn; -if (PageHuge (page)) { -struct page *head =compound_head (page); +struct page *page, *head; +unsigned long skip; -if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) -return pfn; -else { -unsigned long skip; +if (!pfn_valid (pfn)) +continue ; + page =pfn_to_page (pfn); +if (PageLRU (page)) +return pfn; +if (__PageMovable (page)) +return pfn; - skip = (1 <<compound_order (head)) - (page - head); - pfn += skip -1 ; - } - } - } +if (!PageHuge (page)) +continue ; + head =compound_head (page); +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head)) +return pfn; + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; }return 0 ; }
_
On Mon 28-01-19 14:56:17, Andrew Morton wrote:
[...]
Quoted 49 lines
mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +18
-19@@ -1305,28 +1305,27 @@ int test_pages_in_a_zone(unsigned long sstatic unsigned long scan_movable_pages (unsigned long start,unsigned long end) {unsigned long pfn; -struct page *page; +for (pfn = start; pfn < end; pfn++) { -if (pfn_valid (pfn)) { - page =pfn_to_page (pfn); -if (PageLRU (page)) -return pfn; -if (__PageMovable (page)) -return pfn; -if (PageHuge (page)) { -struct page *head =compound_head (page); +struct page *page, *head; +unsigned long skip; -if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) -return pfn; -else { -unsigned long skip; +if (!pfn_valid (pfn)) +continue ; + page =pfn_to_page (pfn); +if (PageLRU (page)) +return pfn; +if (__PageMovable (page)) +return pfn; - skip = (1 <<compound_order (head)) - (page - head); - pfn += skip -1 ; - } - } - } +if (!PageHuge (page)) +continue ; + head =compound_head (page); +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head)) +return pfn; + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; }return 0 ; }
Quoted 2 lines
_
LGTM
Signature · 2 lines
Michal Hocko SUSE Labs
Oscar Salvador
maintainerQuoted 40 lines · Andrew Morton
mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +11
-19@@ -1305,28 +1305,27 @@ int test_pages_in_a_zone(unsigned long sstatic unsigned long scan_movable_pages (unsigned long start,unsigned long end) {unsigned long pfn; -struct page *page; +for (pfn = start; pfn < end; pfn++) { -if (pfn_valid (pfn)) { - page =pfn_to_page (pfn); -if (PageLRU (page)) -return pfn; -if (__PageMovable (page)) -return pfn; -if (PageHuge (page)) { -struct page *head =compound_head (page); +struct page *page, *head; +unsigned long skip; -if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) -return pfn; -else { -unsigned long skip; +if (!pfn_valid (pfn)) +continue ; + page =pfn_to_page (pfn); +if (PageLRU (page)) +return pfn; +if (__PageMovable (page)) +return pfn; - skip = (1 <<compound_order (head)) - (page - head); - pfn += skip -1 ; - } - } - } +if (!PageHuge (page))
mm/memory_hotplug.c int test_pages_in_a_zone(unsigned long s · line 1331
Quoted 10 lines
mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +7
-0+continue ; + head =compound_head (page); +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head)) +return pfn; + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ; }return 0 ; }
It looks much better, thanks a lot for the cleanup Andrew!
Signature · 2 lines
Oscar Salvador SUSE L3
Quoted 117 lines · Andrew Morton
Quoted 40 lines · On Mon, 28 Jan 2019 14:53:09 -0800 Andrew Morton <[email protected]>…
Quoted 12 lines · On Fri, 25 Jan 2019 08:58:33 +0100 Oscar Salvador <[email protected]> wrote:
Quoted 1 line · David Hildenbrand
If you use {} for the else case, please also do so for the if case.
