From nobody Mon Sep 16 20:18:19 2024 Delivered-To: importer2@patchew.org Authentication-Results: mx.zohomail.com; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer2=patchew.org@nongnu.org Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 171273482772384.71409786878303; Wed, 10 Apr 2024 00:40:27 -0700 (PDT) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1ruSNV-0003ZM-5a; Wed, 10 Apr 2024 03:28:17 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1ruSNS-0003SN-8d; Wed, 10 Apr 2024 03:28:14 -0400 Received: from isrv.corpit.ru ([86.62.121.231]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1ruSNQ-0004uS-Bm; Wed, 10 Apr 2024 03:28:13 -0400 Received: from tsrv.corpit.ru (tsrv.tls.msk.ru [192.168.177.2]) by isrv.corpit.ru (Postfix) with ESMTP id 8EF655D6A8; Wed, 10 Apr 2024 10:25:06 +0300 (MSK) Received: from tls.msk.ru (mjt.wg.tls.msk.ru [192.168.177.130]) by tsrv.corpit.ru (Postfix) with SMTP id 2719FB02E8; Wed, 10 Apr 2024 10:23:08 +0300 (MSK) Received: (nullmailer pid 4191817 invoked by uid 1000); Wed, 10 Apr 2024 07:23:04 -0000 From: Michael Tokarev To: qemu-devel@nongnu.org Cc: qemu-stable@nongnu.org, Fiona Ebner , Kevin Wolf , Stefan Hajnoczi , Michael Tokarev Subject: [Stable-8.2.3 56/87] block-backend: fix edge case in bdrv_next() where BDS associated to BB changes Date: Wed, 10 Apr 2024 10:22:29 +0300 Message-Id: <20240410072303.4191455-56-mjt@tls.msk.ru> X-Mailer: git-send-email 2.39.2 In-Reply-To: References: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer2=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=86.62.121.231; envelope-from=mjt@tls.msk.ru; helo=isrv.corpit.ru X-Spam_score_int: -68 X-Spam_score: -6.9 X-Spam_bar: ------ X-Spam_report: (-6.9 / 5.0 requ) BAYES_00=-1.9, RCVD_IN_DNSWL_HI=-5, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer2=patchew.org@nongnu.org Sender: qemu-devel-bounces+importer2=patchew.org@nongnu.org X-ZM-MESSAGEID: 1712734827892100001 Content-Type: text/plain; charset="utf-8" From: Fiona Ebner The old_bs variable in bdrv_next() is currently determined by looking at the old block backend. However, if the block graph changes before the next bdrv_next() call, it might be that the associated BDS is not the same that was referenced previously. In that case, the wrong BDS is unreferenced, leading to an assertion failure later: > bdrv_unref: Assertion `bs->refcnt > 0' failed. In particular, this can happen in the context of bdrv_flush_all(), when polling for bdrv_co_flush() in the generated co-wrapper leads to a graph change (for example with a stream block job [0]). A racy reproducer: > #!/bin/bash > rm -f /tmp/backing.qcow2 > rm -f /tmp/top.qcow2 > ./qemu-img create /tmp/backing.qcow2 -f qcow2 64M > ./qemu-io -c "write -P42 0x0 0x1" /tmp/backing.qcow2 > ./qemu-img create /tmp/top.qcow2 -f qcow2 64M -b /tmp/backing.qcow2 -F qc= ow2 > ./qemu-system-x86_64 --qmp stdio \ > --blockdev qcow2,node-name=3Dnode0,file.driver=3Dfile,file.filename=3D/tm= p/top.qcow2 \ > < {"execute": "qmp_capabilities"} > {"execute": "block-stream", "arguments": { "job-id": "stream0", "device":= "node0" } } > {"execute": "quit"} > EOF [0]: > #0 bdrv_replace_child_tran (child=3D..., new_bs=3D..., tran=3D...) > #1 bdrv_replace_node_noperm (from=3D..., to=3D..., auto_skip=3D..., tran= =3D..., errp=3D...) > #2 bdrv_replace_node_common (from=3D..., to=3D..., auto_skip=3D..., deta= ch_subchain=3D..., errp=3D...) > #3 bdrv_drop_filter (bs=3D..., errp=3D...) > #4 bdrv_cor_filter_drop (cor_filter_bs=3D...) > #5 stream_prepare (job=3D...) > #6 job_prepare_locked (job=3D...) > #7 job_txn_apply_locked (fn=3D..., job=3D...) > #8 job_do_finalize_locked (job=3D...) > #9 job_exit (opaque=3D...) > #10 aio_bh_poll (ctx=3D...) > #11 aio_poll (ctx=3D..., blocking=3D...) > #12 bdrv_poll_co (s=3D...) > #13 bdrv_flush (bs=3D...) > #14 bdrv_flush_all () > #15 do_vm_stop (state=3D..., send_stop=3D...) > #16 vm_shutdown () Signed-off-by: Fiona Ebner Message-ID: <20240322095009.346989-3-f.ebner@proxmox.com> Reviewed-by: Kevin Wolf Reviewed-by: Stefan Hajnoczi Signed-off-by: Kevin Wolf (cherry picked from commit f6d38c9f6dae6fce99dcaf6ca16a1fe5b5e19c4c) Signed-off-by: Michael Tokarev diff --git a/block/block-backend.c b/block/block-backend.c index ec21148806..4a419f75fb 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -613,14 +613,14 @@ BlockDriverState *bdrv_next(BdrvNextIterator *it) /* Must be called from the main loop */ assert(qemu_get_current_aio_context() =3D=3D qemu_get_aio_context()); =20 + old_bs =3D it->bs; + /* First, return all root nodes of BlockBackends. In order to avoid * returning a BDS twice when multiple BBs refer to it, we only return= it * if the BB is the first one in the parent list of the BDS. */ if (it->phase =3D=3D BDRV_NEXT_BACKEND_ROOTS) { BlockBackend *old_blk =3D it->blk; =20 - old_bs =3D old_blk ? blk_bs(old_blk) : NULL; - do { it->blk =3D blk_all_next(it->blk); bs =3D it->blk ? blk_bs(it->blk) : NULL; @@ -634,11 +634,10 @@ BlockDriverState *bdrv_next(BdrvNextIterator *it) if (bs) { bdrv_ref(bs); bdrv_unref(old_bs); + it->bs =3D bs; return bs; } it->phase =3D BDRV_NEXT_MONITOR_OWNED; - } else { - old_bs =3D it->bs; } =20 /* Then return the monitor-owned BDSes without a BB attached. Ignore a= ll --=20 2.39.2