kernel/git/stable/linux.git/fs/btrfs/fs.h, branch master

btrfs: trigger cow fixup via dirty_folio()

2026-07-30T17:28:36+00:00

The problem scenario: If we have a folio mmapped shared and then somebody does a dio read with that folio as the read destination, then it is possible that the dio will see a dirty destination page when it starts (and thus skip dirtying and just GUP pin it) but then while it is doing the read, btrfs finishes writing it back and by the endio, the folio is clean. In that case, the dio read must re-dirty the folio with aops->dirty_folio(): btrfs_check_read_bio() |- __iomap_dio_bio_end_io() from btrfs_bio_end_io() |- bio_check_pages_dirty() |- bio_dirty_fn() |- bio_release_pages(bio, true) |- __bio_release_pages(bio, mark_dirty == true) |- folio_lock() |- folio_mark_dirty() |- aops->dirty_folio() |- folio_unlock() A data block normally moves through writeback as follows: TASK folio_lock write clean -> dirty bit + delalloc folio_unlock WRITEBACK for-each-dirty-folio: folio_lock run_delalloc delalloc consumed -> dirty bit + OE submission dirty bit consumed -> writeback bit + OE folio_unlock ENDIO endio OE bytes accounted OE finish writeback -> clean; destroy OE Three critical invariants that this path maintains are: I1. Any dirty block is covered by delalloc xor an ordered extent I2. Any dirty block covered by an OE will be submitted into that OE I3. Any dirty block already submitted into an OE will not be submitted again into the same OE. These ensure that the block will be written exactly once. It is clear that not reserving delalloc for the re-dirty case violates I1. This situation, even without bs < folio_size, has long required btrfs to fixup such dirty pages during writeback with an asynchronous worker that is allowed to do this expensive work and writeback does not proceed for a folio while it is doing this work. Commit 247e743cbe6e ("Btrfs: Use async helpers to deal with pages that have been improperly dirtied") introduced the COW fixup to catch exactly this class at writeback, way back in 2008. Since then, there have been many advances to prevent most of the causes of such re-dirtying and we thought we could get away with removing the annoying cow-fixup in the hope of simplifying writeback for large folio support. Commit b2a9f217ad3f ("btrfs: remove the COW fixup mechanism") Commit 4927b141877c ("btrfs: remove folio ordered flag and subpage bitmap") Since it turns out this assumption was incorrect, as evidenced by the report and attendant reproducers, we must reintroduce the fixup concept. This is of course critically further complicated by bs < folio_size. In that case, rather than just a folio dirty bit, we have a bitmap for the dirty blocks in the folio. And the (also broken) invariant is: I4. folio dirty IFF at least one block bitmap dirty. The original report of a stall on a misinterpreted empty bitmap is exactly evidence of a violation of I4. It is exactly because of bs < folio_size we don't want to simply revert the removal patches. The original fixup was not properly bs < folio_size aware, which motivated removal in the first place. So we wish to build a bs < folio_size aware fixup. One other important detail from the old design, any normal write that happens after a re-dirty but before a fixup is racing with the cow fixup to do the delalloc reservation, therefore it must cancel the fixup state. If it arrives after the reservation exists, it will be a normal dirty overwrite. This critically informs the design in a pretty clear way. fixup requiring re-dirty has folio granularity, while cancellation has delalloc (block) granularity so while we only ever produce fixup in chunks of folios, we must be able to clear it in blocks. Therefore we must track the blocks needing fixup at block granularity. The obvious way to do this is with a new bitmap in btrfs_folio_state, but it is desirable to avoid that if possible. Unfortunately, I don't think it is possible and the reason is subtle and leans on a sort of extreme reproducer, but I think can be explained relatively succinctly. Consider a folio whose two halves will land in different ordered extents (can be accomplished with tricks using nodatasum) and a dio read is running with it as the shared mmap destination. 1. The front half: a. folio comes clean on a normal write b. dio read completes into the folio marking it fixup. c. a write comes for the previous folio for a range extending into this folio, this is a cancellation of the fixup which reserves space. d. writeback runs on the range *not* overlapping the folio. This half remains dirty but is now covered by an OE and is awaiting writeback running on its range to be submitted and finish the OE. 2. The back half: a. the folio is part of an OE that gets far enough along to clear writeback. b. dio read completes into the folio marking it fixup. After this, the folio's front half is dirty in the "normal" sense, it needs to be submitted to the OE waiting for it. It's a cancelled fixup. Meanwhile, the second half is a true fresh fixup. So at this point if we run writeback on this folio, we genuinely can't know what to do without block level information. If we submit it, we submit unreserved dirty from the back half. If we don't, we will never finish the OE waiting for it. So it's either a corruption or a deadlock. Thus, the full high level design picture: - btrfs_data_dirty_folio(): For out of band non-reserving dirties, mark still-clean blocks inside EOF dirty and set their fixup bits (the event carries no range, so every clean block is suspect). Already-dirty blocks are covered or pending and are left alone. - Writeback: skip fixup blocks and enqueue work for them - writepage_fixup(): for each fixup block do the fixup reservation in a worker, after which the blocks can be written back normally. - Typical reserving write paths cancel fixup state for the ranges they cover with btrfs_folio_cancel_fixup() Link: https://lore.kernel.org/linux-btrfs/20260721191152.101118-1-borntraeger@linux.ibm.com/ Assisted-by: LLM Reviewed-by: Qu Wenruo Signed-off-by: Boris Burkov Signed-off-by: David Sterba