Diff on top:
mm/memory_hotplug.c +0
-0@@ -1338,9 +1338,9 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)struct page *head =compound_head (page);mm/memory_hotplug.c static unsigned long scan_movable_pages(unsigned long start, unsigned long end) · line 1346
Quoted 10 lines
mm/memory_hotplug.c +2
-2if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) +page_huge_active (head)) {return pfn; -else { + }else {unsigned long skip; skip = (1 <<compound_order (head)) - (page - head);The indenting is getting a bit deep also, so how about this?
static unsigned long
scan_movable_pages(unsigned long start, unsigned long end){ unsigned long pfn; for (pfn = start; pfn < end; pfn++) { struct page *page, *head; if (!pfn_valid(pfn)) continue; page = pfn_to_page(pfn); if (PageLRU(page)) return pfn; if (__PageMovable(page))Quoted 8 lines
return pfn; if (!PageHuge(page)) continue; head = compound_head(page); if (hugepage_migration_supported(page_hstate(head)) && page_huge_active(head)) { return pfn;checkpatch pointed out that else-after-return isn't needed so we can do
static unsigned long
scan_movable_pages(unsigned long start, unsigned long end){ unsigned long pfn; for (pfn = start; pfn < end; pfn++) { struct page *page, *head; unsigned long skip; if (!pfn_valid(pfn)) continue; page = pfn_to_page(pfn); if (PageLRU(page)) return pfn; if (__PageMovable(page)) return pfn; if (!PageHuge(page)) continue; head = compound_head(page); if (hugepage_migration_supported(page_hstate(head)) && page_huge_active(head)) return pfn; skip = (1 << compound_order(head)) - (page - head); pfn += skip - 1; } return 0; }mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +10
-17@@ -1305,28 +1305,27 @@ int test_pages_in_a_zone(unsigned long sstatic unsigned long scan_movable_pages (unsigned long start,unsigned long end) {unsigned long pfn; -struct page *page; +for (pfn = start; pfn < end; pfn++) { -if (pfn_valid (pfn)) { - page =pfn_to_page (pfn); -if (PageLRU (page)) -return pfn; -if (__PageMovable (page)) -return pfn; -if (PageHuge (page)) { -struct page *head =compound_head (page); +struct page *page, *head; +unsigned long skip; -if (hugepage_migration_supported (page_hstate (head)) && -page_huge_active (head)) -return pfn; -else { -unsigned long skip; +if (!pfn_valid (pfn)) +continue ; + page =pfn_to_page (pfn); +if (PageLRU (page)) +return pfn; +if (__PageMovable (page)) +return pfn; - skip = (1 <<compound_order (head)) - (page - head); - pfn += skip -1 ; - }
mm/memory_hotplug.c int test_pages_in_a_zone(unsigned long s · line 1328
Quoted 10 lines
mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +8
-2- } - } +if (!PageHuge (page)) +continue ; + head =compound_head (page); +if (hugepage_migration_supported (page_hstate (head)) && +page_huge_active (head)) +return pfn; + skip = (1 <<compound_order (head)) - (page - head); + pfn += skip -1 ;
Not sure if encoding the -1 in the previous line is even better now that we have more space
skip = (1 << compound_order(head)) - (page - head + 1);
Looks good to me.
Quoted 5 lines
mm/memory_hotplug.c renamed from mm/memory_hotplug.c~mmmemory_hotplug-fix-scan_movable_pages-for-gigantic-hugepages-fix +0
-0}return 0 ; }_
Signature · 3 lines
Thanks, David / dhildenb
Oscar Salvador
maintainerQuoted 8 lines · David Hildenbrand
Not sure if encoding the -1 in the previous line is even better now that we have more space
skip = (1 <<
compound_order(head)) - (page - head + 1);Looks good to me.
Quoted 1 line
}
I would rather not do that.
For me looks a bit "subtle" why do we add up 1 after substracting page - head. I think that doing "skip - 1" looks more clear, but I do not have a strong opinion here.
Signature · 2 lines
Oscar Salvador SUSE L3
Andrew Morton
maintainerQuoted 1 line · On Tue, 22 Jan 2019 16:44:07 +0100 Oscar Salvador <[email protected]> wrote:
This is the same sort of error we saw in [1].
I'll replace "[1]" with 17e2e7d7e1b83 ("mm, page_alloc: fix has_unmovable_pages for HugePages").
Quoted 1 line
Signed-off-by Oscar Salvador <[email protected]>
And I'll add cc:stable.