From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:57692) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Ur9Ov-0003GC-5j for qemu-devel@nongnu.org; Mon, 24 Jun 2013 12:13:59 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Ur9Os-0002U6-Cw for qemu-devel@nongnu.org; Mon, 24 Jun 2013 12:13:57 -0400 Received: from mx.ipv6.kamp.de ([2a02:248:0:51::16]:44758 helo=mx01.kamp.de) by eggs.gnu.org with smtp (Exim 4.71) (envelope-from ) id 1Ur9Lv-0001QE-J1 for qemu-devel@nongnu.org; Mon, 24 Jun 2013 12:10:51 -0400 Message-ID: <51C86F7D.401@kamp.de> Date: Mon, 24 Jun 2013 18:10:37 +0200 From: Peter Lieven MIME-Version: 1.0 References: <1371934712-11714-1-git-send-email-pl@kamp.de> <1371934712-11714-8-git-send-email-pl@kamp.de> <51C85808.7000206@redhat.com> In-Reply-To: <51C85808.7000206@redhat.com> Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 7/8] iscsi: assert that sectors are aligned to LUN blocksize List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Paolo Bonzini Cc: Kevin Wolf , ronniesahlberg@gmail.com, qemu-devel@nongnu.org, Stefan Hajnoczi Am 24.06.2013 16:30, schrieb Paolo Bonzini: > Il 22/06/2013 22:58, Peter Lieven ha scritto: >> if the blocksize of an iSCSI LUN is bigger than the BDRV_SECTOR_SIZE >> it is possible that sector_num or nb_sectors are not correctly >> alligned. >> >> for now assert that there is no misalignment to avoid data >> corruption. > You should just fail to open the LUN instead. Ronnie added support for reading fragments of blocksize if the offset is aligned. I don't know what the idea was, maybe for qemu reading the boot sector. If the OS does only read multiple of blocksize at aligned offsets everything should work. I now that our storages support 4K blocksize. I can check if it is usable. I could also just fail the operations instead of asserting. Peter > > Paolo > >> Signed-off-by: Peter Lieven >> --- >> block/iscsi.c | 7 +++++++ >> 1 file changed, 7 insertions(+) >> >> diff --git a/block/iscsi.c b/block/iscsi.c >> index 683692a..2c410b1 100644 >> --- a/block/iscsi.c >> +++ b/block/iscsi.c >> @@ -249,6 +249,7 @@ iscsi_aio_write16_cb(struct iscsi_context *iscsi, int status, >> >> static int64_t sector_qemu2lun(int64_t sector, IscsiLun *iscsilun) >> { >> + assert((sector * BDRV_SECTOR_SIZE) % iscsilun->block_size == 0); >> return sector * BDRV_SECTOR_SIZE / iscsilun->block_size; >> } >> >> @@ -269,6 +270,8 @@ iscsi_aio_writev_acb(IscsiAIOCB *acb) >> acb->status = -EINPROGRESS; >> acb->buf = NULL; >> >> + assert((acb->nb_sectors * BDRV_SECTOR_SIZE) % acb->iscsilun->block_size == 0); >> + >> /* this will allow us to get rid of 'buf' completely */ >> size = acb->nb_sectors * BDRV_SECTOR_SIZE; >> >> @@ -618,6 +621,8 @@ static int iscsi_aio_discard_acb(IscsiAIOCB *acb) { >> acb->buf = NULL; >> >> list[0].lba = sector_qemu2lun(acb->sector_num, acb->iscsilun); >> + >> + assert((acb->nb_sectors * BDRV_SECTOR_SIZE) % acb->iscsilun->block_size == 0); >> list[0].num = acb->nb_sectors * BDRV_SECTOR_SIZE / acb->iscsilun->block_size; >> >> acb->task = iscsi_unmap_task(iscsi, acb->iscsilun->lun, >> @@ -947,6 +952,8 @@ coroutine_fn iscsi_co_write_zeroes(BlockDriverState *bs, int64_t sector_num, >> iTask.complete = 0; >> iTask.bs = bs; >> >> + assert((nb_sectors * BDRV_SECTOR_SIZE) % iscsilun->block_size == 0); >> + >> list[0].lba = sector_qemu2lun(sector_num, iscsilun); >> list[0].num = nb_sectors * BDRV_SECTOR_SIZE / iscsilun->block_size; >> >>