btrfs: skip global block reserve accounting for rescue mounts

2026-07-21T04:40:18+00:00

[BUG] Mounting with rescue=ibadroots after corrupting the block group tree root triggers a NULL pointer dereference: BUG: kernel NULL pointer dereference, address: 0000000000000100 RIP: 0010:btrfs_update_global_block_rsv+0x9d/0x1c0 [btrfs] Call Trace: fill_dummy_bgs+0xd4/0x120 [btrfs] open_ctree+0xc6e/0x1ca0 [btrfs] btrfs_get_tree+0x50d/0xa40 [btrfs] The same crash occurs with a corrupted raid stripe tree root, via btrfs_read_block_groups() instead of fill_dummy_bgs(). [CAUSE] With rescue=ibadroots, btrfs_read_roots() allows the mount to continue when either root cannot be read, leaving the corresponding root pointer NULL while its on-disk feature bit remains set. btrfs_update_global_block_rsv() then dereferences the missing root based on the feature bit alone. [FIX] Rescue mounts are fully read-only and cannot start transactions, so the global reserve is never consumed. Under btrfs_is_full_ro(), mark the reserve as full and return before performing the accounting. And since we need to check if the fs is mount fully RO, export fs_is_full_ro() as btrfs_is_full_ro(), and move it to fs.h. Fixes: 8dbfc14fc736 ("btrfs: account block group tree when calculating global reserve size") Fixes: 515020900d44 ("btrfs: read raid stripe tree from disk") Suggested-by: Qu Wenruo Signed-off-by: Dongjiang Zhu [ Squash the fs_is_full_ro() export commit into this one. ] Reviewed-by: Qu Wenruo Signed-off-by: David Sterba

