From nobody Fri Oct 2 12:25:02 2026 Received: from mail-pj1-f52.google.com (mail-pj1-f52.google.com [209.85.216.52]) (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 7BF5F37E5C2 for ; Fri, 31 Jul 2026 23:06:13 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.216.52 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1785539174; cv=none; b=nLOzj+WRFqPhZvpvqkUw8eh3tTAwHdUL2Mqh70sdYOhT9KQEJ3YOgDujWnXoT5LazRCLoBUK+3ptuPdCwVwEZ7vMm5Nyr6rOvlR+ZxD5xjmkcHyMwhckKkVUcJzCnsSYwY4RTBly19FtD6OWNGTsEjDWlzsQi3kFqNm2dVKpU3k= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1785539174; c=relaxed/simple; bh=vTS5fABI4Si4pXLP98k+xMuVt7ofecoAxFjwM9il9dg=; h=From:To:Cc:Subject:Date:Message-ID:MIME-Version; b=mZiysvniZ5Eb6MN9ZLo8UA/CLJtX72uXX19Qn0yW/Y58O2az0V5s+KkEOxIUdB3mAsNZDDbRHG1vFW2hHHANqogo/Lb4AHSyJGuSKC9VSg9Dtr9JyBWGHRZbceWUnUydgD8A0OT9PP+YjwommIyFVzAzV03XbOdr1cDaGKtuaEk= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=UhFnMk4Z; arc=none smtp.client-ip=209.85.216.52 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="UhFnMk4Z" Received: by mail-pj1-f52.google.com with SMTP id 98e67ed59e1d1-38101f85591so2054719a91.1 for ; Fri, 31 Jul 2026 16:06:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20251104; t=1785539173; x=1786143973; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:from:to:cc:subject:date:message-id:reply-to:content-type; bh=t/+wptgQI/4wuhlfg1bZeWkE5qmDQxqgmBrLggWs8oE=; b=UhFnMk4ZAwyTROKO9JFbDUGVmh+H55SAR6+oDgQm3On2oi+KsmA4Tw988gphamCvCN vujBlMlqG6eiYY3XhL2pW1l3a4Bf5cTndZWoopUOOy5+12noEZ83GIwRUu1mBvT5x3kk 3hcPH5IxubX36nTL6tLdPiXb+2oZP/BT/JksmleOBMQ5Loq+2EROVy0JYDap/Cqv5PxU KTJiq5tSjo/v06wz+SQ3B9o36bQuZ1X0bb0yJdjV+J9SS9roEzSke+scZR7Evx3ZO3XL D4ngF3HFL6rDUeAbgtFpaLxtj3rB/LgCtexVNCszcAP1rMitVms4SD+8ibzhW0JPajQJ QNMA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1785539173; x=1786143973; 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:content-type; bh=t/+wptgQI/4wuhlfg1bZeWkE5qmDQxqgmBrLggWs8oE=; b=TK3NI40hU1ARKtzWtDjmJ01F3Sw8qkZP3D1ECtMowqkLdATmpplOElp842X2U4XQj/ xlBwbFXUVoITwYyjw5+cH1XryEnTl0yINDDlsOxqn+S/Bd5r3xrQprnE2dQ04MonQQWo FEQnSFmW1d/OFdZInRS1FkYNXcBmE04GxPrOxpD4zIK99uDUorOQxdtufdGHPnUuInvH eIakb1hXit9HpI4opmPyzJ4ByO+V/YEb4OI+VKpVPjfrOqRDBm33KDj0dr345ytXRJ09 ydDVPY2Ok8lH0jHBwoJ4FLs82geM6bPq6ZZS3WdF012a8rkVNIt9NY3o6vuTom+U1Pmk HrVA== X-Forwarded-Encrypted: i=1; AHgh+Rpd9K5eK5LiUnK54eWbxpWot0tbwPXx4eHJVqqaSNQLgR17SdQfJGfpbeaoFIrDhNpq33utyVg0fDkxtxc=@vger.kernel.org X-Gm-Message-State: AOJu0YxFz7hmgTVJGuNNa+KADQneNY/p8q2XlIEDmgLLPppT868SwtuG WVBCR+DAXv4kIX2HbWf0tjuBoX1m+QGsnexkp7cFhBHUewthiociiHEQ X-Gm-Gg: AR+sD10CIc1BF0sM5hxlPsIeIa1EBxTKwI6Xide0mhzidQ+2v4+UXUue5ZErSF7XJJp RQO0NBJuzGMgR5IhP9t71shiqE9dsVULBssMNFA/ZXd/1BQvkFvhf8CARpNpEmNLomnRWKdjrVx SLAgUMP1GAFWnNLUidstRXXbKkki+J3lUcjvq31H4+MZ95CG0jzgf1nkAYVHP8okj5pX+GNMZe+ szflKH92KmQsc45/h2Lr6PYy54b3OydDzkEgZR9q6V0HzvjmFl2d6fT6gQRrueCIQGZAGmCHX4o vdvrTZ7Q4KuFMOxfalvJX1DchjPv+qxIVaaq8w3XSacfgZcp5HeMiy7UmkvAvMRvJO/Or9JF2OP t/tCfeU9xLd4QOpAWrXMKoqCUF5ECCRTTFzM8i3JGdBVCXfP1Nd1vb5zjPBOsu9lYj3Oxzms4Z2 1Wpbx6smGo7AUwrTDWKQitCjWTmpClVjJUXfC1qRyxS7ac299RLDAO6KUgVS/IdUOubsFP0+l3E eRV/fe6uyVo50pOHY0gCvIw/34o+NgJ0v6L0wC/nIE= X-Received: by 2002:a17:90b:48c9:b0:369:7491:7b24 with SMTP id 98e67ed59e1d1-38fb2441d4fmr3449982a91.6.1785539172700; Fri, 31 Jul 2026 16:06:12 -0700 (PDT) Received: from deepanshu-kernel-hacker.. ([2405:201:682f:383f:d0a0:b78b:9d8:d130]) by smtp.gmail.com with ESMTPSA id 5a478bee46e88-3153ddda5d9sm12756358eec.13.2026.07.31.16.06.07 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 31 Jul 2026 16:06:11 -0700 (PDT) From: Deepanshu Kartikey To: gregkh@linuxfoundation.org Cc: viro@zeniv.linux.org.uk, jack@suse.cz, mjguzik@gmail.com, brauner@kernel.org, kees@kernel.org, linux-usb@vger.kernel.org, linux-kernel@vger.kernel.org, Deepanshu Kartikey , syzbot+df9e891bf8ea586f846b@syzkaller.appspotmail.com Subject: [PATCH v2] usb: gadgetfs: fix use-after-free in ep_open() Date: Sat, 1 Aug 2026 04:36:03 +0530 Message-ID: <20260731230603.4637-1-kartikey406@gmail.com> X-Mailer: git-send-email 2.43.0 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" ep_open() dereferenced inode->i_private and locked its embedded mutex before checking whether the underlying ep_data was still alive. destroy_ep_files(), called from gadgetfs_unbind(), can free that same ep_data concurrently, since nothing prevented an in-flight open() from racing the teardown. KASAN reports a slab-use-after-free in ep_open(), with the object freed by a concurrent gadgetfs_unbind() -> destroy_ep_files() -> put_ep() -> kfree(). Fix it by adding an inode back-pointer to ep_data, having destroy_ep_files() clear inode->i_private to NULL under dev->lock before the object can be freed, and having ep_open() read inode->i_private and pin the result with get_ep() under that same lock, bailing out with -ENOENT if it finds NULL. This makes liveness self-evident at the point of use instead of being inferred from device state, so it holds regardless of which path frees the endpoint. dev is derived from inode->i_sb->s_fs_info, set in gadgetfs_fill_super(), rather than from the file's the_device singleton, so ep_open() is correctly scoped to the specific mount the inode belongs to. gadgetfs_create_file() now returns the created inode (or an ERR_PTR) directly instead of taking an output parameter, so activate_ep_files() can store it in data->inode without an unexplained NULL at call sites that don't need it. Reported-by: syzbot+df9e891bf8ea586f846b@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=3Ddf9e891bf8ea586f846b Fixes: 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 ("Linux-2.6.12-rc2") Tested-by: syzbot+df9e891bf8ea586f846b@syzkaller.appspotmail.com Signed-off-by: Deepanshu Kartikey --- v2: - Added Fixes: tag; git blame traces the bug to the 2005 git import boundary, confirmed no later commit changed the ordering - Derive dev from inode->i_sb->s_fs_info instead of the new the_device singleton, so the lookup is scoped to the inode's own mount instead of a global (Greg) - gadgetfs_create_file() now returns struct inode * / ERR_PTR instead of taking an inode_out output parameter, so call sites don't need an unexplained NULL argument (Greg) --- drivers/usb/gadget/legacy/inode.c | 61 ++++++++++++++++++++----------- 1 file changed, 39 insertions(+), 22 deletions(-) diff --git a/drivers/usb/gadget/legacy/inode.c b/drivers/usb/gadget/legacy/= inode.c index d87a8ab51510..e5a316777db2 100644 --- a/drivers/usb/gadget/legacy/inode.c +++ b/drivers/usb/gadget/legacy/inode.c @@ -207,6 +207,7 @@ struct ep_data { struct usb_endpoint_descriptor desc, hs_desc; struct list_head epfiles; wait_queue_head_t wait; + struct inode *inode; }; =20 static inline void get_ep (struct ep_data *data) @@ -235,6 +236,7 @@ static void put_ep (struct ep_data *data) =20 static const char *CHIP; static DEFINE_MUTEX(sb_mutex); /* Serialize superblock operations */ +static struct dev_data *the_device; =20 /*----------------------------------------------------------------------*/ =20 @@ -817,25 +819,39 @@ ep_config (struct ep_data *data, const char *buf, siz= e_t len) static int ep_open (struct inode *inode, struct file *fd) { - struct ep_data *data =3D inode->i_private; - int value =3D -EBUSY; + struct dev_data *dev =3D inode->i_sb->s_fs_info; + struct ep_data *data; + int value =3D -ENODEV; + + spin_lock_irq(&dev->lock); + data =3D inode->i_private; + if (data =3D=3D NULL) { + spin_unlock_irq(&dev->lock); + return -ENOENT; + } + get_ep(data); + spin_unlock_irq(&dev->lock); =20 - if (mutex_lock_interruptible(&data->lock) !