stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Willy Tarreau <w@1wt.eu>
To: gregkh@linuxfoundation.org
Cc: geert@linux-m68k.org, marius.gorski@gmail.com,
	miguel.ojeda.sandonis@gmail.com, stable@vger.kernel.org
Subject: Re: FAILED: patch "[PATCH] auxdisplay: charlcd: properly restore atomic counter on error" failed to apply to 4.9-stable tree
Date: Sun, 12 Nov 2017 14:09:49 +0100	[thread overview]
Message-ID: <20171112130949.GA11481@1wt.eu> (raw)
In-Reply-To: <15076478239963@kroah.com>

[-- Attachment #1: Type: text/plain, Size: 214 bytes --]

Hi Greg,

as promised a month ago, here finally come the tested backports of this
fix for 4.9 and 4.4 respectively. This backport is not needed for 3.18
as the regression was introduced in 3.19-rc1.

Cheers,
Willy

[-- Attachment #2: panel-4.4.patch --]
[-- Type: text/plain, Size: 2394 bytes --]

>From 29e391ca59d60b7902b09ed11edbc5c95c372f09 Mon Sep 17 00:00:00 2001
From: Willy Tarreau <w@1wt.eu>
Date: Thu, 7 Sep 2017 11:56:40 +0200
Subject: misc: panel: properly restore atomic counter on error path

Commit f4757af ("staging: panel: Fix single-open policy race condition")
introduced in 3.19-rc1 attempted to fix a race condition on the open, but
failed to properly do it and used to exit without restoring the semaphore.

This results in -EBUSY being returned after the first open error until
the module is reloaded or the system restarted (ie: consecutive to a
dual open resulting in -EBUSY or to a permission error).

Fixes: f4757af85 # 3.19-rc1
Cc: stable@vger.kernel.org [3.19+]
Cc: Mariusz Gorski <marius.gorski@gmail.com>
Signed-off-by: Willy Tarreau <w@1wt.eu>
[wt: driver is in staging/panel in 4.4]
---
 drivers/staging/panel/panel.c | 23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)

diff --git a/drivers/staging/panel/panel.c b/drivers/staging/panel/panel.c
index 70b8f4f..e658e11 100644
--- a/drivers/staging/panel/panel.c
+++ b/drivers/staging/panel/panel.c
@@ -1431,17 +1431,25 @@ static ssize_t lcd_write(struct file *file,
 
 static int lcd_open(struct inode *inode, struct file *file)
 {
+	int ret;
+
+	ret = -EBUSY;
 	if (!atomic_dec_and_test(&lcd_available))
-		return -EBUSY;	/* open only once at a time */
+		goto fail; /* open only once at a time */
 
+	ret = -EPERM;
 	if (file->f_mode & FMODE_READ)	/* device is write-only */
-		return -EPERM;
+		goto fail;
 
 	if (lcd.must_clear) {
 		lcd_clear_display();
 		lcd.must_clear = false;
 	}
 	return nonseekable_open(inode, file);
+
+ fail:
+	atomic_inc(&lcd_available);
+	return ret;
 }
 
 static int lcd_release(struct inode *inode, struct file *file)
@@ -1704,14 +1712,21 @@ static ssize_t keypad_read(struct file *file,
 
 static int keypad_open(struct inode *inode, struct file *file)
 {
+	int ret;
+
+	ret = -EBUSY;
 	if (!atomic_dec_and_test(&keypad_available))
-		return -EBUSY;	/* open only once at a time */
+		goto fail;	/* open only once at a time */
 
+	ret = -EPERM;
 	if (file->f_mode & FMODE_WRITE)	/* device is read-only */
-		return -EPERM;
+		goto fail;
 
 	keypad_buflen = 0;	/* flush the buffer on opening */
 	return 0;
+ fail:
+	atomic_inc(&keypad_available);
+	return ret;
 }
 
 static int keypad_release(struct inode *inode, struct file *file)
-- 
1.7.12.1


[-- Attachment #3: panel-4.9.patch --]
[-- Type: text/plain, Size: 2346 bytes --]

>From 7ddfddb3fc0afb2aa36e89b554f06c901c5fa36f Mon Sep 17 00:00:00 2001
From: Willy Tarreau <w@1wt.eu>
Date: Thu, 7 Sep 2017 11:56:40 +0200
Subject: misc: panel: properly restore atomic counter on error path

Commit f4757af ("staging: panel: Fix single-open policy race condition")
introduced in 3.19-rc1 attempted to fix a race condition on the open, but
failed to properly do it and used to exit without restoring the semaphore.

This results in -EBUSY being returned after the first open error until
the module is reloaded or the system restarted (ie: consecutive to a
dual open resulting in -EBUSY or to a permission error).

Fixes: f4757af85 # 3.19-rc1
Cc: stable@vger.kernel.org [3.19+]
Cc: Mariusz Gorski <marius.gorski@gmail.com>
Signed-off-by: Willy Tarreau <w@1wt.eu>
[wt: driver is in misc/panel in 4.9]
---
 drivers/misc/panel.c | 23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)

diff --git a/drivers/misc/panel.c b/drivers/misc/panel.c
index 6030ac5..a9fa4c0 100644
--- a/drivers/misc/panel.c
+++ b/drivers/misc/panel.c
@@ -1423,17 +1423,25 @@ static ssize_t lcd_write(struct file *file,
 
 static int lcd_open(struct inode *inode, struct file *file)
 {
+	int ret;
+
+	ret = -EBUSY;
 	if (!atomic_dec_and_test(&lcd_available))
-		return -EBUSY;	/* open only once at a time */
+		goto fail; /* open only once at a time */
 
+	ret = -EPERM;
 	if (file->f_mode & FMODE_READ)	/* device is write-only */
-		return -EPERM;
+		goto fail;
 
 	if (lcd.must_clear) {
 		lcd_clear_display();
 		lcd.must_clear = false;
 	}
 	return nonseekable_open(inode, file);
+
+ fail:
+	atomic_inc(&lcd_available);
+	return ret;
 }
 
 static int lcd_release(struct inode *inode, struct file *file)
@@ -1696,14 +1704,21 @@ static ssize_t keypad_read(struct file *file,
 
 static int keypad_open(struct inode *inode, struct file *file)
 {
+	int ret;
+
+	ret = -EBUSY;
 	if (!atomic_dec_and_test(&keypad_available))
-		return -EBUSY;	/* open only once at a time */
+		goto fail;	/* open only once at a time */
 
+	ret = -EPERM;
 	if (file->f_mode & FMODE_WRITE)	/* device is read-only */
-		return -EPERM;
+		goto fail;
 
 	keypad_buflen = 0;	/* flush the buffer on opening */
 	return 0;
+ fail:
+	atomic_inc(&keypad_available);
+	return ret;
 }
 
 static int keypad_release(struct inode *inode, struct file *file)
-- 
1.7.12.1


  parent reply	other threads:[~2017-11-12 13:09 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-10 15:03 FAILED: patch "[PATCH] auxdisplay: charlcd: properly restore atomic counter on error" failed to apply to 4.9-stable tree gregkh
2017-10-10 20:22 ` Willy Tarreau
2017-11-12 13:09 ` Willy Tarreau [this message]
2017-11-16 17:41   ` Greg KH

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=20171112130949.GA11481@1wt.eu \
    --to=w@1wt.eu \
    --cc=geert@linux-m68k.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=marius.gorski@gmail.com \
    --cc=miguel.ojeda.sandonis@gmail.com \
    --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).