btrfs: fix use-after-free after relocation failure with concurrent COW

2026-06-09T16:22:47+00:00

If we get a failure during relocation, before we update all the extent buffers that have file extent items pointing to extents from the block group being relocated, we can trigger a user-after-free on the reloc control structure (fs_info->reloc_control) if we have a concurrent task that is COWing a subvolume leaf. This happens like this: 1) Relocation of data block group X starts; 2) Relocation changes its state to UPDATE_DATA_PTRS; 3) A task doing a rename for example, COWs leaf A from a subvolume tree and ends up at btrfs_reloc_cow_block() and extracts fs_info->reloc_ctl into a local variable, which then passes to replace_file_extents(); 4) The relocation task gets an error and under the label 'out_put_bg' in btrfs_relocate_block_group() calls free_reloc_control(), which frees the reloc control structure that the rename task is using; 5) The rename task triggers a use-after-free on the reloc control structure that was just freed. Syzbot reported this recently, with the following stack trace: [ 88.389822][ T5325] BTRFS error (device loop0 state A): Transaction aborted (error -5) [ 88.389842][ T5325] BTRFS: error (device loop0 state A) in cleanup_transaction:2067: errno=-5 IO failure [ 88.389864][ T5325] BTRFS info (device loop0 state EA): forced readonly [ 88.392277][ T5324] BTRFS: error (device loop0 state EA) in btrfs_sync_log:3572: errno=-5 IO failure [ 88.396630][ T5325] BTRFS info (device loop0 state EA): balance: ended with status: -5 [ 88.400135][ T5346] ================================================================== [ 88.400148][ T5346] BUG: KASAN: slab-use-after-free in replace_file_extents+0x85f/0x1590 [ 88.400288][ T5346] Read of size 8 at addr ffff888012312010 by task syz.0.0/5346 [ 88.400299][ T5346] [ 88.400306][ T5346] CPU: 0 UID: 0 PID: 5346 Comm: syz.0.0 Not tainted syzkaller #0 PREEMPT(full) [ 88.400319][ T5346] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2 04/01/2014 [ 88.400325][ T5346] Call Trace: [ 88.400331][ T5346] [ 88.400336][ T5346] dump_stack_lvl+0xe8/0x150 [ 88.400351][ T5346] print_address_description+0x55/0x1e0 [ 88.400364][ T5346] ? replace_file_extents+0x85f/0x1590 [ 88.400378][ T5346] print_report+0x58/0x70 [ 88.400389][ T5346] kasan_report+0x117/0x150 [ 88.400405][ T5346] ? replace_file_extents+0x85f/0x1590 [ 88.400420][ T5346] replace_file_extents+0x85f/0x1590 [ 88.400440][ T5346] ? __pfx_replace_file_extents+0x10/0x10 [ 88.400452][ T5346] ? update_ref_for_cow+0xa71/0x1270 [ 88.400473][ T5346] btrfs_force_cow_block+0xa4d/0x2450 [ 88.400492][ T5346] ? __pfx_btrfs_force_cow_block+0x10/0x10 [ 88.400508][ T5346] ? __pfx_btrfs_get_32+0x10/0x10 [ 88.400523][ T5346] btrfs_cow_block+0x3c4/0xa90 [ 88.400542][ T5346] push_leaf_left+0x2ac/0x4a0 [ 88.400561][ T5346] split_leaf+0xd16/0x12e0 [ 88.400574][ T5346] ? btrfs_bin_search+0x924/0xc70 [ 88.400592][ T5346] ? __pfx_split_leaf+0x10/0x10 [ 88.400602][ T5346] ? leaf_space_used+0x177/0x1e0 [ 88.400618][ T5346] ? btrfs_leaf_free_space+0x14a/0x2f0 [ 88.400634][ T5346] btrfs_search_slot+0x2641/0x2d20 [ 88.400654][ T5346] ? __pfx_btrfs_search_slot+0x10/0x10 [ 88.400669][ T5346] ? rcu_is_watching+0x15/0xb0 [ 88.400681][ T5346] ? trace_kmem_cache_alloc+0x29/0xe0 [ 88.400694][ T5346] btrfs_insert_empty_items+0x9c/0x190 [ 88.400711][ T5346] btrfs_insert_inode_ref+0x229/0xcb0 [ 88.400724][ T5346] ? __pfx_btrfs_insert_inode_ref+0x10/0x10 [ 88.400736][ T5346] ? __pfx_btrfs_qgroup_convert_reserved_meta+0x10/0x10 [ 88.400751][ T5346] ? btrfs_record_root_in_trans+0x124/0x180 [ 88.400767][ T5346] ? start_transaction+0x8a0/0x1820 [ 88.400778][ T5346] ? btrfs_set_inode_index+0x5e/0x100 [ 88.400787][ T5346] btrfs_rename2+0x17bb/0x40d0 [ 88.400800][ T5346] ? check_noncircular+0xda/0x150 [ 88.400814][ T5346] ? add_lock_to_list+0xc7/0x100 [ 88.400828][ T5346] ? __pfx_btrfs_rename2+0x10/0x10 [ 88.400842][ T5346] ? lockdep_hardirqs_on+0x7a/0x110 [ 88.400901][ T5346] ? lock_acquire+0x221/0x350 [ 88.400915][ T5346] ? down_write_nested+0x174/0x210 [ 88.400931][ T5346] ? __pfx_down_write_nested+0x10/0x10 [ 88.400941][ T5346] ? do_raw_spin_unlock+0x4d/0x210 [ 88.400952][ T5346] ? try_break_deleg+0x5b/0x180 [ 88.400963][ T5346] ? __pfx_btrfs_rename2+0x10/0x10 [ 88.400973][ T5346] vfs_rename+0xa96/0xeb0 [ 88.400992][ T5346] ? __pfx_vfs_rename+0x10/0x10 [ 88.401010][ T5346] ovl_fill_super+0x46b7/0x5e20 [ 88.401030][ T5346] ? __pfx_ovl_fill_super+0x10/0x10 [ 88.401042][ T5346] ? xas_create+0x1902/0x1b90 [ 88.401060][ T5346] ? __pfx___mutex_trylock_common+0x10/0x10 [ 88.401076][ T5346] ? trace_contention_end+0x3d/0x140 [ 88.401094][ T5346] ? shrinker_register+0x124/0x230 [ 88.401111][ T5346] ? __mutex_unlock_slowpath+0x1be/0x6f0 [ 88.401127][ T5346] ? shrinker_register+0x61/0x230 [ 88.401143][ T5346] ? __pfx___mutex_lock+0x10/0x10 [ 88.401158][ T5346] ? __pfx___mutex_unlock_slowpath+0x10/0x10 [ 88.401177][ T5346] ? __raw_spin_lock_init+0x45/0x100 [ 88.401196][ T5346] ? sget_fc+0x962/0xa40 [ 88.401208][ T5346] ? __pfx_set_anon_super_fc+0x10/0x10 [ 88.401222][ T5346] ? __pfx_ovl_fill_super+0x10/0x10 [ 88.401241][ T5346] get_tree_nodev+0xbb/0x150 [ 88.401257][ T5346] vfs_get_tree+0x92/0x2a0 [ 88.401272][ T5346] do_new_mount+0x341/0xd30 [ 88.401283][ T5346] ? apparmor_capable+0x126/0x170 [ 88.401301][ T5346] ? __pfx_do_new_mount+0x10/0x10 [ 88.401311][ T5346] ? ns_capable+0x89/0xe0 [ 88.401322][ T5346] ? path_mount+0x690/0x10e0 [ 88.401333][ T5346] ? user_path_at+0xd4/0x160 [ 88.401346][ T5346] __se_sys_mount+0x31d/0x420 [ 88.401358][ T5346] ? __pfx___se_sys_mount+0x10/0x10 [ 88.401370][ T5346] ? __x64_sys_mount+0x20/0xc0 [ 88.401381][ T5346] ? entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 88.401391][ T5346] do_syscall_64+0x15f/0xf80 [ 88.401403][ T5346] ? trace_irq_disable+0x3b/0x140 [ 88.401413][ T5346] ? clear_bhb_loop+0x40/0x90 [ 88.401421][ T5346] entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 88.401429][ T5346] RIP: 0033:0x7fa1ff79ce59 [ 88.401436][ T5346] Code: ff c3 66 (...) [ 88.401443][ T5346] RSP: 002b:00007fa2005affe8 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5 [ 88.401456][ T5346] RAX: ffffffffffffffda RBX: 00007fa1ffa16180 RCX: 00007fa1ff79ce59 [ 88.401464][ T5346] RDX: 0000200000000100 RSI: 0000200000002240 RDI: 0000000000000000 [ 88.401474][ T5346] RBP: 00007fa1ff832d6f R08: 0000200000000440 R09: 0000000000000000 [ 88.401481][ T5346] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 [ 88.401488][ T5346] R13: 00007fa1ffa16218 