linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Mateusz Guzik <mjguzik@gmail.com>
To: viro@zeniv.linux.org.uk
Cc: brauner@kernel.org, jack@suse.cz, linux-kernel@vger.kernel.org,
	linux-fsdevel@vger.kernel.org, Mateusz Guzik <mjguzik@gmail.com>
Subject: [PATCH] close_files(): reimplement based on do_close_on_exec()
Date: Mon, 12 Aug 2024 09:56:58 +0200	[thread overview]
Message-ID: <20240812075659.1399447-1-mjguzik@gmail.com> (raw)
In-Reply-To: <20240812064427.240190-3-viro@zeniv.linux.org.uk>

While here take more advantage of the fact nobody should be messing with
the table anymore and don't clear the fd slot.

Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
---

how about this instead, I think it's a nicer clean up.

It's literally do_close_on_exec except locking and put fd are deleted.

boots & does not blow up, but admittedly I did not bother with ltp or
any serious testing

 fs/file.c | 37 +++++++++++++++++++++----------------
 1 file changed, 21 insertions(+), 16 deletions(-)

diff --git a/fs/file.c b/fs/file.c
index 74d7ad676579..3ff2e8265156 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -389,33 +389,38 @@ struct files_struct *dup_fd(struct files_struct *oldf, unsigned int max_fds)
 	return newf;
 }
 
-static struct fdtable *close_files(struct files_struct * files)
+static struct fdtable *close_files(struct files_struct *files)
 {
 	/*
 	 * It is safe to dereference the fd table without RCU or
 	 * ->file_lock because this is the last reference to the
 	 * files structure.
+	 *
+	 * For the same reason we can skip locking.
 	 */
 	struct fdtable *fdt = rcu_dereference_raw(files->fdt);
-	unsigned int i, j = 0;
+	unsigned i;
 
-	for (;;) {
+	for (i = 0; ; i++) {
 		unsigned long set;
-		i = j * BITS_PER_LONG;
-		if (i >= fdt->max_fds)
+		unsigned fd = i * BITS_PER_LONG;
+		fdt = files_fdtable(files);
+		if (fd >= fdt->max_fds)
 			break;
-		set = fdt->open_fds[j++];
-		while (set) {
-			if (set & 1) {
-				struct file * file = xchg(&fdt->fd[i], NULL);
-				if (file) {
-					filp_close(file, files);
-					cond_resched();
-				}
-			}
-			i++;
-			set >>= 1;
+		set = fdt->open_fds[i];
+		if (!set)
+			continue;
+		for ( ; set ; fd++, set >>= 1) {
+			struct file *file;
+			if (!(set & 1))
+				continue;
+			file = fdt->fd[fd];
+			if (!file)
+				continue;
+			filp_close(file, files);
+			cond_resched();
 		}
+
 	}
 
 	return fdt;
-- 
2.43.0


  reply	other threads:[~2024-08-12  7:57 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-12  6:42 [PATCHES] fs/file.c stuff Al Viro
2024-08-12  6:44 ` [PATCH 01/11] get rid of ...lookup...fdget_rcu() family Al Viro
2024-08-12  6:44   ` [PATCH 02/11] remove pointless includes of <linux/fdtable.h> Al Viro
2024-08-12  9:25     ` Christian Brauner
2024-08-12  6:44   ` [PATCH 03/11] close_files(): don't bother with xchg() Al Viro
2024-08-12  7:56     ` Mateusz Guzik [this message]
2024-08-14  5:24       ` [PATCH] close_files(): reimplement based on do_close_on_exec() Al Viro
2024-08-14  5:34         ` Mateusz Guzik
2024-08-12  6:44   ` [PATCH 04/11] proc_fd_getattr(): don't bother with S_ISDIR() check Al Viro
2024-08-12  6:44   ` [PATCH 05/11] move close_range(2) into fs/file.c, fold __close_range() into it Al Viro
2024-08-12  6:44   ` [PATCH 06/11] sane_fdtable_size(): don't bother looking at descriptors we are not going to copy Al Viro
2024-08-12  9:30     ` Christian Brauner
2024-08-12  6:44   ` [PATCH 07/11] fs/file.c: remove sanity_check and add likely/unlikely in alloc_fd() Al Viro
2024-08-12  6:44   ` [PATCH 08/11] fs/file.c: conditionally clear full_fds Al Viro
2024-08-12  6:44   ` [PATCH 09/11] fs/file.c: add fast path in find_next_fd() Al Viro
2024-08-12  6:44   ` [PATCH 10/11] alloc_fdtable(): change calling conventions Al Viro
2024-08-12  9:35     ` Christian Brauner
2024-08-12  6:44   ` [PATCH 11/11] dup_fd(): " Al Viro
2024-08-12  9:32     ` Christian Brauner
2024-08-12  9:24   ` [PATCH 01/11] get rid of ...lookup...fdget_rcu() family Christian Brauner

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=20240812075659.1399447-1-mjguzik@gmail.com \
    --to=mjguzik@gmail.com \
    --cc=brauner@kernel.org \
    --cc=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=viro@zeniv.linux.org.uk \
    /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).