linux-kernel - Re: [linux-usb-devel] 2.6.19-rc5-mm1

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 9 Nov 2006 14:51:00 -0800
From:	Andrew Morton <akpm@...l.org>
To:	Alan Stern <stern@...land.harvard.edu>
Cc:	Benoit Boissinot <bboissin@...il.com>,
	Mattia Dongili <malattia@...ux.it>,
	USB development list <linux-usb-devel@...ts.sourceforge.net>,
	Kernel development list <linux-kernel@...r.kernel.org>,
	SCSI development list <linux-scsi@...r.kernel.org>
Subject: Re: [linux-usb-devel] 2.6.19-rc5-mm1

On Thu, 9 Nov 2006 16:58:31 -0500 (EST)
Alan Stern <stern@...land.harvard.edu> wrote:

> On Thu, 9 Nov 2006, Mattia Dongili wrote:
> 
> > On Thu, Nov 09, 2006 at 11:04:53AM -0800, Andrew Morton wrote:
> > > 
> > > (added linux-scsi)
> > [...]
> > > > [27526.232000] EIP: [<e8074e26>]
> > > > scsi_device_dev_release_usercontext+0x36/0x100 [scsi_mod] SS:ESP
> > > > 0068:dfdb1e3c
> > > > 
> > > > full dmesg attached, I can test patches and provide any useful
> > > > information if needed (just not now because the dock is at work).
> > > 
> > > You're the second or third person to report this (to no effect, btw). 
> > 
> > oh, great. I was going to report the same (had with usb key unplug).
> > Linux version 2.6.19-rc5-mm1-1 (mattia@...amune) (gcc version 4.1.2 20060901 (prerelease) (Debian 4.1.1-13)) #4 SMP Wed Nov 8 22:46:11 CET 2006
> 
> I don't know exactly where the problem lies, but I have narrowed it down.
> 
> In drivers/scsi/sd.c:sd_probe(), the call to add_disk() increases the 
> device's refcount by 1.  However in sd_remove(), the call to del_gendisk() 
> decreases the device's refcount by 2.  Consequently the structure is 
> deallocated too early, causing the oops.
> 
> Somebody who knows more than I do about add_disk() and del_gendisk() will 
> have to figure what's going wrong.
> 

hm.  Maybe it's the disk_sysfs_symlinks() changes.

Could someone who can reproduce this please try this revert, on
2.6.19-rc2-mm2 through 2.6.19-rc5-mm1?



 fs/partitions/check.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff -puN fs/partitions/check.c~revert-fix-ide-cs-hang-after-device-removal fs/partitions/check.c
--- a/fs/partitions/check.c~revert-fix-ide-cs-hang-after-device-removal
+++ a/fs/partitions/check.c
@@ -416,7 +416,7 @@ static char *make_block_name(struct gend
 
 static int disk_sysfs_symlinks(struct gendisk *disk)
 {
-	struct device *target = disk->driverfs_dev;
+	struct device *target = get_device(disk->driverfs_dev);
 	int err;
 	char *disk_name = NULL;
 
@@ -452,8 +452,9 @@ err_out_dev_link:
 		sysfs_remove_link(&disk->kobj, "device");
 err_out_disk_name:
 		kfree(disk_name);
-	}
 err_out:
+		put_device(target);
+	}
 	return err;
 }
 
_

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/