R14: 00007fa1ffa16180 R15: 00007ffc734fba78 [ 88.401500][ T5346] [ 88.401506][ T5346] [ 88.401510][ T5346] Allocated by task 5325: [ 88.401516][ T5346] kasan_save_track+0x3e/0x80 [ 88.401529][ T5346] __kasan_kmalloc+0x93/0xb0 [ 88.401542][ T5346] __kmalloc_cache_noprof+0x31c/0x660 [ 88.401554][ T5346] btrfs_relocate_block_group+0x217/0xc40 [ 88.401568][ T5346] btrfs_relocate_chunk+0x115/0x820 [ 88.401577][ T5346] __btrfs_balance+0x1db0/0x2ae0 [ 88.401587][ T5346] btrfs_balance+0xaf3/0x11b0 [ 88.401596][ T5346] btrfs_ioctl_balance+0x3d3/0x610 [ 88.401612][ T5346] __se_sys_ioctl+0xfc/0x170 [ 88.401626][ T5346] do_syscall_64+0x15f/0xf80 [ 88.401640][ T5346] entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 88.401650][ T5346] [ 88.401653][ T5346] Freed by task 5325: [ 88.401659][ T5346] kasan_save_track+0x3e/0x80 [ 88.401671][ T5346] kasan_save_free_info+0x46/0x50 [ 88.401680][ T5346] __kasan_slab_free+0x5c/0x80 [ 88.401692][ T5346] kfree+0x1c5/0x640 [ 88.401703][ T5346] btrfs_relocate_block_group+0x95d/0xc40 [ 88.401715][ T5346] btrfs_relocate_chunk+0x115/0x820 [ 88.401724][ T5346] __btrfs_balance+0x1db0/0x2ae0 [ 88.401733][ T5346] btrfs_balance+0xaf3/0x11b0 [ 88.401742][ T5346] btrfs_ioctl_balance+0x3d3/0x610 [ 88.401757][ T5346] __se_sys_ioctl+0xfc/0x170 [ 88.401770][ T5346] do_syscall_64+0x15f/0xf80 [ 88.401785][ T5346] entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 88.401795][ T5346] [ 88.401798][ T5346] The buggy address belongs to the object at ffff888012312000 [ 88.401798][ T5346] which belongs to the cache kmalloc-2k of size 2048 [ 88.401807][ T5346] The buggy address is located 16 bytes inside of [ 88.401807][ T5346] freed 2048-byte region [ffff888012312000, ffff888012312800) [ 88.401819][ T5346] [ 88.401822][ T5346] The buggy address belongs to the physical page: [ 88.401829][ T5346] page: refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x12310 [ 88.401840][ T5346] head: order:3 mapcount:0 entire_mapcount:0 nr_pages_mapped:0 pincount:0 [ 88.401849][ T5346] flags: 0xfff00000000040(head|node=0|zone=1|lastcpupid=0x7ff) [ 88.401860][ T5346] page_type: f5(slab) [ 88.401871][ T5346] raw: 00fff00000000040 ffff88801ac42000 dead000000000100 dead000000000122 [ 88.401881][ T5346] raw: 0000000000000000 0000000800080008 00000000f5000000 0000000000000000 [ 88.401892][ T5346] head: 00fff00000000040 ffff88801ac42000 dead000000000100 dead000000000122 [ 88.401902][ T5346] head: 0000000000000000 0000000800080008 00000000f5000000 0000000000000000 [ 88.401913][ T5346] head: 00fff00000000003 fffffffffffffe01 00000000ffffffff 00000000ffffffff [ 88.401923][ T5346] head: ffffffffffffffff 0000000000000000 00000000ffffffff 0000000000000008 [ 88.401929][ T5346] page dumped because: kasan: bad access detected [ 88.401935][ T5346] page_owner tracks the page as allocated [ 88.401941][ T5346] page last allocated via order 3, migratetype Unmovable, gfp_mask 0xd20c0(__GFP_IO|__GFP_FS|__GFP_NOWARN|__GFP_NORETRY|__GFP_COMP|__GFP_NOMEMALLOC), pid 9, tgid 9 (kworker/0:0), ts 83905464494, free_ts 83674944822 [ 88.401961][ T5346] post_alloc_hook+0x231/0x280 [ 88.401975][ T5346] get_page_from_freelist+0x24ba/0x2540 [ 88.401990][ T5346] __alloc_frozen_pages_noprof+0x18d/0x380 [ 88.402004][ T5346] allocate_slab+0x77/0x660 [ 88.402019][ T5346] refill_objects+0x339/0x3d0 [ 88.402033][ T5346] __pcs_replace_empty_main+0x321/0x720 [ 88.402043][ T5346] __kmalloc_node_track_caller_noprof+0x572/0x7b0 [ 88.402055][ T5346] __alloc_skb+0x2c1/0x7d0 [ 88.402067][ T5346] mld_newpack+0x14c/0xc90 [ 88.402080][ T5346] add_grhead+0x5a/0x2a0 [ 88.402093][ T5346] add_grec+0x1452/0x1740 [ 88.402105][ T5346] mld_ifc_work+0x6e6/0xe70 [ 88.402116][ T5346] process_scheduled_works+0xb5d/0x1860 [ 88.402127][ T5346] worker_thread+0xa53/0xfc0 [ 88.402138][ T5346] kthread+0x389/0x470 [ 88.402150][ T5346] ret_from_fork+0x514/0xb70 [ 88.402161][ T5346] page last free pid 5282 tgid 5282 stack trace: [ 88.402168][ T5346] __free_frozen_pages+0xbc7/0xd30 [ 88.402180][ T5346] __slab_free+0x274/0x2c0 [ 88.402191][ T5346] qlist_free_all+0x99/0x100 [ 88.402201][ T5346] kasan_quarantine_reduce+0x148/0x160 [ 88.402211][ T5346] __kasan_slab_alloc+0x22/0x80 [ 88.402221][ T5346] __kmalloc_cache_noprof+0x2ba/0x660 [ 88.402231][ T5346] kernfs_fop_open+0x3f0/0xda0 [ 88.402253][ T5346] do_dentry_open+0x785/0x14e0 [ 88.402262][ T5346] vfs_open+0x3b/0x340 [ 88.402270][ T5346] path_openat+0x2e08/0x3860 [ 88.402281][ T5346] do_file_open+0x23e/0x4a0 [ 88.402292][ T5346] do_sys_openat2+0x113/0x200 [ 88.402300][ T5346] __x64_sys_openat+0x138/0x170 [ 88.402309][ T5346] do_syscall_64+0x15f/0xf80 [ 88.402326][ T5346] entry_SYSCALL_64_after_hwframe+0x77/0x7f [ 88.402336][ T5346] [ 88.402339][ T5346] Memory state around the buggy address: [ 88.402345][ T5346] ffff888012311f00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc [ 88.402352][ T5346] ffff888012311f80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc [ 88.402359][ T5346] >ffff888012312000: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 88.402365][ T5346] ^ [ 88.402370][ T5346] ffff888012312080: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 88.402380][ T5346] ffff888012312100: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 88.402385][ T5346] ================================================================== Fix this by: 1) Making the reloc control structure ref counted; 2) Make revery place that access fs_info->reloc_ctl outside the relocation code, which at the moment it's only replace_file_extents() and btrfs_init_reloc_root(), get a reference count on the structure. There's also btrfs_update_reloc_root() that is called outside the relocation code, but this case is safe because it's only called in the transaction commit path while under the fs_info->reloc_mutex protection, but nevertheless grab a reference to make the code more consistent and avoid false alerts from AI reviews; 3) Add a spinlock to protect fs_info->reloc_ctl, since we can not take the fs_info->reloc_mutex as that would cause a deadlock since that lock is taken in the transaction commit path. That spinlock is taken before setting fs_info->reloc_ctl to an allocated structure, setting it to NULL and reading fs_info->reloc_ctl; 4) Make sure the structure is freed only when its reference count drops to zero. Reported-by: syzbot+0eea49bba18051dea35e@syzkaller.appspotmail.com Link: https://lore.kernel.org/linux-btrfs/6a1df323.bb0696ed.125a22.000a.GAE@google.com/ Reviewed-by: Qu Wenruo Signed-off-by: Filipe Manana Signed-off-by: David Sterba

