linux-kernel - race between SCSI uevents and media detection

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:	Tue, 18 Mar 2008 13:21:13 -0400
From:	Bill Nottingham <notting@...hat.com>
To:	linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org
Cc:	kay.sievers@...y.org
Subject: race between SCSI uevents and media detection

For a typical livecd boot using udev, you'd check for the root device
with a udev rule like:

KERNEL=="sr[0-9]", PROGRAM=="/lib/udev/vol_id -l %N", RESULT="<some label>",
	SYMLINK+="root"

This creates a /dev/root symlink to whichever CD device holds the proper
CD/DVD.

However, this currently fails repeatedly on some machines, as the run
of vol_id gets ENOMEDIUM attempting to open the CD device. I suspect
this is because the initial medium scan is actually done after the
uevent for /dev/sr0 is emitted.

Can the open path be fixed to block until the medium state is initially
known? I don't see how this can be worked around from userspace - you
don't get another uevent for the media detection.

Bill
--
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/