From patchwork Wed Aug 21 16:01:59 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Josef Bacik X-Patchwork-Id: 13771872 Received: from mail-yw1-f175.google.com (mail-yw1-f175.google.com [209.85.128.175]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id EB36B14C5A4 for ; Wed, 21 Aug 2024 16:02:08 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.128.175 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724256131; cv=none; b=E+yZLEfme2R4BT3Bzw9Jtspx5C6LAcawYZ5kxJuHYxQCiRWvCWavxJmbuq94Prc/WSE2t9uLEmjcqrr3e8z63EW1xoubFeZYOMahnHqKIltDAcyl9LJFamDa+Nh4cr4A9M4vBWj+DAib2MKBloQStVf/rmEI4PTsTqp1Nnu8MuM= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724256131; c=relaxed/simple; bh=ek+YnISNLCkr94REAIlDaJPta9nkySQcBN6LBQmyefk=; h=From:To:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=t7/7OpR5454yTOAVZJeuRXyBmQ4RsbJoQl/YyR+VMlNqv1g23P5mHTD4Ghf2a4LamVjkNrt2Z32m5c7Tdb1jM6l6KWK5SHaac0zPucbzvNQkCTgeJjEuVL2AvCDvXjsm0Zh0xqV8y6hcbuEqYH5CPzwXKTIBczQ3USswqMX0qjo= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=toxicpanda.com; spf=none smtp.mailfrom=toxicpanda.com; dkim=pass (2048-bit key) header.d=toxicpanda-com.20230601.gappssmtp.com header.i=@toxicpanda-com.20230601.gappssmtp.com header.b=2aZVEozO; arc=none smtp.client-ip=209.85.128.175 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=toxicpanda.com Authentication-Results: smtp.subspace.kernel.org; spf=none smtp.mailfrom=toxicpanda.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=toxicpanda-com.20230601.gappssmtp.com header.i=@toxicpanda-com.20230601.gappssmtp.com header.b="2aZVEozO" Received: by mail-yw1-f175.google.com with SMTP id 00721157ae682-69df49d92b8so63550157b3.3 for ; Wed, 21 Aug 2024 09:02:08 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toxicpanda-com.20230601.gappssmtp.com; s=20230601; t=1724256127; x=1724860927; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:from:to:cc:subject:date:message-id :reply-to; bh=WVo/r/XTSk0LReYd69ratXomFBFmGj3FiEQW0Pg38Qc=; b=2aZVEozO3ejeaIIOyxBlL29oiP6B+3KvqijtXh8uAiWGUTvVS96GYgYHmhVSVbzfad RRXtLmjIZqxIE/xn5TL2ZvzA2idREQxnQ0AgRiMcfw2EG2C8STlXqH4zC+6rWxp/u+mh IRNunB5QgxXMiE9iMtiJQrJD6nYm2zrOioFskzknhFYF88vSIDvpDcwU4zuJObx7qxZY tfK0a0Gc90aUBx7AQngDbJBPEMJhzKHQLJfojeGjrdJrr1VpISfjqGNUyf9DfRHu6Rjl b5ksjwF7gc8u+ghdU7CgNQXM0OlPvmAvJ/d9fepsdPDR5piN28/JoW7vqMUvybfekd4X WYPw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1724256127; x=1724860927; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=WVo/r/XTSk0LReYd69ratXomFBFmGj3FiEQW0Pg38Qc=; b=ARFOb5Rt/dCu/l3GD9PJjDA0qK2jDjUmeNc+zkS5eonYJuzhNQ4pC+vf9/idaaNyXA guYv2AcTx/k0U8kWttkZld9IrMIJ7xuXQLcca1fqcf1HyAXtfAwJANt5VfqCjHqmOegi 59DV0w5Z07FNIfR39dXqbg0siLo4y3B+eueQtiTeZuhBTwKjmSmd5rS4hit7NIHX1ddF qwnEGX+49ZJgxbtCLL3AqWqMZZIDqqi0jXy0JQyeU+Ffvxl9A5ThG1VI9xVxADdC9FFE lNOIDMH9Km+VH3pO4r0JAtoVg/sIh6O5rFDtCXpS1DFpIBhLMH3T9dpzD3vQTUozemFU KQ2A== X-Gm-Message-State: AOJu0YwFQVKK25oH1ZLvq7m67QAhomu655N9xuf7dsAjbCd/wp6GVqPl p4N5SuPUXq426n2MO5vSsfOsxwRFooMJEcM64GVA09OCAR6/i3trn4EZmdMTRQsYjKfMXIhOOFY m X-Google-Smtp-Source: AGHT+IH1Q+Ne/lTYbC74+eMzKsLejfXPJU/u4jMV7kbsRheHsjNdLVjtVgfZC7wI9rTZb8+LiwY+tQ== X-Received: by 2002:a05:690c:7093:b0:6b6:4d2e:1fe4 with SMTP id 00721157ae682-6c09d7e77b3mr40018447b3.19.1724256127435; Wed, 21 Aug 2024 09:02:07 -0700 (PDT) Received: from localhost (syn-076-182-020-124.res.spectrum.com. [76.182.20.124]) by smtp.gmail.com with ESMTPSA id 00721157ae682-6b9d327a582sm10804837b3.123.2024.08.21.09.02.06 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 21 Aug 2024 09:02:07 -0700 (PDT) From: Josef Bacik To: linux-btrfs@vger.kernel.org, kernel-team@fb.com Subject: [PATCH 1/3] btrfs: introduce EXTENT_DIO_LOCKED Date: Wed, 21 Aug 2024 12:01:59 -0400 Message-ID: <15f44426712f2443bf335604e78839ab715a83ed.1724255475.git.josef@toxicpanda.com> X-Mailer: git-send-email 2.43.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: linux-btrfs@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 In order to support dropping the extent lock during a read we need a way to make sure that direct reads and direct writes for overlapping ranges are protected from each other. To accomplish this introduce another lock bit specifically for direct io. Subsequent patches will utilize this to protect direct IO operations. Signed-off-by: Josef Bacik --- fs/btrfs/extent-io-tree.c | 52 ++++++++++++++++++--------------------- fs/btrfs/extent-io-tree.h | 38 +++++++++++++++++++++++++--- 2 files changed, 58 insertions(+), 32 deletions(-) diff --git a/fs/btrfs/extent-io-tree.c b/fs/btrfs/extent-io-tree.c index c54c5d7a5cd5..383e55e0f62e 