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]
Message-ID: <951969.99629.qm@web31806.mail.mud.yahoo.com>
Date:	Mon, 25 Apr 2011 02:03:41 -0700 (PDT)
From:	Luben Tuikov <ltuikov@...oo.com>
To:	linux-usb@...r.kernel.org, linux-scsi@...r.kernel.org,
	linux-kernel@...r.kernel.org, Greg KH <greg@...ah.com>
Subject: UASP (UAS) for Linux: Updates

The USB Attached SCSI Protocol driver (UAS, UASP) which can be found here:
https://github.com/ltuikov/linux-2.6 off of master, saw the following update:

commit 1cc384f0b3a9422237ff218e3be10a4a4c15f9b4
Author: Luben Tuikov <ltuikov@...oo.com>
Date:   Fri Apr 22 19:25:07 2011 -0700

    [USB] UASP: Driver generated tags on per I_T nexus
    
    In UAS, a single tag space exists per I_T nexus,
    that is, tags are unique on per I_T nexus. Also,
    tags are shared between COMMAND IUs and TASK
    MANAGEMENT IUs. The technical reason for these
    requirements can be deduced by reading the USB 3.0
    and UAS specs.
    
    This commit implements an infrastructure for
    driver generated tags on per I_T nexus.
    
    A few of the reasons to do this in the driver
    instead of using the tags generated by the block
    layer are:
    
    * Tags exists per I_T nexus. Linux generates I_T_L
      nexus tags. Although some provision is made for
      "host" based tags in Linux, it doesn't scale for
      the reasons described in the following
      bullet-points, and for the reason that if in the
      future multiple UAS interfaces of the same USB
      device are mapped as separate target ports into
      the same "SCSI host (template)", then the tag
      space breaks, as it must not be shared between
      ports. The solution provided by this commit
      solves the general case by providing I_T tags.
    
    * Linux doesn't provide for a driver to ask for
      tags to use for Task Management Functions, which
      are generated by the driver. In Linux tags come
      in already assigned to commands which the driver
      should send out to the transport. Generating
      tags in the driver assures that we can generate
      (unique, non-duplicate with commands) tags for
      TMFs too.
    
    * When doing error recovery, after an ABORT TASK
      TMF, Linux generates a TEST UNIT READY with a
      tag number the same as the task which was
      aborted (the Task To Be Managed, TTBM) by the
      preceding ABORT TASK TMF. Clearly an abuse of a
      tag number.
    
    * Tags should form an increasing sequence wrapping
      when the highest numbered tag has been used.
    
    Signed-off-by: Luben Tuikov <ltuikov@...oo.com>

Original posting of the UASP driver can be found here:
http://marc.info/?l=linux-usb&m=129165511732388&w=2

    Luben

--
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/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