Re: [Qemu-devel] [PATCH v2 0/9] raw: Prohibit dangerous writes for probed images

[Stefan Hajnoczi <stefanha@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 3224 bytes --]

On Fri, Nov 07, 2014 at 08:39:16PM +0100, Kevin Wolf wrote:
> See the commit message of patch 7 for the why and how. This series
> will probably be only part of the solution and doesn't mean that we
> should stop looking for other patches which improve different parts of
> the problem.
> 
> See the mailing list thread "Image probing: how it can be insecure, and
> what we could do about it" for the complete context.
> 
> v2:
> - Fixed offset in qemu_iovec_concat [Kevin]
> - Added paragraph to patch 7 explaining that we're not breaking
>   additional cases, but only change the failure mode of already
>   broken scenarios [Max]
> - Added a warning when opening an image in "restricted raw" mode,
>   which required a few more patches to make the test cases avoid
>   this warning [Markus]
> 
> 
> Kevin Wolf (8):
>   qemu-io: Allow explicitly specifying format
>   qemu-iotests: Use qemu-io -f $IMGFMT
>   qemu-iotests: Add qemu-io format option in Python tests
>   qtests: Specify image format explicitly
>   block: Read only one sector for format probing
>   raw: Prohibit dangerous writes for probed images
>   qemu-iotests: Fix stderr handling in common.qemu
>   qemu-iotests: Test writing non-raw image headers to raw image
> 
> Markus Armbruster (1):
>   block: Factor bdrv_probe_all() out of find_image_format()
> 
>  block.c                        |  48 +++++++++----
>  block/raw_bsd.c                |  57 +++++++++++++++-
>  include/block/block_int.h      |   5 ++
>  qemu-io.c                      |  28 +++++---
>  tests/ahci-test.c              |   3 +-
>  tests/bios-tables-test.c       |   2 +-
>  tests/drive_del-test.c         |   2 +-
>  tests/fdc-test.c               |   2 +-
>  tests/hd-geo-test.c            |   2 +-
>  tests/i440fx-test.c            |   5 +-
>  tests/ide-test.c               |   9 +--
>  tests/nvme-test.c              |   2 +-
>  tests/qemu-iotests/016         |  11 +--
>  tests/qemu-iotests/030         |  22 +++---
>  tests/qemu-iotests/040         |  32 ++++-----
>  tests/qemu-iotests/048         |   2 +-
>  tests/qemu-iotests/055         |  18 ++---
>  tests/qemu-iotests/058         |  11 +--
>  tests/qemu-iotests/071         |  10 +--
>  tests/qemu-iotests/071.out     |   6 +-
>  tests/qemu-iotests/077         |   2 +-
>  tests/qemu-iotests/081         |   8 ++-
>  tests/qemu-iotests/081.out     |   2 +-
>  tests/qemu-iotests/089         |   6 +-
>  tests/qemu-iotests/109         | 100 +++++++++++++++++++++++++++
>  tests/qemu-iotests/109.out     | 149 +++++++++++++++++++++++++++++++++++++++++
>  tests/qemu-iotests/common      |   2 +-
>  tests/qemu-iotests/common.qemu |   3 +-
>  tests/qemu-iotests/group       |   1 +
>  tests/usb-hcd-uhci-test.c      |   2 +-
>  tests/usb-hcd-xhci-test.c      |   2 +-
>  tests/virtio-blk-test.c        |   4 +-
>  tests/virtio-scsi-test.c       |   4 +-
>  33 files changed, 460 insertions(+), 102 deletions(-)
>  create mode 100755 tests/qemu-iotests/109
>  create mode 100644 tests/qemu-iotests/109.out

QEMU 2.3 material but looks pretty close.  Eric and Max had comments, I
also posted a question about the nb_sectors == 0 edge-case.

Stefan

[-- Attachment #2: Type: application/pgp-signature, Size: 473 bytes --]