100644 --- a/fs/btrfs/extent-io-tree.c +++ b/fs/btrfs/extent-io-tree.c @@ -126,7 +126,7 @@ void extent_io_tree_init(struct btrfs_fs_info *fs_info, * Empty an io tree, removing and freeing every extent state record from the * tree. This should be called once we are sure no other task can access the * tree anymore, so no tree updates happen after we empty the tree and there - * aren't any waiters on any extent state record (EXTENT_LOCKED bit is never + * aren't any waiters on any extent state record (EXTENT_LOCK_BITS are never * set on any extent state when calling this function). */ void extent_io_tree_release(struct extent_io_tree *tree) @@ -141,7 +141,7 @@ void extent_io_tree_release(struct extent_io_tree *tree) rbtree_postorder_for_each_entry_safe(state, tmp, &root, rb_node) { /* Clear node to keep free_extent_state() happy. */ RB_CLEAR_NODE(&state->rb_node); - ASSERT(!(state->state & EXTENT_LOCKED)); + ASSERT(!(state->state & EXTENT_LOCK_BITS)); /* * No need for a memory barrier here, as we are holding the tree * lock and we only change the waitqueue while holding that lock @@ -399,7 +399,7 @@ static void merge_next_state(struct extent_io_tree *tree, struct extent_state *s */ static void merge_state(struct extent_io_tree *tree, struct extent_state *state) { - if (state->state & (EXTENT_LOCKED | EXTENT_BOUNDARY)) + if (state->state & (EXTENT_LOCK_BITS | EXTENT_BOUNDARY)) return; merge_prev_state(tree, state); @@ -445,7 +445,7 @@ static struct extent_state *insert_state(struct extent_io_tree *tree, struct rb_node *parent = NULL; const u64 start = state->start - 1; const u64 end = state->end + 1; - const bool try_merge = !(bits & (EXTENT_LOCKED | EXTENT_BOUNDARY)); + const bool try_merge = !(bits & (EXTENT_LOCK_BITS | EXTENT_BOUNDARY)); set_state_bits(tree, state, bits, changeset); @@ -616,9 +616,6 @@ static void set_gfp_mask_from_bits(u32 *bits, gfp_t *mask) * inserting elements in the tree, so the gfp mask is used to indicate which * allocations or sleeping are allowed. * - * Pass 'wake' == 1 to kick any sleepers, and 'delete' == 1 to remove the given - * range from the tree regardless of state (ie for truncate). - * * The range [start, end] is inclusive. * * This takes the tree lock, and returns 0 on success and < 0 on error. @@ -647,8 +644,8 @@ int __clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, if (bits & EXTENT_DELALLOC) bits |= EXTENT_NORESERVE; - wake = (bits & EXTENT_LOCKED) ? 1 : 0; - if (bits & (EXTENT_LOCKED | EXTENT_BOUNDARY)) + wake = (bits & EXTENT_LOCK_BITS) ? 1 : 0; + if (bits & (EXTENT_LOCK_BITS | EXTENT_BOUNDARY)) clear = 1; again: if (!prealloc) { @@ -862,7 +859,7 @@ static void cache_state(struct extent_state *state, struct extent_state **cached_ptr) { return cache_state_if_flags(state, cached_ptr, - EXTENT_LOCKED | EXTENT_BOUNDARY); + EXTENT_LOCK_BITS | EXTENT_BOUNDARY); } /* @@ -1063,7 +1060,7 @@ static int __set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int ret = 0; u64 last_start; u64 last_end; - u32 exclusive_bits = (bits & EXTENT_LOCKED); + u32 exclusive_bits = (bits & EXTENT_LOCK_BITS); gfp_t mask; set_gfp_mask_from_bits(&bits, &mask); @@ -1812,12 +1809,11 @@ int set_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, u32 bits, struct extent_changeset *changeset) { /* - * We don't support EXTENT_LOCKED yet, as current changeset will - * record any bits changed, so for EXTENT_LOCKED case, it will - * either fail with -EEXIST or changeset will record the whole - * range. + * We don't support EXTENT_LOCK_BITS yet, as current changeset will + * record any bits changed, so for EXTENT_LOCK_BITS case, it will either + * fail with -EEXIST or changeset will record the whole range. */ - ASSERT(!(bits & EXTENT_LOCKED)); + ASSERT(!(bits & EXTENT_LOCK_BITS)); return __set_extent_bit(tree, start, end, bits, NULL, NULL, NULL, changeset); } @@ -1826,26 +1822,26 @@ int clear_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, u32 bits, struct extent_changeset *changeset) { /* - * Don't support EXTENT_LOCKED case, same reason as + * Don't support EXTENT_LOCK_BITS case, same reason as * set_record_extent_bits(). */ - ASSERT(!(bits & EXTENT_LOCKED)); + ASSERT(!