public inbox for gfs2@lists.linux.dev
 help / color / mirror / Atom feed
From: Andreas Gruenbacher <agruenba@redhat.com>
To: gfs2@lists.linux.dev
Cc: Andreas Gruenbacher <agruenba@redhat.com>
Subject: [PATCH 10/16] gfs2: Clean up delete work processing
Date: Thu, 26 Sep 2024 00:03:20 +0200	[thread overview]
Message-ID: <20240925220331.417856-11-agruenba@redhat.com> (raw)
In-Reply-To: <20240925220331.417856-1-agruenba@redhat.com>

Function delete_work_func() was previously assuming that the
GLF_TRY_TO_EVICT and GLF_VERIFY_DELETE flags won't both be set at the
same time, but there probably are races in which that can happen, so
handle that case correctly.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/glock.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 9ec305b8d4d4..95f082f13a8c 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -1028,6 +1028,7 @@ static void delete_work_func(struct work_struct *work)
 	struct delayed_work *dwork = to_delayed_work(work);
 	struct gfs2_glock *gl = container_of(dwork, struct gfs2_glock, gl_delete);
 	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
+	bool verify_delete = test_and_clear_bit(GLF_VERIFY_DELETE, &gl->gl_flags);
 
 	if (test_and_clear_bit(GLF_TRY_TO_EVICT, &gl->gl_flags)) {
 		/*
@@ -1048,15 +1049,15 @@ static void delete_work_func(struct work_struct *work)
 		 * step entirely.
 		 */
 		if (gfs2_try_evict(gl)) {
-			if (test_bit(SDF_KILL, &sdp->sd_flags))
-				goto out;
-			if (gfs2_queue_verify_delete(gl, true))
-				return;
+			if (!test_bit(SDF_KILL, &sdp->sd_flags)) {
+				gfs2_glock_hold(gl);
+				if (!gfs2_queue_verify_delete(gl, true))
+					gfs2_glock_put(gl);
+			}
 		}
-		goto out;
 	}
 
-	if (test_and_clear_bit(GLF_VERIFY_DELETE, &gl->gl_flags)) {
+	if (verify_delete) {
 		u64 no_addr = gl->gl_name.ln_number;
 		struct inode *inode;
 
@@ -1073,7 +1074,6 @@ static void delete_work_func(struct work_struct *work)
 		}
 	}
 
-out:
 	gfs2_glock_put(gl);
 }
 
-- 
2.46.0


  parent reply	other threads:[~2024-09-25 22:03 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-09-25 22:03 [PATCH 00/16] gfs2: unlinked inodes Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 01/16] gfs2: Rename GLF_VERIFY_EVICT to GLF_VERIFY_DELETE Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 02/16] gfs2: Initialize gl_no_formal_ino earlier Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 03/16] gfs2: Allow immediate GLF_VERIFY_DELETE work Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 04/16] gfs2: Fix unlinked inode cleanup Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 05/16] gfs2: Faster gfs2_upgrade_iopen_glock wakeups Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 06/16] gfs2: Rename GIF_{DEFERRED -> DEFER}_DELETE Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 07/16] gfs2: Rename dinode_demise to evict_behavior Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 08/16] gfs2: Return enum evict_behavior from gfs2_upgrade_iopen_glock Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 09/16] gfs2: Minor delete_work_func cleanup Andreas Gruenbacher
2024-09-25 22:03 ` Andreas Gruenbacher [this message]
2024-09-25 22:03 ` [PATCH 11/16] gfs2: Call gfs2_queue_verify_delete from gfs2_evict_inode Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 12/16] gfs2: Update to the evict / remote delete documentation Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 13/16] gfs2: Use mod_delayed_work in gfs2_queue_try_to_evict Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 14/16] gfs2: Randomize GLF_VERIFY_DELETE work delay Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 15/16] gfs2: Use get_random_u32 in gfs2_orlov_skip Andreas Gruenbacher
2024-09-25 22:03 ` [PATCH 16/16] gfs2: Make gfs2_inode_refresh static Andreas Gruenbacher
2024-09-26 12:18 ` [PATCH 00/16] gfs2: unlinked inodes Andrew Price

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=20240925220331.417856-11-agruenba@redhat.com \
    --to=agruenba@redhat.com \
    --cc=gfs2@lists.linux.dev \
    /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