fix signed/unsigned overflows in SCSI disk (Rik van Riel)
commite035b43d7c98b1509420d41b017abcdac0de0442
authoraliguori <aliguori@c046a42c-6fe2-441c-8c8c-71466251a162>
Wed, 28 Jan 2009 21:58:22 +0000 (28 21:58 +0000)
committeraliguori <aliguori@c046a42c-6fe2-441c-8c8c-71466251a162>
Wed, 28 Jan 2009 21:58:22 +0000 (28 21:58 +0000)
tree2dccc84bb2394e3b5d5b56474abea2ab24c72a64
parentbaa8c602295b1d33844503ce0a172b85f84646d3
fix signed/unsigned overflows in SCSI disk (Rik van Riel)

Sector numbers can overflow on a virtual scsi disk of over 1TB
in size.  Qemu's bdrv_read expects an int64_t, so fix the overflow
by going to that data type.

On large disks, we clip the capacity to 2TB instead of returning
"capacity modulo 2TB".

Turn sector_count into an unsigned to prevent a signed/unsigned
overflow with SCSI transfers larger than 2TB.  We're unlikely to
ever hit this bug, but fixing it is just one line.

Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@6467 c046a42c-6fe2-441c-8c8c-71466251a162
hw/scsi-disk.c