btrfs: add message format for qgroupid

2026-06-09T16:22:44+00:00

The qgroupid has a specific format, add common format specifier, similar to what we have for checksums and keys. Reviewed-by: Boris Burkov Signed-off-by: David Sterba

btrfs: introduce support for huge folios

2026-06-09T10:49:26+00:00

With all the previous preparations, it's finally time to enable the huge folio support. - The max folio size Here we define BTRFS_MAX_FOLIO_SIZE, which is fixed at 2MiB. This will ensure we have a large enough but not too large folio for btrfs. This limit applies to all systems regardless of page size. Then we also define BTRFS_MAX_BLOCKS_PER_FOLIO, which depends on CONFIG_BTRFS_EXPERIMENTAL. If it's an experimental build, BTRFS_MAX_BLOCKS_PER_FOLIO is 512, otherwise it's BITS_PER_LONG. The filemap max order will be calculated using both BTRFS_MAX_FOLIO_SIZE and BTRFS_MAX_BLOCKS_PER_FOLIO. E.g. for 64K page size with 64K fs block size, the limit will be BTRFS_MAX_FOLIO_SIZE (2M), which limits the filemap max order to 5. This will be lower than the old order (6), but folios larger than 2M are rarely any better for IO performance. Meanwhile excessively large folios can cause other problems like stalling the IO pipeline for too long. For 4K page size and 4K fs block size, the limit will be increased to 2M from the old 256K. This new size is constrained by both BTRFS_MAX_FOLIO_SIZE (2M) and BTRFS_MAX_BLOCKS_PER_FOLIO (512 * 4K), allowing x86_64 to achieve huge folio support, and the filemap max order will be 9. - btrfs_bio_ctrl::submit_bitmap This will be enlarged to contain BTRFS_MAX_BLOCKS_PER_FOLIO bits, and this will be on-stack memory. This will increase on-stack memory usage by 56 bytes compared to the baseline (before the first patch in the series). - Local @delalloc_bitmap inside writepage_delalloc() Unfortunately we cannot afford to handle an allocation error here, thus again we use on-stack memory. Thus this will increase on-stack memory usage by 56 bytes again. So unfortunately this means during the delalloc window, the writeback path will have +112 bytes on-stack memory usage, and for other cases the writeback path will have +56 bytes on-stack memory usage. The +56 bytes (btrfs_bio_ctrl::submit_bitmap) can be removed after we have reworked the compression submission, so the current on-stack submit_bitmap is mostly a workaround until then. Signed-off-by: Qu Wenruo Signed-off-by: David Sterba

btrfs: remove folio ordered flag and subpage bitmap

2026-06-08T13:53:32+00:00

Btrfs has an internal flag/subpage bitmap called ordered, which is to indicate that a block has corresponding ordered extent covering it. However this requires extra synchronization between the inode ordered tree, and the folio flag/subpage bitmap, not to mention we need to maintain the extra folio flag with subpage bitmap. As a step to align btrfs_folio_state more closely to iomap_folio_state, remove the btrfs specific ordered flag/bitmap. This will also save us 64 bytes for the bitmap of a huge folio. Since we're here, also update the ASCII graph of the bitmap, as there are only 3 sub-bitmaps now, show all sub-bitmaps directly. Signed-off-by: Qu Wenruo Signed-off-by: David Sterba

btrfs: limit size of bios submitted from writeback

2026-06-08T13:53:30+00:00

Currently btrfs_writepages() just accumulates as large bio as possible (within writeback_control constraints) and then submits it. This can however lead to significant latency in writeback IO submission (I have observed tens of milliseconds) because the submitted bio easily has over hundred of megabytes. Consequently this leads to IO pipeline stalls and reduced throughput. At the same time beyond certain size submitting so large bio provides diminishing returns because the bio is split by the block layer immediately anyway. So compute (estimate of) bio size beyond which we are unlikely to improve performance and just submit the bio for writeback once we accumulate that much to keep the IO pipeline busy. This improves writeback throughput for sequential writes by about 15% on the test machine I was using. Reviewed-by: Qu Wenruo Signed-off-by: Jan Kara [ Fix the handling of missing device to avoid NULL pointer dereference. ] Signed-off-by: Qu Wenruo Signed-off-by: David Sterba