=3D 0) + if (mutex_lock_interruptible(&data->lock) !=3D 0) { + put_ep(data); return -EINTR; - spin_lock_irq (&data->dev->lock); - if (data->dev->state =3D=3D STATE_DEV_UNBOUND) + } + + value =3D -EBUSY; + spin_lock_irq(&dev->lock); + if (dev->state =3D=3D STATE_DEV_UNBOUND) value =3D -ENOENT; else if (data->state =3D=3D STATE_EP_DISABLED) { value =3D 0; data->state =3D STATE_EP_READY; - get_ep (data); fd->private_data =3D data; - VDEBUG (data->dev, "%s ready\n", data->name); + VDEBUG(dev, "%s ready\n", data->name); } else - DBG (data->dev, "%s state %d\n", - data->name, data->state); - spin_unlock_irq (&data->dev->lock); + DBG(dev, "%s state %d\n", data->name, data->state); + spin_unlock_irq(&dev->lock); mutex_unlock(&data->lock); + if (value) + put_ep(data); return value; } =20 @@ -1563,6 +1579,7 @@ static void destroy_ep_files (struct dev_data *dev) /* break link to FS */ ep =3D list_first_entry (&dev->epfiles, struct ep_data, epfiles); list_del_init (&ep->epfiles); + ep->inode->i_private =3D NULL; spin_unlock_irq (&dev->lock); =20 /* break link to controller */ @@ -1587,14 +1604,13 @@ static void destroy_ep_files (struct dev_data *dev) } =20 =20 -static int gadgetfs_create_file (struct super_block *sb, char const *name, +static struct inode *gadgetfs_create_file(struct super_block *sb, char con= st *name, void *data, const struct file_operations *fops); =20 static int activate_ep_files (struct dev_data *dev) { struct usb_ep *ep; struct ep_data *data; - int err; =20 gadget_for_each_ep (ep, dev->gadget) { =20 @@ -1617,9 +1633,9 @@ static int activate_ep_files (struct dev_data *dev) if (!data->req) goto enomem1; =20 - err =3D gadgetfs_create_file (dev->sb, data->name, + data->inode =3D gadgetfs_create_file(dev->sb, data->name, data, &ep_io_operations); - if (err) + if (IS_ERR(data->inode)) goto enomem2; list_add_tail (&data->epfiles, &dev->epfiles); } @@ -1663,8 +1679,6 @@ gadgetfs_unbind (struct usb_gadget *gadget) put_dev (dev); } =20 -static struct dev_data *the_device; - static int gadgetfs_bind(struct usb_gadget *gadget, struct usb_gadget_driver *driver) { @@ -1983,7 +1997,7 @@ gadgetfs_make_inode (struct super_block *sb, /* creates in fs root directory, so non-renamable and non-linkable. * so inode and dentry are paired, until device reconfig. */ -static int gadgetfs_create_file (struct super_block *sb, char const *name, +static struct inode *gadgetfs_create_file(struct super_block *sb, char con= st *name, void *data, const struct file_operations *fops) { struct dentry *dentry; @@ -1992,18 +2006,18 @@ static int gadgetfs_create_file (struct super_block= *sb, char const *name, inode =3D gadgetfs_make_inode (sb, data, fops, S_IFREG | (default_perm & S_IRWXUGO)); if (!inode) - return -ENOMEM; + return ERR_PTR(-ENOMEM); =20 dentry =3D simple_start_creating(sb->s_root, name); if (IS_ERR(dentry)) { iput(inode); - return PTR_ERR(dentry); + return ERR_CAST(dentry); } =20 d_make_persistent(dentry, inode); =20 simple_done_creating(dentry); - return 0; + return inode; } =20 static const struct super_operations gadget_fs_operations =3D { @@ -2056,8 +2070,11 @@ gadgetfs_fill_super (struct super_block *sb, struct = fs_context *fc) goto Enomem; =20 dev->sb =3D sb; - rc =3D gadgetfs_create_file(sb, CHIP, dev, &ep0_operations); - if (rc) { + sb->s_fs_info =3D dev; + inode =3D gadgetfs_create_file(sb, CHIP, dev, &ep0_operations); + if (IS_ERR(inode)) { + rc =3D PTR_ERR(inode); + sb->s_fs_info =3D NULL; put_dev(dev); goto Enomem; } --=20 2.43.0