From patchwork Wed Mar 30 01:51:47 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ryan Harper X-Patchwork-Id: 88873 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [199.232.76.165]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 0E37FB6F73 for ; Wed, 30 Mar 2011 12:52:47 +1100 (EST) Received: from localhost ([127.0.0.1]:33175 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Q4kaR-00065G-Br for incoming@patchwork.ozlabs.org; Tue, 29 Mar 2011 21:52:43 -0400 Received: from [140.186.70.92] (port=38921 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Q4kZk-000654-EY for qemu-devel@nongnu.org; Tue, 29 Mar 2011 21:52:01 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Q4kZi-0008Gi-Ol for qemu-devel@nongnu.org; Tue, 29 Mar 2011 21:52:00 -0400 Received: from e1.ny.us.ibm.com ([32.97.182.141]:50070) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Q4kZi-0008G7-M4 for qemu-devel@nongnu.org; Tue, 29 Mar 2011 21:51:58 -0400 Received: from d01dlp02.pok.ibm.com (d01dlp02.pok.ibm.com [9.56.224.85]) by e1.ny.us.ibm.com (8.14.4/8.13.1) with ESMTP id p2U1fdwA000451 for ; Tue, 29 Mar 2011 21:41:39 -0400 Received: from d01relay02.pok.ibm.com (d01relay02.pok.ibm.com [9.56.227.234]) by d01dlp02.pok.ibm.com (Postfix) with ESMTP id 099C86E8036 for ; Tue, 29 Mar 2011 21:51:54 -0400 (EDT) Received: from d03av03.boulder.ibm.com (d03av03.boulder.ibm.com [9.17.195.169]) by d01relay02.pok.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id p2U1prnm478688 for ; Tue, 29 Mar 2011 21:51:53 -0400 Received: from d03av03.boulder.ibm.com (loopback [127.0.0.1]) by d03av03.boulder.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id p2U1pqFq009834 for ; Tue, 29 Mar 2011 19:51:53 -0600 Received: from localhost.localdomain (frylock.austin.ibm.com [9.53.41.12]) by d03av03.boulder.ibm.com (8.14.4/8.13.1/NCO v10.0 AVin) with ESMTP id p2U1pqiJ009824; Tue, 29 Mar 2011 19:51:52 -0600 Received: by localhost.localdomain (Postfix, from userid 1000) id 037E21BF83; Tue, 29 Mar 2011 20:51:48 -0500 (CDT) Date: Tue, 29 Mar 2011 20:51:47 -0500 From: Ryan Harper To: Markus Armbruster Message-ID: <20110330015147.GG5445@us.ibm.com> Mime-Version: 1.0 Content-Disposition: inline User-Agent: Mutt/1.5.6+20040907i X-Content-Scanned: Fidelis XPS MAILER X-detected-operating-system: by eggs.gnu.org: GNU/Linux 2.6, seldom 2.4 (older, 4) X-Received-From: 32.97.182.141 Cc: Stefan Hajnoczi , Kevin Wolf , qemu-devel@nongnu.org Subject: [Qemu-devel] [PATCH v4] Do not delete BlockDriverState when deleting the drive X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org When removing a drive from the host-side via drive_del we currently have the following path: drive_del qemu_aio_flush() bdrv_close() // zaps bs->drv, which makes any subsequent I/O get // dropped. Works as designed drive_uninit() bdrv_delete() // frees the bs. Since the device is still connected to // bs, any subsequent I/O is a use-after-free. The value of bs->drv becomes unpredictable on free. As long as it remains null, I/O still gets dropped, however it could become non-null at any point after the free resulting SEGVs or other QEMU state corruption. To resolve this issue as simply as possible, we can chose to not actually delete the BlockDriverState pointer. Since bdrv_close() handles setting the drv pointer to NULL, we just need to remove the BlockDriverState from the QLIST that is used to enumerate the block devices. This is currently handled within bdrv_delete, so move this into its own function, bdrv_make_anon(). The result is that we can now invoke drive_del, this closes the file descriptors and sets BlockDriverState->drv to NULL which prevents futher IO to the device, and since we do not free BlockDriverState, we don't have to worry about the copy retained in the block devices. We also don't attempt to remove the qdev property since we are no longer deleting the BlockDriverState on drives with associated drives. This also allows for removing Drives with no devices associated either. Reported-by: Markus Armbruster Signed-off-by: Ryan Harper Acked-by: Markus Armbruster --- v3->v4 - add back missing nulling of device_name in v2 - use drive_uninit() when removing drive with no peer - align commit message on 72 char boundary v2->v3 - Update drive_del use after free description - s/bdrv_remove/bdrv_make_anon/g - Don't remove qdev property since we don't delete bs any more - If (bs->peer) bdrv_make_anon else bdrv_delete to handle removing drives with no device. v1->v2 - NULL bs->device_name after removing from list to prevent second removal. block.c | 14 +++++++++++--- block.h | 1 + blockdev.c | 25 ++++++++----------------- 3 files changed, 20 insertions(+), 20 deletions(-) diff --git a/block.c b/block.c index c8e2f97..c93ec6d 100644 --- a/block.c +++ b/block.c @@ -697,14 +697,22 @@ void bdrv_close_all(void) } } +/* make a BlockDriverState anonymous by removing from bdrv_state list. + Also, NULL terminate the device_name to prevent double remove */ +void bdrv_make_anon(BlockDriverState *bs) +{ + if (bs->device_name[0] != '\0') { + QTAILQ_REMOVE(&bdrv_states, bs, list); + } + bs->device_name[0] = '\0'; +} + void bdrv_delete(BlockDriverState *bs) { assert(!bs->peer); /* remove from list, if necessary */ - if (bs->device_name[0] != '\0') { - QTAILQ_REMOVE(&bdrv_states, bs, list); - } + bdrv_make_anon(bs); bdrv_close(bs); if (bs->file != NULL) { diff --git a/block.h b/block.h index 5d78fc0..52e9cad 100644 --- a/block.h +++ b/block.h @@ -66,6 +66,7 @@ int bdrv_create(BlockDriver *drv, const char* filename, QEMUOptionParameter *options); int bdrv_create_file(const char* filename, QEMUOptionParameter *options); BlockDriverState *bdrv_new(const char *device_name); +void bdrv_make_anon(BlockDriverState *bs); void bdrv_delete(BlockDriverState *bs); int bdrv_file_open(BlockDriverState **pbs, const char *filename, int flags); int bdrv_open(BlockDriverState *bs, const char *filename, int flags, diff --git a/blockdev.c b/blockdev.c index ecf2252..c810b16 100644 --- a/blockdev.c +++ b/blockdev.c @@ -737,8 +737,6 @@ int do_drive_del(Monitor *mon, const QDict *qdict, QObject **ret_data) { const char *id = qdict_get_str(qdict, "id"); BlockDriverState *bs; - BlockDriverState **ptr; - Property *prop; bs = bdrv_find(id); if (!bs) { @@ -755,24 +753,17 @@ int do_drive_del(Monitor *mon, const QDict *qdict, QObject **ret_data) bdrv_flush(bs); bdrv_close(bs); - /* clean up guest state from pointing to host resource by - * finding and removing DeviceState "drive" property */ + /* if we have a device associated with this BlockDriverState (bs->peer) + * then we need to make the drive anonymous until the device + * can be removed. If this is a drive with no device backing + * then we can just get rid of the block driver state right here. + */ if (bs->peer) { - for (prop = bs->peer->info->props; prop && prop->name; prop++) { - if (prop->info->type == PROP_TYPE_DRIVE) { - ptr = qdev_get_prop_ptr(bs->peer, prop); - if (*ptr == bs) { - bdrv_detach(bs, bs->peer); - *ptr = NULL; - break; - } - } - } + bdrv_make_anon(bs); + } else { + drive_uninit(drive_get_by_blockdev(bs)); } - /* clean up host side */ - drive_uninit(drive_get_by_blockdev(bs)); - return 0; }