(bits & EXTENT_LOCK_BITS)); return __clear_extent_bit(tree, start, end, bits, NULL, changeset); } -int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, - struct extent_state **cached) +bool __try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, + u32 bits, struct extent_state **cached) { int err; u64 failed_start; - err = __set_extent_bit(tree, start, end, EXTENT_LOCKED, &failed_start, + err = __set_extent_bit(tree, start, end, bits, &failed_start, NULL, cached, NULL); if (err == -EEXIST) { if (failed_start > start) clear_extent_bit(tree, start, failed_start - 1, - EXTENT_LOCKED, cached); + bits, cached); return 0; } return 1; @@ -1855,23 +1851,23 @@ int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, * Either insert or lock state struct between start and end use mask to tell * us if waiting is desired. */ -int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, - struct extent_state **cached_state) +int __lock_extent(struct extent_io_tree *tree, u64 start, u64 end, u32 bits, + struct extent_state **cached_state) { struct extent_state *failed_state = NULL; int err; u64 failed_start; - err = __set_extent_bit(tree, start, end, EXTENT_LOCKED, &failed_start, + err = __set_extent_bit(tree, start, end, bits, &failed_start, &failed_state, cached_state, NULL); while (err == -EEXIST) { if (failed_start != start) clear_extent_bit(tree, start, failed_start - 1, - EXTENT_LOCKED, cached_state); + bits, cached_state); - wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED, + wait_extent_bit(tree, failed_start, end, bits, &failed_state); - err = __set_extent_bit(tree, start, end, EXTENT_LOCKED, + err = __set_extent_bit(tree, start, end, bits, &failed_start, &failed_state, cached_state, NULL); } diff --git a/fs/btrfs/extent-io-tree.h b/fs/btrfs/extent-io-tree.h index 9d3a52d8f59a..def953f441f9 100644 --- a/fs/btrfs/extent-io-tree.h +++ b/fs/btrfs/extent-io-tree.h @@ -19,6 +19,7 @@ enum { ENUM_BIT(EXTENT_DIRTY), ENUM_BIT(EXTENT_UPTODATE), ENUM_BIT(EXTENT_LOCKED), + ENUM_BIT(EXTENT_DIO_LOCKED), ENUM_BIT(EXTENT_NEW), ENUM_BIT(EXTENT_DELALLOC), ENUM_BIT(EXTENT_DEFRAG), @@ -67,6 +68,8 @@ enum { EXTENT_ADD_INODE_BYTES | \ EXTENT_CLEAR_ALL_BITS) +#define EXTENT_LOCK_BITS (EXTENT_LOCKED | EXTENT_DIO_LOCKED) + /* * Redefined bits above which are used only in the device allocation tree, * shouldn't be using EXTENT_LOCKED / EXTENT_BOUNDARY / EXTENT_CLEAR_META_RESV @@ -134,12 +137,22 @@ const struct btrfs_fs_info *extent_io_tree_to_fs_info(const struct extent_io_tre void extent_io_tree_init(struct btrfs_fs_info *fs_info, struct extent_io_tree *tree, unsigned int owner); void extent_io_tree_release(struct extent_io_tree *tree); +int __lock_extent(struct extent_io_tree *tree, u64 start, u64 end, + u32 bits, struct extent_state **cached); +bool __try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, + u32 bits, struct extent_state **cached); -int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, - struct extent_state **cached); +static inline int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, + struct extent_state **cached) +{ + return __lock_extent(tree, start, end, EXTENT_LOCKED, cached); +} -int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, - struct extent_state **cached); +static inline bool try_lock_extent(struct extent_io_tree *tree, u64 start, + u64 end, struct extent_state **cached) +{ + return __try_lock_extent(tree, start, end, EXTENT_LOCKED, cached); +} int __init extent_state_init_cachep(void); void __cold extent_state_free_cachep(void); @@ -212,5 +225,22 @@ int find_contiguous_extent_bit(struct extent_io_tree *tree, u64 start, bool btrfs_find_delalloc_range(struct extent_io_tree *tree, u64 *start, u64 *end, u64 max_bytes, struct extent_state **cached_state); +static inline int lock_dio_extent(struct extent_io_tree *tree, u64 start, + u64 end, struct extent_state **cached) +{ + return __lock_extent(tree, start, end, EXTENT_DIO_LOCKED, cached); +} + +static inline bool try_lock_dio_extent(struct extent_io_tree *tree, u64 start, + u64 end, struct extent_state **cached) +{ + return __try_lock_extent(tree, start, end, EXTENT_DIO_LOCKED, cached); +} + +static inline int unlock_dio_extent(struct extent_io_tree *tree, u64 start, + u64 end, struct extent_state **cached) +{ + return __clear_extent_bit(tree, start, end, EXTENT_DIO_LOCKED, cached, NULL); +} #endif /* BTRFS_EXTENT_IO_TREE_H */ From patchwork Wed Aug 21 16:02:00 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Josef Bacik X-Patchwork-Id: 13771873 Received: from mail-ot1-f53.google.com (mail-ot1-f53.google.com [209.85.210.53]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C3F4714EC64 for ; Wed, 21 Aug 2024 16:02:09 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.210.53 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724256132; cv=none; b=T6644J/wTdKQlw2MKDz+MEXRU5tM2usOZPlE5qc243qZ8cP3lJjdRd0+5A+SeShj/GiQbQegsZ1B5sc10JCJ225IQxcOu9Nr4Uiw81F8P/AbGL7iBceALzIsGVQ/rA0tepBcPNQML8YzgBvkSJuJEKm6OyugbcC0925LeWAhhpI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724256132; c=relaxed/simple; bh=6sYAZUCmjpxoqIQU0tHk/Q9iiNTZFAHgAOzHJ4Fpm2U=; h=From:To:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=JSyDp9yVcKjU0bXFz47rElu1wVJcEc8XiFb6ehICAVnqcTR4Jb8ox1zR0VLdN/wNtE5+8p0Xi3EF7O7MtkVyjpTt8DEnHBqfz8CYdejNfalWANlHEALHEQga8wZUciTBD5yDdVlK/j1G+t+KSo6b10TYiP7zfKNW2VOS8adlJTo= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=toxicpanda.com; spf=none smtp.mailfrom=toxicpanda.com; dkim=pass (2048-bit key) header.d=toxicpanda-com.20230601.gappssmtp.com header.i=@toxicpanda-com.20230601.gappssmtp.com header.b=QKBo5gwC; arc=none smtp.client-ip=209.85.210.53 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=toxicpanda.com Authentication-Results: smtp.subspace.kernel.org; spf=none smtp.mailfrom=toxicpanda.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=toxicpanda-com.20230601.gappssmtp.com header.i=@toxicpanda-com.20230601.gappssmtp.com header.b="QKBo5gwC" Received: by mail-ot1-f53.google.com with SMTP id 46e09a7af769-709428a9469so4326623a34.3 for ; Wed, 21 Aug 2024 09:02:09 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toxicpanda-com.20230601.gappssmtp.com; s=20230601; t=1724256129; x=1724860929; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:from:to:cc:subject:date:message-id :reply-to; bh=Bm7qw22RjMkP1zkkz1J+rH69AVj2mfYOeGv3FTOwVz8=; b=QKBo5gwCS5T/o8onWJf4Z4thKlpBYg3k5wZnh3hP+GAwF3FAMlSwwJ8+cOXPxdV13v MfJBS238+owaCR/NoOv5zxsOOBd9aBFTJTKVvc6bTXaGhkJ4aqSq0k2IOh5rn9+JivGV aW/FEDsGsBpZyZMLWecfDszLIZ1pXfohNxYxQuhJwUbUMU90FMELCcaQ2yF8OZRXJGlZ AB4Z1UqsuXSXnyBX8DJaHPPEMz/qtRKtn3RGOVESZzK+PQriz2BsGoUeyPiYrbrTu1uK pRaEWm9nIczhOa1f1HqYenGsN0xdRs4XMTZYfbF/NsN70zm7v2RVvRmLrI2lTh81Sx6k Ck1w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1724256129; x=1724860929; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=Bm7qw22RjMkP1zkkz1J+rH69AVj2mfYOeGv3FTOwVz8=; b=Ni+/9auYIVepP9rc0SQIqH5o9zooBCcdeh9wioop12pg88RSeie6S2QqQfSxCUxwr9 LOm0hyIK4GYHl/mWTXww7Gc+SAPexSfOufs/lrr+Zl/+V2ZIkvreiwDElM0P2e83f8i/ RKMefaKJQnDOkyCeXTYumXO4XMGOPn8e3efMzg6y/VK2mILsahl7hmnroTIaM4s2ITvN H2Hdw4xmLrycUoEtJnYaZma4Q/tFCfMJyZi2h2U2S0bxLDqMW9PQESv5XUv8LprDpiVy gTv/FicqaIYoiUpsR4kmbExVyei0sxa6cLwtdJe/8ZcsfkmXYBhi7Jrr2A0p6bAwYj1z Ynpw== X-Gm-Message-State: AOJu0YyNxlvkrl8HVbMrl24dhzTbM1dDEsJxKe7LqDJGS/QB5OdCkbUF DsgTXHlyJsLXw++sbX/suAHp50X4LItlPyRuaj/ylMK32rt2XkL4uU29UfbcSGWJjN9j3rfniha 1 X-Google-Smtp-Source: AGHT+IHydACgd70cMX4uDt4PS/BbZMEoKFkZ7XsAH54/fOy7sgSjbWypBVJxcel+iwLiA5eMZLyvXA== X-Received: by 2002:a05:6830:698b:b0:703:79f5:fabb with SMTP id 46e09a7af769-70df8700014mr3230157a34.11.1724256128476; Wed, 21 Aug 2024 09:02:08 -0700 (PDT) Received: from localhost (syn-076-182-020-124.res.spectrum.com. [76.182.20.124]) by smtp.gmail.com with ESMTPSA id 00721157ae682-6af9d823bbesm22423257b3.106.2024.08.21.09.02.08 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 21 Aug 2024 09:02:08 -0700 (PDT) From: Josef Bacik To: linux-btrfs@vger.kernel.org, kernel-team@fb.com Subject: [PATCH 2/3] btrfs: take the dio extent lock during O_DIRECT operations Date: Wed, 21 Aug 2024 12:02:00 -0400 Message-ID: X-Mailer: git-send-email 2.43.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: linux-btrfs@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Currently we hold the extent lock for the entire duration of a read. This isn't really necessary in the buffered case, we're protected by the page lock, however it's necessary for O_DIRECT. For O_DIRECT reads, if we only locked the extent for the part where we get the extent, we could potentially race with an O_DIRECT write in the same region. This isn't really a problem, unless the read is delayed so much that the write does the CoW, unpins the old extent, and some other application re-allocates the extent before the read is actually able to be submitted. At that point at best we'd have a csum mismatch, but at worse we could read data that doesn't belong to us. To address this potential race we need to make sure we don't have overlapping, concurrent direct io reads and writes. To accomplish this use the new EXTENT_DIO_LOCKED bit in the direct IO case in the same spot as the current extent lock. The writes will take this while they're creating the ordered extent, which is also used to make sure concurrent buffered reads or concurrent direct reads are not allowed to occur, and drop it after the ordered extent is taken. For reads it will act as the current read behavior for the EXTENT_LOCKED bit, we set it when we're starting the read, we clear it in the end_io to allow other direct writes