linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Jens Axboe <axboe@kernel.dk>,
	Pavel Begunkov <asml.silence@gmail.com>
Subject: [PATCH 5.10 10/54] io_uring: replace inflight_wait with tctx->wait
Date: Thu, 11 Feb 2021 16:01:54 +0100	[thread overview]
Message-ID: <20210211150153.326935464@linuxfoundation.org> (raw)
In-Reply-To: <20210211150152.885701259@linuxfoundation.org>

From: Pavel Begunkov <asml.silence@gmail.com>

[ Upstream commit c98de08c990e190fc7cc3aaf8079b4a0674c6425 ]

As tasks now cancel only theirs requests, and inflight_wait is awaited
only in io_uring_cancel_files(), which should be called with ->in_idle
set, instead of keeping a separate inflight_wait use tctx->wait.

That will add some spurious wakeups but actually is safer from point of
not hanging the task.

e.g.
task1                   | IRQ
                        | *start* io_complete_rw_common(link)
                        |        link: req1 -> req2 -> req3(with files)
*cancel_files()         |
io_wq_cancel(), etc.    |
                        | put_req(link), adds to io-wq req2
schedule()              |

So, task1 will never try to cancel req2 or req3. If req2 is
long-standing (e.g. read(empty_pipe)), this may hang.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 fs/io_uring.c |   13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -286,7 +286,6 @@ struct io_ring_ctx {
 		struct list_head	timeout_list;
 		struct list_head	cq_overflow_list;
 
-		wait_queue_head_t	inflight_wait;
 		struct io_uring_sqe	*sq_sqes;
 	} ____cacheline_aligned_in_smp;
 
@@ -1220,7 +1219,6 @@ static struct io_ring_ctx *io_ring_ctx_a
 	INIT_LIST_HEAD(&ctx->iopoll_list);
 	INIT_LIST_HEAD(&ctx->defer_list);
 	INIT_LIST_HEAD(&ctx->timeout_list);
-	init_waitqueue_head(&ctx->inflight_wait);
 	spin_lock_init(&ctx->inflight_lock);
 	INIT_LIST_HEAD(&ctx->inflight_list);
 	INIT_DELAYED_WORK(&ctx->file_put_work, io_file_put_work);
@@ -5894,6 +5892,7 @@ static int io_req_defer(struct io_kiocb
 static void io_req_drop_files(struct io_kiocb *req)
 {
 	struct io_ring_ctx *ctx = req->ctx;
+	struct io_uring_task *tctx = req->task->io_uring;
 	unsigned long flags;
 
 	if (req->work.flags & IO_WQ_WORK_FILES) {
@@ -5905,8 +5904,8 @@ static void io_req_drop_files(struct io_
 	spin_unlock_irqrestore(&ctx->inflight_lock, flags);
 	req->flags &= ~REQ_F_INFLIGHT;
 	req->work.flags &= ~IO_WQ_WORK_FILES;
-	if (waitqueue_active(&ctx->inflight_wait))
-		wake_up(&ctx->inflight_wait);
+	if (atomic_read(&tctx->in_idle))
+		wake_up(&tctx->wait);
 }
 
 static void __io_clean_op(struct io_kiocb *req)
@@ -8605,8 +8604,8 @@ static void io_uring_cancel_files(struct
 			break;
 		}
 		if (found)
-			prepare_to_wait(&ctx->inflight_wait, &wait,
-						TASK_UNINTERRUPTIBLE);
+			prepare_to_wait(&task->io_uring->wait, &wait,
+					TASK_UNINTERRUPTIBLE);
 		spin_unlock_irq(&ctx->inflight_lock);
 
 		/* We need to keep going until we don't find a matching req */
@@ -8619,7 +8618,7 @@ static void io_uring_cancel_files(struct
 		/* cancellations _may_ trigger task work */
 		io_run_task_work();
 		schedule();
-		finish_wait(&ctx->inflight_wait, &wait);
+		finish_wait(&task->io_uring->wait, &wait);
 	}
 }
 



  parent reply	other threads:[~2021-02-11 15:31 UTC|newest]

