From: Stanley Jhu <stanleyjhu@google.com>
To: "Martin K . Petersen" <mkp@kernel.org>,
"James E . J . Bottomley"
<James.Bottomley@HansenPartnership.com>,
linux-scsi@vger.kernel.org
Cc: Brian Kao <powenkao@google.com>, Bean Huo <beanhuo@micron.com>,
Bart Van Assche <bvanassche@acm.org>,
Alim Akhtar <alim.akhtar@samsung.com>,
Avri Altman <avri.altman@sandisk.com>,
Can Guo <can.guo@oss.qualcomm.com>,
Peter Wang <peter.wang@mediatek.com>,
linux-kernel@vger.kernel.org,
Stanley Jhu <stanleyjhu@google.com>
Subject: [PATCH 0/2] scsi: ufs: rpmb: Fix device lifetime UAF and unregistered bus regression
Date: Fri, 4 Sep 2026 19:00:15 +0800 [thread overview]
Message-ID: <20260904110017.3444852-1-stanleyjhu@google.com> (raw)
This series fixes two architectural defects in the UFS OP-TEE RPMB
driver: a use-after-free (UAF) during controller unbind or shutdown, and
an unregistered bus rejection by the driver core.
Patch 1: Decouple RPMB device lifecycle from devres
- Problem: struct ufs_rpmb_dev embeds a struct device but was allocated
via devres tied to the parent controller. When the controller unbinds,
devres prematurely frees the structure while external references
remain, causing a UAF Oops on SLUB-poisoned memory
(0x006b6b6b6b6b6b9c). Additionally, in-flight I/O lacks transport
pinning to the underlying SCSI device.
- Solution: Transition to a reference-counted device lifecycle, pin the
underlying SCSI device and host hierarchy during probe, and enforce
proper subsystem unregistration and offline state validation.
Patch 2: Register and unregister ufs_rpmb_bus_type
- Problem: The driver assigns devices to ufs_rpmb_bus_type without
registering the bus with the driver core, causing device_register()
to fail with -EINVAL during probe.
- Solution: Register the bus during UFS core initialization and
unregister it on exit, using IS_REACHABLE(CONFIG_RPMB) to provide clean
fallback stubs and prevent modular build breakages.
Verification:
- QEMU ARM64 with virtual UFS 4.0 PCI controller:
* Verified /sys/bus/ufs_rpmb registration and successful device probe.
* Verified probe, unbind, and shutdown complete without leaks or Oops.
- checkpatch.pl: 0 errors, 0 warnings across all patches.
Stanley Jhu (2):
scsi: ufs: rpmb: Decouple device lifecycle from devres to avoid UAF
scsi: ufs: rpmb: Register and unregister ufs_rpmb_bus_type
drivers/ufs/core/ufs-rpmb.c | 96 ++++++++++++++++++++------------
drivers/ufs/core/ufshcd-priv.h | 11 ++++-
drivers/ufs/core/ufshcd.c | 16 ++++++-
3 files changed, 84 insertions(+), 39 deletions(-)
--
2.55.0.979.g7e5102b832-goog
next reply other threads:[~2026-09-04 11:00 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-04 11:00 Stanley Jhu [this message]
2026-09-04 11:00 ` [PATCH 1/2] scsi: ufs: rpmb: Decouple device lifecycle from devres to avoid UAF Stanley Jhu
2026-09-04 11:14 ` sashiko-bot
2026-09-04 11:00 ` [PATCH 2/2] scsi: ufs: rpmb: Register and unregister ufs_rpmb_bus_type Stanley Jhu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260904110017.3444852-1-stanleyjhu@google.com \
--to=stanleyjhu@google.com \
--cc=James.Bottomley@HansenPartnership.com \
--cc=alim.akhtar@samsung.com \
--cc=avri.altman@sandisk.com \
--cc=beanhuo@micron.com \
--cc=bvanassche@acm.org \
--cc=can.guo@oss.qualcomm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=mkp@kernel.org \
--cc=peter.wang@mediatek.com \
--cc=powenkao@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox