All of lore.kernel.org
 help / color / mirror / Atom feed
From: P J P <1883083@bugs.launchpad.net>
To: qemu-devel@nongnu.org
Subject: [Bug 1883083] [NEW] QEMU: block/vvfat driver issues
Date: Thu, 11 Jun 2020 10:03:36 -0000	[thread overview]
Message-ID: <159186981638.8308.16825672026309340419.malonedeb@gac.canonical.com> (raw)

Public bug reported:

Nathan Huckleberry <nhuck15@gmail.com> has reported following issues in
the block/vvfat driver for the virtual VFAT file system image, used to
share a host system directory with a guest VM.

Please note:
  -> https://www.qemu.org/docs/master/system/images.html#virtual-fat-disk-images

Virtual VFAT read/write support is available only for (beta) testing
purposes.

Following issues are reproducible with:

   host)$ ./bin/qemu-system-x86_64 -nographic -enable-kvm \
              -drive file=fat:rw:/tmp/var/run/,index=2  -m 2048 /var/lib/libvirt/images/f27vm.qcow2

  guest)# mount -t vfat /dev/sdb1 /mnt/

The attached reproducers (run inside a guest) include:

1. dir.sh: - directory traversal on the host
   - It creates a file under /mnt/yyyy
   - Then edits the VFAT directory entry to make it -> /mnt/../y
   - The handle_renames_and_mkdirs() routine does not check this new file name
     and creates a file outside of the shared directory on the host

2. dos.sh: hits an assertion failure in vvfat driver
   - Creates a deep directory tree like - /mnt/0/1/2/3/4/5/6/../29/30/
   - While updating vvfat commits, driver hits an assertion in
     handle_renames_and_mkdirs
       ...
       } else if (commit->action == ACTION_MKDIR) {
           ...
           assert(j < s->mapping.next);    <== it fails

3. read.sh: reads past vvfat directory entries
   - Creates a file with: echo "x" > /mnt/a
   - Reads past VVFAT directory entry structure with

       # head -c 1000000 $MNTDEV | xxd | grep x -A 512

   - It may disclose some heap addresses.

4. write.sh: heap buffer overflow
   - Creates large number of files as /mnt/file[1..35]
   - while syncing directory tree with the host, driver hits an overflow
     while doing memmove(3) in array_roll() routine

** Affects: qemu
     Importance: Undecided
         Status: New


** Tags: qemu

** Attachment added: "vvfat-reproducers-shared-by-Nathan"
   https://bugs.launchpad.net/bugs/1883083/+attachment/5382870/+files/vvfat-issues.tar.xz

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1883083

Title:
  QEMU: block/vvfat driver issues

Status in QEMU:
  New

Bug description:
  Nathan Huckleberry <nhuck15@gmail.com> has reported following issues
  in the block/vvfat driver for the virtual VFAT file system image, used
  to share a host system directory with a guest VM.

  Please note:
    -> https://www.qemu.org/docs/master/system/images.html#virtual-fat-disk-images

  Virtual VFAT read/write support is available only for (beta) testing
  purposes.

  Following issues are reproducible with:

     host)$ ./bin/qemu-system-x86_64 -nographic -enable-kvm \
                -drive file=fat:rw:/tmp/var/run/,index=2  -m 2048 /var/lib/libvirt/images/f27vm.qcow2

    guest)# mount -t vfat /dev/sdb1 /mnt/

  The attached reproducers (run inside a guest) include:

  1. dir.sh: - directory traversal on the host
     - It creates a file under /mnt/yyyy
     - Then edits the VFAT directory entry to make it -> /mnt/../y
     - The handle_renames_and_mkdirs() routine does not check this new file name
       and creates a file outside of the shared directory on the host

  2. dos.sh: hits an assertion failure in vvfat driver
     - Creates a deep directory tree like - /mnt/0/1/2/3/4/5/6/../29/30/
     - While updating vvfat commits, driver hits an assertion in
       handle_renames_and_mkdirs
         ...
         } else if (commit->action == ACTION_MKDIR) {
             ...
             assert(j < s->mapping.next);    <== it fails

  3. read.sh: reads past vvfat directory entries
     - Creates a file with: echo "x" > /mnt/a
     - Reads past VVFAT directory entry structure with

         # head -c 1000000 $MNTDEV | xxd | grep x -A 512

     - It may disclose some heap addresses.

  4. write.sh: heap buffer overflow
     - Creates large number of files as /mnt/file[1..35]
     - while syncing directory tree with the host, driver hits an overflow
       while doing memmove(3) in array_roll() routine

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1883083/+subscriptions


             reply	other threads:[~2020-06-11 10:11 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-11 10:03 P J P [this message]
2021-04-29  9:54 ` [Bug 1883083] Re: QEMU: block/vvfat driver issues Thomas Huth
2021-05-11  5:48 ` Thomas Huth

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=159186981638.8308.16825672026309340419.malonedeb@gac.canonical.com \
    --to=1883083@bugs.launchpad.net \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.