Thread overview: 68+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-11 15:01 [PATCH 5.10 00/54] 5.10.16-rc1 review Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 01/54] io_uring: simplify io_task_match() Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 02/54] io_uring: add a {task,files} pair matching helper Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 03/54] io_uring: dont iterate io_uring_cancel_files() Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 04/54] io_uring: pass files into kill timeouts/poll Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 05/54] io_uring: always batch cancel in *cancel_files() Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 06/54] io_uring: fix files cancellation Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 07/54] io_uring: account io_uring internal files as REQ_F_INFLIGHT Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 08/54] io_uring: if we see flush on exit, cancel related tasks Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 09/54] io_uring: fix __io_uring_files_cancel() with TASK_UNINTERRUPTIBLE Greg Kroah-Hartman
2021-02-11 15:01 ` Greg Kroah-Hartman [this message]
2021-02-11 15:01 ` [PATCH 5.10 11/54] io_uring: fix cancellation taking mutex while TASK_UNINTERRUPTIBLE Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 12/54] io_uring: fix flush cqring overflow list while TASK_INTERRUPTIBLE Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 13/54] io_uring: fix list corruption for splice file_get Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 14/54] io_uring: fix sqo ownership false positive warning Greg Kroah-Hartman
2021-02-11 15:01 ` [PATCH 5.10 15/54] io_uring: reinforce cancel on flush during exit Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 16/54] io_uring: drop mm/files between task_work_submit Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 17/54] gpiolib: cdev: clear debounce period if line set to output Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 18/54] powerpc/64/signal: Fix regression in __kernel_sigtramp_rt64() semantics Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 19/54] af_key: relax availability checks for skb size calculation Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 20/54] regulator: core: avoid regulator_resolve_supply() race condition Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 21/54] ASoC: wm_adsp: Fix control name parsing for multi-fw Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 22/54] drm/nouveau/nvif: fix method count when pushing an array Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 23/54] mac80211: 160MHz with extended NSS BW in CSA Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 24/54] ASoC: Intel: Skylake: Zero snd_ctl_elem_value Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 25/54] chtls: Fix potential resource leak Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 26/54] pNFS/NFSv4: Try to return invalid layout in pnfs_layout_process() Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 27/54] pNFS/NFSv4: Improve rejection of out-of-order layouts Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 28/54] ALSA: hda: intel-dsp-config: add PCI id for TGL-H Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 29/54] ASoC: ak4458: correct reset polarity Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 30/54] ASoC: Intel: sof_sdw: set proper flags for Dell TGL-H SKU 0A5E Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 31/54] iwlwifi: mvm: skip power command when unbinding vif during CSA Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 32/54] iwlwifi: mvm: take mutex for calling iwl_mvm_get_sync_time() Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 33/54] iwlwifi: pcie: add a NULL check in iwl_pcie_txq_unmap Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 34/54] iwlwifi: pcie: fix context info memory leak Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 35/54] iwlwifi: mvm: invalidate IDs of internal stations at mvm start Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 36/54] iwlwifi: pcie: add rules to match Qu with Hr2 Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 37/54] iwlwifi: mvm: guard against device removal in reprobe Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 38/54] iwlwifi: queue: bail out on invalid freeing Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 39/54] SUNRPC: Move simple_get_bytes and simple_get_netobj into private header Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 40/54] SUNRPC: Handle 0 length opaque XDR object data properly Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 41/54] i2c: mediatek: Move suspend and resume handling to NOIRQ phase Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 42/54] blk-cgroup: Use cond_resched() when destroy blkgs Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 43/54] regulator: Fix lockdep warning resolving supplies Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 44/54] bpf: Fix verifier jmp32 pruning decision logic Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 45/54] bpf: Fix 32 bit src register truncation on div/mod Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 46/54] bpf: Fix verifier jsgt branch analysis on max bound Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 47/54] drm/i915: Fix ICL MG PHY vswing handling Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 48/54] drm/i915: Skip vswing programming for TBT Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 49/54] nilfs2: make splice write available again Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 50/54] Revert "mm: memcontrol: avoid workload stalls when lowering memory.high" Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 51/54] squashfs: avoid out of bounds writes in decompressors Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 52/54] squashfs: add more sanity checks in id lookup Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 53/54] squashfs: add more sanity checks in inode lookup Greg Kroah-Hartman
2021-02-11 15:02 ` [PATCH 5.10 54/54] squashfs: add more sanity checks in xattr id lookup Greg Kroah-Hartman
2021-02-12  3:16 ` [PATCH 5.10 00/54] 5.10.16-rc1 review Naresh Kamboju
2021-02-13 12:58   ` Greg Kroah-Hartman
2021-02-12 16:17 ` Shuah Khan
2021-02-13 12:58   ` Greg Kroah-Hartman
2021-02-17 22:45     ` Shuah Khan
2021-02-12 18:08 ` Guenter Roeck
2021-02-13 12:58   ` Greg Kroah-Hartman
2021-02-12 19:21 ` Florian Fainelli
2021-02-13 12:58   ` Greg Kroah-Hartman
2021-02-12 19:54 ` Pavel Machek
2021-02-13 12:58   ` Greg Kroah-Hartman
2021-02-13  3:20 ` Ross Schmidt
2021-02-13 12:57   ` Greg Kroah-Hartman

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210211150153.326935464@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=asml.silence@gmail.com \
    --cc=axboe@kernel.dk \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).