From nobody Wed Jun 10 12:31:33 2026 Received: from mail-yw1-f180.google.com (mail-yw1-f180.google.com [209.85.128.180]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 7B2B41C69D for ; Wed, 22 Apr 2026 05:00:26 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.128.180 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1776834028; cv=none; b=GdRuxphKIQNdeSJwpR/2KenvzWic88rHZqKjj5F2JtDdvUxswcn1OjJAggQmTDupYQ3pkMjgDDy+d2dM4SlfACJrTktAHccHwE9Km+pHmozQChljTCegYhMD4vqU3dM2sCnjcTWUQV00blXobpXeknk72/M9FSrOR+29q1TXPiw= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1776834028; c=relaxed/simple; bh=JVg6WuJdpnrGYPBRN8h7I89sVuilX1TIZd3bOZQvQ+U=; h=From:To:Cc:Subject:Date:Message-ID:MIME-Version; b=l5S4KzX7o1V3WHmaXuX72hsUwldXYenDEC+gN+AkjCJBDWpJZg1ZNsJEwaopyiLpdXLASQBtj3RxFqrD27oKF9VrXX3T2uILemNhUCxCBR2p+dV5lpO1qnH4g1Lziae+W693QXbmeWlYlvNmxZxdrt9a6rHmoZxwlz4RYB/xM9Y= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=fail (p=none dis=none) header.from=sung-woo.kim; spf=pass smtp.mailfrom=gmail.com; arc=none smtp.client-ip=209.85.128.180 Authentication-Results: smtp.subspace.kernel.org; dmarc=fail (p=none dis=none) header.from=sung-woo.kim Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Received: by mail-yw1-f180.google.com with SMTP id 00721157ae682-79a7109f568so58502447b3.1 for ; Tue, 21 Apr 2026 22:00:26 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1776834025; x=1777438825; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:x-gm-gg:x-gm-message-state:from:to:cc:subject:date :message-id:reply-to; bh=Hcov1izHTpncQzuxQpNFTuDzi6szui6lCG7Na0Ey9Hc=; b=iWou8kuRPslB4Hft9fejFqCoeXv6BxDmSZ1cGK4yVxoL3NXT5qqeCCF81afX7FGsJA c8y5cOosb6pgLwh73kkVs/qFNaUedd44WofbyRlUz0Uu+jEQ9ytlcK78l01wrV7FCY0Q YOVWmpWxBs2gpFVwvRZS4uEPbUb4nLfmkSJcnTXUxBZSGHqLS9nXblQYQe+xfudnk1Nf R1LXxxd2NUPgE0wQpSlHuYGKz6FEUY10W4n4C2KxANR61zrPFRDO26Dag7ygSIQbFxOW O1GrF+MzzYzcjT9M183V19thexiMhp4kqDQc+lu+742+T2m0Rk+0qgO3fVlyxIsbBhJ8 ZbOQ== X-Forwarded-Encrypted: i=1; AFNElJ/3zRq6MLzBXfgLIg3NIKJlVGaz/rvLum6w0vC2G2dA/D2/7YwHsJUtWOZF3YStMQeL56NSARbqDgWKlKs=@vger.kernel.org X-Gm-Message-State: AOJu0YwTrEmhdP1zNgCEDohZBmSeArVNjEzY4a90+bKH+g4bBK9k+8Rf jPNK86ncd/bxciFeHz12bMkLiqA8d8npUaGyFIzZ4UzmKgXyGuswL+l9 X-Gm-Gg: AeBDietlHkdRHBGXMF9pRuSs0nwBHzoB7PojmHnWFJUOnuxDZ5795qzNdXb/RGu/QMa QkxqVD0loLp3e1s6RH6y5Qgt2kKhb9nOLWyrCRZImF49UnFm5EWcVsCBg0dcfU7gLqSrRWETToW BAO11997++FHYNPTERhosJ6D95UvnyAD1rv7KNj5PNrI7nQe0R3B+Q2HaWfoUaDLdHpxkGbOSm8 bkzbk4ta3lZM+P9rzrOInpvyrnhoA4ACORcqJsTv4abgenhmLFNh9naTy7UR+tjtYq0K3UDoP7r +4i0IvPNud6/25m5Rfe3TTQbdl0spFSefJEAIPEZ24wZoajC6H+jE2NgmJBy2zJfQ+19P+MBuEO 9aM7X+RtBL5mBGacsWiEBLBU/b9u0aUgrvHVT3p6kYo76k693zy5ceMqa3YgZOdunZt3/5cy9Z3 e6fny21JS53UbJeiPLTvt+vsBZ7GDUNNM= X-Received: by 2002:a05:690c:c4f9:b0:7b2:dad2:5e0d with SMTP id 00721157ae682-7b9ecfe760cmr207509197b3.33.1776834025393; Tue, 21 Apr 2026 22:00:25 -0700 (PDT) Received: from tofu.. ([128.210.0.165]) by smtp.googlemail.com with ESMTPSA id 00721157ae682-7baeeaa2cb5sm33636607b3.21.2026.04.21.22.00.22 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 21 Apr 2026 22:00:24 -0700 (PDT) From: Sungwoo Kim To: Davidlohr Bueso , Jonathan Cameron , Dave Jiang , Alison Schofield , Vishal Verma , Ira Weiny , Dan Williams , Ben Widawsky Cc: Dave Tian , Sungwoo Kim , Dan Williams , Jonathan Cameron , linux-cxl@vger.kernel.org, linux-kernel@vger.kernel.org Subject: [PATCH v2] cxl/region: Fix a race bug in delete_region_store Date: Wed, 22 Apr 2026 00:56:36 -0400 Message-ID: <20260422045637.3048249-2-iam@sung-woo.kim> X-Mailer: git-send-email 2.47.3 Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" devm_release_action() cannot find a matching entry in the following two race scenarios. scenario 1: delete two same regions concurrently CPU 0 CPU 1 =3D=3D=3D=3D=3D=3D =3D=3D=3D=3D=3D=3D delete_region_store() cxlr =3D cxl_find_region_by_name() delete_region_store() cxlr =3D cxl_find_region_by_name() devm_release_action() devm_release_action() // cannot find the action, WARN_ON() scenario 2: delete parent and child concurrently [1] CPU0 CPU1 devres_release_all() // take devres_lock remove_nodes(devres_head) // mv to local todo // drop devres_lock delete_region_store() cxlr =3D cxl_find_region_by_name= () // success devm_release_action(unregister_r= egion) devres_release() devres_remove() // hold devres_lock find_dr(devres_head) // do= es not find it WARN_ON(-ENOENT) release_nodes() // drain todo unregister_region(cxlr) // release() cb device_del() To fix scenario 1, delete_region_store() directly calls unregister_region() with a test_and_set_bit(CXL_REGION_F_UNREGISTER). Also, replace devm_release_action() to devm_remove_action() as unregister_region() is now called directly. To fix scenario 2, delete_region_store() removes actions only if the driver is still attached. To ensure this, scoped_guard(device, port->uport_dev) is required to check port->uport_dev->driver. To hold this lock, a workqueue is required for a clean context. Splat: WARNING: drivers/base/devres.c:824 at devm_release_action drivers/base/devr= es.c:824 [inline], CPU#0: syz.1.12224/47589 WARNING: drivers/base/devres.c:824 at devm_release_action+0x2b2/0x360 drive= rs/base/devres.c:817, CPU#0: syz.1.12224/47589 [1] https://lore.kernel.org/linux-cxl/20260310183644.4rwc7ilmzy4t5xp6@offwo= rld/ Fixes: 779dd20cfb56 ("cxl/region: Add region creation support") Suggested-by: Dan Williams Signed-off-by: Sungwoo Kim --- V1: https://lore.kernel.org/linux-cxl/20260308185958.2453707-2-iam@sung-woo= .kim/ V1->V2: - Made devm_remove_action() asynchronous. - Made unregister_region() idempotent. - Addressed Dan's comments and added the suggested-by tag. drivers/cxl/core/region.c | 45 ++++++++++++++++++++++++++++++++++++--- drivers/cxl/cxl.h | 8 +++++++ 2 files changed, 50 insertions(+), 3 deletions(-) diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c index e50dc716d4e8..64db0d332c13 100644 --- a/drivers/cxl/core/region.c +++ b/drivers/cxl/core/region.c @@ -39,6 +39,7 @@ static nodemask_t nodemask_region_seen =3D NODE_MASK_NONE; =20 static struct cxl_region *to_cxl_region(struct device *dev); +static void remove_devm_actions_work(struct work_struct *work); =20 #define __ACCESS_ATTR_RO(_level, _name) { \ .attr =3D { .name =3D __stringify(_name), .mode =3D 0444 }, \ @@ -2543,6 +2544,9 @@ static void unregister_region(void *_cxlr) struct cxl_region_params *p =3D &cxlr->params; int i; =20 + if (test_and_set_bit(CXL_REGION_F_UNREGISTER, &cxlr->flags)) + return; + device_del(&cxlr->dev); =20 /* @@ -2589,6 +2593,8 @@ static struct cxl_region *cxl_region_alloc(struct cxl= _root_decoder *cxlrd, int i dev->type =3D &cxl_region_type; cxl_region_setup_flags(cxlr, &cxlrd->cxlsd.cxld); =20 + INIT_WORK(&cxlr->remove_work, remove_devm_actions_work); + return cxlr; } =20 @@ -2831,20 +2837,53 @@ cxl_find_region_by_name(struct cxl_root_decoder *cx= lrd, const char *name) return to_cxl_region(region_dev); } =20 +static void remove_devm_actions_work(struct work_struct *work) +{ + struct cxl_region *cxlr =3D container_of(work, typeof(*cxlr), remove_work= ); + struct cxl_root_decoder *cxlrd =3D cxlr->cxlrd; + struct cxl_port *port =3D to_cxl_port(cxlrd->cxlsd.cxld.dev.parent); + + if (test_and_set_bit(CXL_REGION_F_DEVM_REMOVE, &cxlr->flags)) { + put_device(&cxlr->dev); + return; + } + + scoped_guard(device, port->uport_dev) { + if (port->uport_dev->driver) + devm_remove_action(port->uport_dev, unregister_region, cxlr); + } + + put_device(&cxlr->dev); +} + +static int remove_devm_actions(struct cxl_region *cxlr) +{ + if (!schedule_work(&cxlr->remove_work)) + return -EBUSY; + + return 0; +} + static ssize_t delete_region_store(struct device *dev, struct device_attribute *attr, const char *buf, size_t len) { struct cxl_root_decoder *cxlrd =3D to_cxl_root_decoder(dev); - struct cxl_port *port =3D to_cxl_port(dev->parent); struct cxl_region *cxlr; + int rc; =20 + /* remove_devm_actions_work() will put cxlr->dev. */ cxlr =3D cxl_find_region_by_name(cxlrd, buf); if (IS_ERR(cxlr)) return PTR_ERR(cxlr); =20 - devm_release_action(port->uport_dev, unregister_region, cxlr); - put_device(&cxlr->dev); + unregister_region(cxlr); + + rc =3D remove_devm_actions(cxlr); + if (rc) { + put_device(&cxlr->dev); + return rc; + } =20 return len; } diff --git a/drivers/cxl/cxl.h b/drivers/cxl/cxl.h index 1297594beaec..75ec292a9f42 100644 --- a/drivers/cxl/cxl.h +++ b/drivers/cxl/cxl.h @@ -447,6 +447,12 @@ struct cxl_region_params { */ #define CXL_REGION_F_NORMALIZED_ADDRESSING 3 =20 +/* Indicate that this region is being unregistered to prevent a race. */ +#define CXL_REGION_F_UNREGISTER 4 + +/* Indicate that this region called devm_remove_action. */ +#define CXL_REGION_F_DEVM_REMOVE 5 + /** * struct cxl_region - CXL region * @dev: This region's device @@ -462,6 +468,7 @@ struct cxl_region_params { * @coord: QoS access coordinates for the region * @node_notifier: notifier for setting the access coordinates to node * @adist_notifier: notifier for calculating the abstract distance of node + * @remove_work: trigger the remove action in a safe context to acquire lo= cks */ struct cxl_region { struct device dev; @@ -477,6 +484,7 @@ struct cxl_region { struct access_coordinate coord[ACCESS_COORDINATE_MAX]; struct notifier_block node_notifier; struct notifier_block adist_notifier; + struct work_struct remove_work; }; =20 struct cxl_nvdimm_bridge { --=20 2.47.3