to continue. This still has the drawback of disallowing concurrent overlapping direct reads from occurring, but that exists with the current extent locking. Signed-off-by: Josef Bacik --- fs/btrfs/direct-io.c | 47 +++++++++++++++++++++++++++++++++----------- 1 file changed, 35 insertions(+), 12 deletions(-) diff --git a/fs/btrfs/direct-io.c b/fs/btrfs/direct-io.c index 67adbe9d294a..576f469cacee 100644 --- a/fs/btrfs/direct-io.c +++ b/fs/btrfs/direct-io.c @@ -40,11 +40,22 @@ static int lock_extent_direct(struct inode *inode, u64 lockstart, u64 lockend, struct btrfs_ordered_extent *ordered; int ret = 0; + /* Direct lock must be taken before the extent lock. */ + if (nowait) { + if (!try_lock_dio_extent(io_tree, lockstart, lockend, + cached_state)) + return -EAGAIN; + } else { + lock_dio_extent(io_tree, lockstart, lockend, cached_state); + } + while (1) { if (nowait) { if (!try_lock_extent(io_tree, lockstart, lockend, - cached_state)) - return -EAGAIN; + cached_state)) { + ret = -EAGAIN; + break; + } } else { lock_extent(io_tree, lockstart, lockend, cached_state); } @@ -120,6 +131,8 @@ static int lock_extent_direct(struct inode *inode, u64 lockstart, u64 lockend, cond_resched(); } + if (ret) + unlock_dio_extent(io_tree, lockstart, lockend, cached_state); return ret; } @@ -546,8 +559,9 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start, } if (unlock_extents) - unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend, - &cached_state); + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, + EXTENT_LOCKED | EXTENT_DIO_LOCKED, + &cached_state); else free_extent_state(cached_state); @@ -572,8 +586,13 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start, return 0; unlock_err: - unlock_extent(&BTRFS_I(inode)->io_tree, lockstart, lockend, - &cached_state); + /* + * Don't use EXTENT_LOCK_BITS here in case we extend it later and forget + * to update this, be explicit that we expect EXTENT_LOCKED and + * EXTENT_DIO_LOCKED to be set here, and so that's what we're clearing. + */ + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, + EXTENT_LOCKED | EXTENT_DIO_LOCKED, &cached_state); err: if (dio_data->data_space_reserved) { btrfs_free_reserved_data_space(BTRFS_I(inode), @@ -596,8 +615,9 @@ static int btrfs_dio_iomap_end(struct inode *inode, loff_t pos, loff_t length, if (!write && (iomap->type == IOMAP_HOLE)) { /* If reading from a hole, unlock and return */ - unlock_extent(&BTRFS_I(inode)->io_tree, pos, pos + length - 1, - NULL); + clear_extent_bit(&BTRFS_I(inode)->io_tree, pos, + pos + length - 1, + EXTENT_LOCKED | EXTENT_DIO_LOCKED, NULL); return 0; } @@ -608,8 +628,10 @@ static int btrfs_dio_iomap_end(struct inode *inode, loff_t pos, loff_t length, btrfs_finish_ordered_extent(dio_data->ordered, NULL, pos, length, false); else - unlock_extent(&BTRFS_I(inode)->io_tree, pos, - pos + length - 1, NULL); + clear_extent_bit(&BTRFS_I(inode)->io_tree, pos, + pos + length - 1, + EXTENT_LOCKED | EXTENT_DIO_LOCKED, + NULL); ret = -ENOTBLK; } if (write) { @@ -641,8 +663,9 @@ static void btrfs_dio_end_io(struct btrfs_bio *bbio) dip->file_offset, dip->bytes, !bio->bi_status); } else { - unlock_extent(&inode->io_tree, dip->file_offset, - dip->file_offset + dip->bytes - 1, NULL); + clear_extent_bit(&inode->io_tree, dip->file_offset, + dip->file_offset + dip->bytes - 1, + EXTENT_LOCKED | EXTENT_DIO_LOCKED, NULL); } bbio->bio.bi_private = bbio->private; From patchwork Wed Aug 21 16:02:01 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Josef Bacik X-Patchwork-Id: 13771874 Received: from mail-yw1-f170.google.com (mail-yw1-f170.google.com [209.85.128.170]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 372B314F131 for ; Wed, 21 Aug 2024 16:02:11 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.128.170 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724256133; cv=none; b=to/xYxs3xRmkWpZXEKQVVmAXKkVfkDGiyYXk3Ava259xECDlbQmF4CLOg8IRk0iX+nRUpxYO1Uoj4EEqSz+RMHfjv9SDVgNE10tyNfgT48+aEMN9x1D7wi24kdjxOnXh7GL0pv7QkXDP3BNSwlnlXabvVAUdBHBvR1vsswERlyA= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724256133; c=relaxed/simple; bh=rR9ypka7eqGS5zoDgzgixATAiy7VyJ17s1AMcGO2cHM=; h=From:To:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=EvjQC8cyD300RF/G2gILXueDfsZJJFlkLltneNTESgXz27GMtcxRHBNWgI8rZQI2Sj1Fkax/GFEaLn/hsiLdRntr+UWLUsq3sjcEpEx704auchNSlurE0/1XT/I/eeYqnfqiszFg44xzCziOnh6o0Iio39ogUJtn9622s7ARoGk= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=toxicpanda.com; spf=none smtp.mailfrom=toxicpanda.com; dkim=pass (2048-bit key) header.d=toxicpanda-com.20230601.gappssmtp.com header.i=@toxicpanda-com.20230601.gappssmtp.com header.b=G2/it2Ul; arc=none smtp.client-ip=209.85.128.170 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=toxicpanda.com Authentication-Results: smtp.subspace.kernel.org; spf=none smtp.mailfrom=toxicpanda.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=toxicpanda-com.20230601.gappssmtp.com header.i=@toxicpanda-com.20230601.gappssmtp.com header.b="G2/it2Ul" Received: by mail-yw1-f170.google.com with SMTP id 00721157ae682-6aab656687cso8229417b3.1 for ; Wed, 21 Aug 2024 09:02:11 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toxicpanda-com.20230601.gappssmtp.com; s=20230601; t=1724256130; x=1724860930; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:from:to:cc:subject:date:message-id :reply-to; bh=hGa73gjIoW7xEA9h9DWBm/Q9njXwKtq+rS04x/d+vGA=; b=G2/it2Ulp/e2FsIp96XP6NMIRuGQH59yK+LSNL8SR+SDLENy39iIvf9wcnmXXaWWY7 BqrRm+2G8sOCyxi9/CJ0VQM6/WqdyJVxBDR08R+XINm0aylVRDGrCsHJKkzSDmQPxK6M /OMmL/Lki0g6eTxF/XdaL2IMR0KqcwOTndAULgrEbmqwtaP3oR61uh/Nm9q05bYYAbpT XZSt3/SY0ybEGvNbKMTUWXmWefyLb3pBeg1NM32g4NQpLiCpOEg4ra40+I6iYpALYww1 LvMu/wX/hFjr3bVxYcXY8Zkk6kFWSw+PfDmllghG1Je7k5mr+oBU8AWng8AqPtP8SqAY z3gw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1724256130; x=1724860930; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=hGa73gjIoW7xEA9h9DWBm/Q9njXwKtq+rS04x/d+vGA=; b=CwPkYplcKVWyqKmOJIHWbeBHEeAB1fSUed5dpRaDxOZ+PlyrLQsnm7Az7OGiom2mGN IAZd6dPFDyTOzhC1tGP5Qq3I3sjSekC5b8VFdIn1RXtbr2M3BgYtpvZH246PePtFjsbB RJjqfxr0Et8GQyi6OI6/O7NJzG3UCKevyu23+eP1QLDXdh7OWiQDa7oB8UPqVts+rG8A Qn1/JxxhzGQ+43+0itlhsy+BluY/j3lfjkidVO0RBxhRJOk6nTX3Htw4L3oz43KUxNPf jNJ2oZRMF/TDAr/KOJbBgxAmZN5Wg4w+jIUBmLhX8hJC5HwfysarwdZK8LL9Paa3djfr ygTQ== X-Gm-Message-State: AOJu0Yx2QhnyjGlGpMi4RZmlTZxA6gf4am63usgIQLxB/cfSgo1UFPXt QZmYZLPYxKBVqFuQjXh2W5O+DO+vGi32jALC3JQK1U7UyS36qF1wTrINbPNIh1sJRHlDZci7Pwb n X-Google-Smtp-Source: AGHT+IGj8jmLhVWrvtusPFgRfQgUo/1L0Fy7xvW7g5Oco+QL4kHNcbZ0evq7UkgJgsZ4PjyCge5GQA== X-Received: by 2002:a05:690c:5009:b0:6b7:f467:e0f5 with SMTP id 00721157ae682-6c3041c4029mr1131517b3.9.1724256129736; Wed, 21 Aug 2024 09:02:09 -0700 (PDT) Received: from localhost (syn-076-182-020-124.res.spectrum.com. [76.182.20.124]) by smtp.gmail.com with ESMTPSA id 00721157ae682-6af99f9e012sm22769797b3.41.2024.08.21.09.02.09 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 21 Aug 2024 09:02:09 -0700 (PDT) From: Josef Bacik To: linux-btrfs@vger.kernel.org, kernel-team@fb.com Subject: [PATCH 3/3] btrfs: do not hold the extent lock for entire read Date: Wed, 21 Aug 2024 12:02:01 -0400 Message-ID: <174b23dd8cf1a03d54f5ca3dc03c3f5011988d8a.1724255475.git.josef@toxicpanda.com> X-Mailer: git-send-email 2.43.0 In-Reply-To: References: Precedence: bulk X-Mailing-List: linux-btrfs@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Historically we've held the extent lock throughout the entire read. There's been a few reasons for this, but it's mostly just caused us problems. For example, this prevents us from allowing page faults during direct io reads, because we could deadlock. This has forced us to only allow 4k reads at a time for io_uring NOWAIT requests because we have no idea if we'll be forced to page fault and thus have to do a whole lot of work. On the buffered side we are protected by the page lock, as long as we're reading things like buffered writes, punch hole, and even direct IO to a certain degree will get hung up on the page lock while the page is in flight. On the direct side we have the dio extent lock, which acts much like the way the extent lock worked previously to this patch, however just for direct reads. This protects direct reads from concurrent direct writes, while we're protected from buffered writes via the inode lock. Now that we're protected in all cases, narrow the extent lock to the part where we're getting the extent map to submit the reads, no longer holding the extent lock for the entire read operation. Signed-off-by: Josef Bacik --- fs/btrfs/compression.c | 2 +- fs/btrfs/direct-io.c | 51 ++++++++++++------------ fs/btrfs/extent_io.c | 88 +++--------------------------------------- 3 files changed, 31 insertions(+), 110 deletions(-) diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 832ab8984c41..511f81f312af 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -521,6 +521,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, } add_size = min(em->start + em->len, page_end + 1) - cur; free_extent_map(em); + unlock_extent(tree, cur, page_end, NULL); if (folio->index == end_index) { size_t zero_offset = offset_in_folio(folio, isize); @@ -534,7 +535,6 @@ static noinline int add_ra_bio_pages(struct inode *inode, if (!bio_add_folio(orig_bio, folio, add_size, offset_in_folio(folio, cur))) { - unlock_extent(tree, cur, page_end, NULL); folio_unlock(folio); folio_put(folio); break; diff --git a/fs/btrfs/direct-io.c b/fs/btrfs/direct-io.c index 576f469cacee..66f1ce5fcfd2 100644 --- a/fs/btrfs/direct-io.c +++ b/fs/btrfs/direct-io.c @@ -366,7 +366,7 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start, int ret = 0; u64 len = length; const u64 data_alloc_len = length; - bool unlock_extents = false; + u32 unlock_bits = EXTENT_LOCKED; /* * We could potentially fault if we have a buffer > PAGE_SIZE, and if @@ -527,7 +527,6 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start, start, &len, flags); if (ret < 0) goto unlock_err; - unlock_extents = true; /* Recalc len in case the new em is smaller than requested */ len = min(len, em->len - (start - em->start)); if (dio_data->data_space_reserved) { @@ -548,23 +547,8 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start, release_offset, release_len); } - } else { - /* - * We need to unlock only the end area that we aren't using. - * The rest is going to be unlocked by the endio routine. - */ - lockstart = start + len; - if (lockstart < lockend) - unlock_extents = true; } - if (unlock_extents) - clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, - EXTENT_LOCKED | EXTENT_DIO_LOCKED, - &cached_state); - else - free_extent_state(cached_state); - /* * Translate extent map information to iomap. * We trim the extents (and move the addr) even though iomap code does @@ -583,6 +567,23 @@ static int btrfs_dio_iomap_begin(struct inode *inode, loff_t start, iomap->length = len; free_extent_map(em); + /* + * Reads will hold the EXTENT_DIO_LOCKED bit until the io is completed, + * writes only hold it for this part. We hold the extent lock until + * we're completely done with the extent map to make sure it remains + * valid. + */ + if (write) + unlock_bits |= EXTENT_DIO_LOCKED; + + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, + unlock_bits, &cached_state); + + /* We didn't use everything, unlock the dio extent for the remainder. */ + if (!write && (start + len) < lockend) + unlock_dio_extent(&BTRFS_I(inode)->io_tree, start + len, + lockend, NULL); + return 0; unlock_err: @@ -615,9 +616,8 @@ static int btrfs_dio_iomap_end(struct inode *inode, loff_t pos, loff_t length, if (!write && (iomap->type == IOMAP_HOLE)) { /* If reading from a hole, unlock and return */ - clear_extent_bit(&BTRFS_I(inode)->io_tree, pos, - pos + length - 1, - EXTENT_LOCKED | EXTENT_DIO_LOCKED, NULL); + unlock_dio_extent(&BTRFS_I(inode)->io_tree, pos, + pos + length - 1, NULL); return 0; } @@ -628,10 +628,8 @@ static int btrfs_dio_iomap_end(struct inode *inode, loff_t pos, loff_t length, btrfs_finish_ordered_extent(dio_data->ordered, NULL, pos, length, false); else - clear_extent_bit(&BTRFS_I(inode)->io_tree, pos, - pos + length - 1, - EXTENT_LOCKED | EXTENT_DIO_LOCKED, - NULL); + unlock_dio_extent(&BTRFS_I(inode)->io_tree, pos, + pos + length - 1, NULL); ret = -ENOTBLK; } if (write) { @@ -663,9 +661,8 @@ static void btrfs_dio_end_io(struct btrfs_bio *bbio) dip->file_offset, dip->bytes, !bio->bi_status); } else { - clear_extent_bit(&inode->io_tree, dip->file_offset, - dip->file_offset + dip->bytes - 1, - EXTENT_LOCKED | EXTENT_DIO_LOCKED, NULL); + unlock_dio_extent(&inode->io_tree, dip->file_offset, + dip->file_offset + dip->bytes - 1, NULL); } bbio->bio.bi_private = bbio->private; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 822e2bf8bc99..3f3baec23717 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -480,75 +480,6 @@ static void end_bbio_data_write(struct btrfs_bio *bbio) bio_put(bio); } -/* - * Record previously processed extent range - * - * For endio_readpage_release_extent() to handle a full extent range, reducing - * the extent io operations. - */ -struct processed_extent { - struct btrfs_inode *inode; - /* Start of the range in @inode */ - u64 start; - /* End of the range in @inode */ - u64 end; - bool uptodate; -}; - -/* - * Try to release processed extent range - * - * May not release the extent range right now if the current range is - * contiguous to processed extent. - * - * Will release processed extent when any of @inode, @uptodate, the range is - * no longer contiguous to the processed range. - * - * Passing @inode == NULL will force processed extent to be released. - */ -static void endio_readpage_release_extent(struct processed_extent *processed, - struct btrfs_inode *inode, u64 start, u64 end, - bool uptodate) -{ - struct extent_state *cached = NULL; - struct extent_io_tree *tree; - - /* The first extent, initialize @processed */ - if (!processed->inode) - goto update; - - /* - * Contiguous to processed extent, just uptodate the end. - * - * Several things to notice: - * - * - bio can be merged as long as on-disk bytenr is contiguous - * This means we can have page belonging to other inodes, thus need to - * check if the inode still matches. - * - bvec can contain range beyond current page for multi-page bvec - * Thus we need to do processed->end + 1 >= start check - */ - if (processed->inode == inode && processed->uptodate == uptodate && - processed->end + 1 >= start && end >= processed->end) { - processed->end = end; - return; - } - - tree = &processed->inode->io_tree; - /* - * Now we don't have range contiguous to the processed range, release - * the processed range now. - */ - unlock_extent(tree, processed->start, processed->end, &cached); - -update: - /* Update processed to current range */ - processed->inode = inode; - processed->start = start; - processed->end = end; - processed->uptodate = uptodate; -} - static void begin_folio_read(struct btrfs_fs_info *fs_info, struct folio *folio) { ASSERT(folio_test_locked(folio)); @@ -575,7 +506,6 @@ static void end_bbio_data_read(struct btrfs_bio *bbio) { struct btrfs_fs_info *fs_info = bbio->fs_info; struct bio *bio = &bbio->bio; - struct processed_extent processed = { 0 }; struct folio_iter fi; const u32 sectorsize = fs_info->sectorsize; @@ -640,11 +570,7 @@ static void end_bbio_data_read(struct btrfs_bio *bbio) /* Update page status and unlock. */ end_folio_read(folio, uptodate, start, len); - endio_readpage_release_extent(&processed, BTRFS_I(inode), - start, end, uptodate); } - /* Release the last extent */ - endio_readpage_release_extent(&processed, NULL, 0, 0, false); bio_put(bio); } @@ -1019,11 +945,9 @@ static int btrfs_do_readpage(struct folio *folio, struct extent_map **em_cached, size_t pg_offset = 0; size_t iosize; size_t blocksize = fs_info->sectorsize; - struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree; ret = set_folio_extent_mapped(folio); if (ret < 0) { - unlock_extent(tree, start, end, NULL); folio_unlock(folio); return ret; } @@ -1047,14 +971,12 @@ static int btrfs_do_readpage(struct folio *folio, struct extent_map **em_cached, if (cur >= last_byte) { iosize = folio_size(folio) - pg_offset; folio_zero_range(folio, pg_offset, iosize); - unlock_extent(tree, cur, cur + iosize - 1, NULL); end_folio_read(folio, true, cur, iosize); break; } em = __get_extent_map(inode, folio, cur, end - cur + 1, em_cached); if (IS_ERR(em)) { - unlock_extent(tree, cur, end, NULL); end_folio_read(folio, false, cur, end + 1 - cur); return PTR_ERR(em); } @@ -1123,7 +1045,6 @@ static int btrfs_do_readpage(struct folio *folio, struct extent_map **em_cached, if (block_start == EXTENT_MAP_HOLE) { folio_zero_range(folio, pg_offset, iosize); - unlock_extent(tree, cur, cur + iosize - 1, NULL); end_folio_read(folio, true, cur, iosize); cur = cur + iosize; pg_offset += iosize; @@ -1131,7 +1052,6 @@ static int btrfs_do_readpage(struct folio *folio, struct extent_map **em_cached, } /* the get_extent function already copied into the folio */ if (block_start == EXTENT_MAP_INLINE) { - unlock_extent(tree, cur, cur + iosize - 1, NULL); end_folio_read(folio, true, cur, iosize); cur = cur + iosize; pg_offset += iosize; @@ -1161,11 +1081,13 @@ int btrfs_read_folio(struct file *file, struct folio *folio) u64 end = start + folio_size(folio) - 1; struct btrfs_bio_ctrl bio_ctrl = { .opf = REQ_OP_READ }; struct extent_map *em_cached = NULL; + struct extent_state *cached = NULL; int ret; - btrfs_lock_and_flush_ordered_range(inode, start, end, NULL); + btrfs_lock_and_flush_ordered_range(inode, start, end, &cached); ret = btrfs_do_readpage(folio, &em_cached, &bio_ctrl, NULL); + unlock_extent(&inode->io_tree, start, end, &cached); free_extent_map(em_cached); /* @@ -2350,13 +2272,15 @@ void btrfs_readahead(struct readahead_control *rac) u64 start = readahead_pos(rac); u64 end = start + readahead_length(rac) - 1; struct extent_map *em_cached = NULL; + struct extent_state *cached = NULL; u64 prev_em_start = (u64)-1; - btrfs_lock_and_flush_ordered_range(inode, start, end, NULL); + btrfs_lock_and_flush_ordered_range(inode, start, end, &cached);; while ((folio = readahead_folio(rac)) != NULL) btrfs_do_readpage(folio, &em_cached, &bio_ctrl, &prev_em_start); + unlock_extent(&inode->io_tree, start, end, &cached); if (em_cached) free_extent_map(em_cached); submit_one_bio(&bio_ctrl);