From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:56598) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cwEyP-00049b-Qc for qemu-devel@nongnu.org; Thu, 06 Apr 2017 17:29:47 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1cwEyO-0000wM-9f for qemu-devel@nongnu.org; Thu, 06 Apr 2017 17:29:45 -0400 Date: Thu, 6 Apr 2017 22:29:36 +0100 From: "Richard W.M. Jones" Message-ID: <20170406212936.GA28793@redhat.com> References: <1488314205-16264-1-git-send-email-kwolf@redhat.com> <1488314205-16264-44-git-send-email-kwolf@redhat.com> <20170406205939.GA28584@redhat.com> <20170406211537.GR30620@redhat.com> <616fdb8b-2c41-9718-d5b3-e335e78745e4@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <616fdb8b-2c41-9718-d5b3-e335e78745e4@redhat.com> Subject: Re: [Qemu-devel] [PULL 43/46] block: Assertions for write permissions List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eric Blake Cc: Kevin Wolf , qemu-devel@nongnu.org, qemu-block@nongnu.org On Thu, Apr 06, 2017 at 04:23:19PM -0500, Eric Blake wrote: > On 04/06/2017 04:15 PM, Richard W.M. Jones wrote: > > >>>> +++ b/block/io.c > >>>> @@ -945,6 +945,8 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child, > >>>> size_t skip_bytes; > >>>> int ret; > >>>> > >>>> + assert(child->perm & (BLK_PERM_WRITE_UNCHANGED | BLK_PERM_WRITE)); > >>> > > > https://bugzilla.redhat.com/show_bug.cgi?id=1439922 > > > > There is also a minimal reproducer in comment 2. > > If it helps, backtrace shows: > > #4 0x0000555555c09f10 in bdrv_co_do_copy_on_readv > (child=0x555556990180, offset=0, bytes=512, qiov=0x7fffffffd0c0) at > block/io.c:948 > #5 0x0000555555c0a33d in bdrv_aligned_preadv (child=0x555556990180, > req=0x7fffb6dffec0, offset=0, bytes=512, align=1, qiov=0x7fffffffd0c0, > flags=1) > at block/io.c:1058 > #6 0x0000555555c0a8c3 in bdrv_co_preadv (child=0x555556990180, > offset=0, bytes=512, qiov=0x7fffffffd0c0, flags=BDRV_REQ_COPY_ON_READ) > at block/io.c:1166 > #7 0x0000555555bf7ccf in blk_co_preadv (blk=0x555556a2bc20, offset=0, > bytes=512, qiov=0x7fffffffd0c0, flags=0) at block/block-backend.c:927 > #8 0x0000555555bf7e19 in blk_read_entry (opaque=0x7fffffffd0e0) > at block/block-backend.c:974 > #9 0x0000555555cc3015 in coroutine_trampoline (i0=1485566720, i1=21845) > at util/coroutine-ucontext.c:79 The patch that fixes this is simply: diff --git a/block/io.c b/block/io.c index 2709a7007f..4f1a730e45 100644 --- a/block/io.c +++ b/block/io.c @@ -945,8 +945,6 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child, size_t skip_bytes; int ret; - assert(child->perm & (BLK_PERM_WRITE_UNCHANGED | BLK_PERM_WRITE)); - /* Cover entire cluster so no additional backing file I/O is required when * allocating cluster in the image file. */ It has no ill effects that I can see, and fixes the libguestfs test suite, but I'm assuming the assert was added for a reason so I'm not proposing this as a patch. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com Fedora Windows cross-compiler. Compile Windows programs, test, and build Windows installers. Over 100 libraries supported. http://fedoraproject.org/wiki/MinGW