From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:55821) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fBI74-0005qR-Ea for qemu-devel@nongnu.org; Wed, 25 Apr 2018 06:57:27 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fBI73-0007yV-IP for qemu-devel@nongnu.org; Wed, 25 Apr 2018 06:57:26 -0400 Date: Wed, 25 Apr 2018 12:57:08 +0200 From: Kevin Wolf Message-ID: <20180425105708.GB4652@localhost.localdomain> References: <20180424192506.149089-1-eblake@redhat.com> <20180424192506.149089-3-eblake@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180424192506.149089-3-eblake@redhat.com> Subject: Re: [Qemu-devel] [PATCH v2 2/6] file-win32: Switch to byte-based callbacks List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eric Blake Cc: qemu-devel@nongnu.org, jsnow@redhat.com, qemu-block@nongnu.org, Max Reitz , Stefan Weil Am 24.04.2018 um 21:25 hat Eric Blake geschrieben: > We are gradually moving away from sector-based interfaces, towards > byte-based. Make the change for the last few sector-based callbacks > in the file-win32 driver. > > Note that the driver was already using byte-based calls for > performing actual I/O, so this just gets rid of a round trip > of scaling; however, as I don't know if Windows is tolerant of > non-sector AIO operations, I went with the conservative approach > of modifying .bdrv_refresh_limits to override the block layer > defaults back to the pre-patch value of 512. > > Signed-off-by: Eric Blake > --- > Compile-tested via 'make docker-test-mingw@fedora', but I don't > have a sane way to test whether it actually works. Tried to test it, and the only result is that something was broken even before your patch: $ ./qemu-img.exe create -f raw /tmp/test.raw 128M Formatting '/tmp/test.raw', fmt=raw size=134217728 $ ./qemu-io.exe -f raw -c 'read 0 4k' /tmp/test.raw read failed: Input/output error For some reason, doing the same with qcow2 works fine. qemu-iotests for qcow2 starts hanging in 013. Maybe someone should look into this, qemu-iotests was working fairly well with mingw builds some time ago. Nothing that will hold up this series, though. Kevin