btrfs: remove 2K block size support

2026-06-08T13:53:30+00:00

Originally 2K block size support was introduced to test subpage (block size < page size) on x86_64 where the page size is exactly the original minimal block size. However that 2K block size support has some problems: - No 2K nodesize support This is critical, as there is still no way to exercise the subpage metadata routine. - Very easy to test subpage data path now With the currently experimental large folio support, it's very easy to test the subpage data folio path already, as when a folio larger than 4K is encountered on x86_64, we will need all the subpage folio states and bitmaps. So there is no need to use 2K block size just to verify subpage data path even on x86_64. And with the incoming huge folio (2M on x86_64) support, the 2K block size will easily double the bitmap size, considering the burden to maintain and the limited extra coverage, I believe it's time to remove it for the incoming huge folio support. Signed-off-by: Qu Wenruo Reviewed-by: David Sterba Signed-off-by: David Sterba

btrfs: remove op field from struct btrfs_free_space_ctl

2026-06-08T13:53:29+00:00

The op field always points to the same use_bitmap function, the only exception is during self tests where we make it temporarily point to a different function. So just because of this op pointer field we are increasing the structure size by 8 bytes. Instead of storing a pointer to a use_bitmap function in struct btrfs_free_space_ctl, move the pointer to struct btrfs_info, make insert_into_bitmap() use that pointer if we are running the self tests and initialize that pointer to the current, default use_bitmap function (now exported for the tests as btrfs_use_bitmap). This way we reduce the size of struct btrfs_free_space_ctl from 136 to 128 bytes and can now fit 32 structures in a 4K page instead of 30. This also avoids the cost of the indirection of a function pointer call when we are not running the self tests. Signed-off-by: Filipe Manana Reviewed-by: David Sterba Signed-off-by: David Sterba

btrfs: remove the COW fixup mechanism

2026-06-08T13:53:27+00:00

[BACKGROUND] Btrfs has a special mechanism called COW fixup, which detects dirty pages without an ordered extent (folio ordered flag). Normally a dirty folio must go through delayed allocation (delalloc) before it can be submitted, and delalloc will create an ordered extent for it and mark the range with ordered flag. However in older kernels, there are bugs related to get_user_pages() which can lead to some page marked dirty but without notifying the fs to properly prepare them for writeback. In that case without an ordered extent btrfs is unable to properly submit such dirty folios, thus the COW fixup mechanism is introduced, which do the extra space reservation so that they can be written back properly. [MODERN SOLUTIONS] The MM layer has solved it properly now with the introduction of pin_user_pages*(), so we're handling cases that are no longer valid. So commit 7ca3e84980ef ("btrfs: reject out-of-band dirty folios during writeback") is introduced to change the behavior from going through COW fixup to rejecting them directly for experimental builds. So far it works fine, but when errors are injected into the IO path, we have random failures triggering the new warnings. It looks like we have error path that cleared the ordered flag but leaves the folio dirty flag, which later triggers the warning. [REMOVAL OF COW FIXUP] Although I hope to fix all those known warnings cases, I just can not figure out the root cause yet. But on the other hand, if we remove the ordered and checked flags in the future, and purely rely on the dirty flags and ordered extent search, we can get a much cleaner handling. Considering it's no longer hitting the COW fixup for normal IO paths, I think it's finally the time to remove the COW fixup completely. Furthermore, the function name "btrfs_writepage_cow_fixup()" is no longer meaningful, and since it's pretty small, only a folio flag check with error message, there is no need to put it as a dedicated helper, just open code it inside extent_writepage_io(). Signed-off-by: Qu Wenruo Reviewed-by: David Sterba Signed-off-by: David Sterba