Archive-only list for patches
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	patches@lists.linux.dev, Tudor Ambarus <tudor.ambarus@linaro.org>,
	Krzysztof Kozlowski <krzk@kernel.org>,
	Sasha Levin <sashal@kernel.org>
Subject: [PATCH 7.0 08/49] firmware: samsung: acpm: Fix missing LKMM barriers in sequence allocator
Date: Thu, 25 Jun 2026 14:03:20 +0100	[thread overview]
Message-ID: <20260625125638.666531247@linuxfoundation.org> (raw)
In-Reply-To: <20260625125637.527552689@linuxfoundation.org>

7.0-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Tudor Ambarus <tudor.ambarus@linaro.org>

[ Upstream commit bf296f83a3ddab1ab875edc4e8862cb10553064f ]

Sashiko identified memory ordering races in [1].

The ACPM driver uses a globally shared 'bitmap_seqnum' to track
available sequence numbers. Even though threads now strictly free their
own sequence numbers, the allocation and freeing of these bits across
concurrent threads are effectively lockless operations and require
explicit LKMM memory barriers.

Previously, the driver used plain bitwise operators (test_bit, set_bit,
clear_bit), which lack ordering guarantees. This creates two race
conditions on weakly ordered architectures like ARM64:

1. Polling Release Violation: The polling thread copies its payload and
   calls clear_bit(). Without a release barrier, the CPU can reorder
   the memory operations, making the cleared bit globally visible
   before the payload reads have fully completed.
2. TX Acquire Violation: The TX thread loops on test_bit(), calls
   set_bit(), and then wipes the payload buffer via memset(). Without
   an acquire barrier, the CPU can speculatively execute the memset()
   before the bit is safely and formally claimed.

If these reorderings overlap, a new TX thread can claim the sequence
number and overwrite the buffer while the original polling thread is
still actively reading from it.

Fix this by upgrading the bitwise operators. Wrap the TX allocation in
test_and_set_bit_lock() to establish formal LKMM Acquire semantics, and
pair it with clear_bit_unlock() in the polling path to enforce Release
semantics.

Cc: stable@vger.kernel.org
Fixes: a88927b534ba ("firmware: add Exynos ACPM protocol driver")
Closes: https://sashiko.dev/#/patchset/20260423-acpm-fixes-sashiko-reports-v1-0-2217b790925e%40linaro.org [1]
Signed-off-by: Tudor Ambarus <tudor.ambarus@linaro.org>
Link: https://patch.msgid.link/20260505-acpm-fixes-sashiko-reports-v5-6-43b5ee7f1674@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 drivers/firmware/samsung/exynos-acpm.c |   18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)

--- a/drivers/firmware/samsung/exynos-acpm.c
+++ b/drivers/firmware/samsung/exynos-acpm.c
@@ -7,7 +7,7 @@
 
 #include <linux/bitfield.h>
 #include <linux/bitmap.h>
-#include <linux/bits.h>
+#include <linux/bitops.h>
 #include <linux/cleanup.h>
 #include <linux/container_of.h>
 #include <linux/delay.h>
@@ -340,7 +340,7 @@ static int acpm_dequeue_by_polling(struc
 				acpm_get_saved_rx(achan, xfer, seqnum);
 
 			/* Relinquish ownership of the sequence slot */
-			clear_bit(seqnum - 1, achan->bitmap_seqnum);
+			clear_bit_unlock(seqnum - 1, achan->bitmap_seqnum);
 			return 0;
 		}
 
@@ -397,11 +397,18 @@ static void acpm_prepare_xfer(struct acp
 	struct acpm_rx_data *rx_data;
 	u32 *txd = (u32 *)xfer->txd;
 
-	/* Prevent chan->seqnum from being re-used */
+	/*
+	 * Prevent chan->seqnum from being re-used.
+	 * test_and_set_bit_lock() provides formal LKMM Acquire semantics.
+	 * It pairs with the RX thread's clear_bit_unlock() to ensure the CPU
+	 * does not speculatively execute the rx_data buffer wipe (memset)
+	 * before the sequence number is safely claimed.
+	 */
 	do {
 		if (++achan->seqnum == ACPM_SEQNUM_MAX)
 			achan->seqnum = 1;
-	} while (test_bit(achan->seqnum - 1, achan->bitmap_seqnum));
+		/* Flag the index based on seqnum. (seqnum: 1~63, bitmap: 0~62) */
+	} while (test_and_set_bit_lock(achan->seqnum - 1, achan->bitmap_seqnum));
 
 	txd[0] |= FIELD_PREP(ACPM_PROTOCOL_SEQNUM, achan->seqnum);
 
@@ -411,9 +418,6 @@ static void acpm_prepare_xfer(struct acp
 	memset(rx_data->cmd, 0, sizeof(*rx_data->cmd) * rx_data->n_cmd);
 	/* zero means no response expected */
 	rx_data->rxcnt = xfer->rxcnt;
-
-	/* Flag the index based on seqnum. (seqnum: 1~63, bitmap: 0~62) */
-	set_bit(achan->seqnum - 1, achan->bitmap_seqnum);
 }
 
 /**



  parent reply	other threads:[~2026-06-25 13:09 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-25 13:03 [PATCH 7.0 00/49] 7.0.14-rc1 review Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 01/49] io_uring/net: Avoid msghdr on op_connect/op_bind async data Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 02/49] arm64/entry: Fix arm64-specific rseq brokenness Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 03/49] lockd: fix TEST handling when not all permissions are available Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 04/49] firmware: exynos-acpm: Count number of commands in acpm_xfer Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 05/49] firmware: exynos-acpm: Count acpm_xfer buffers with __counted_by_ptr Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 06/49] firmware: samsung: acpm: Fix cross-thread RX length corruption Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 07/49] firmware: samsung: acpm: Fix false timeouts and Use-After-Free in polling Greg Kroah-Hartman
2026-06-25 13:03 ` Greg Kroah-Hartman [this message]
2026-06-25 13:03 ` [PATCH 7.0 09/49] fuse: re-lock request before replacing page cache folio Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 10/49] Revert "NFSD: Defer sub-object cleanup in export put callbacks" Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 11/49] RDMA/bnxt_re: zero shared page before exposing to userspace Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 12/49] bpf: Fix NULL pointer dereference in bpf_sk_storage_clone and diag paths Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 13/49] i2c: stub: Reject I2C block transfers with invalid length Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 14/49] net: qualcomm: rmnet: fix endpoint use-after-free in rmnet_dellink() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 15/49] agp/amd64: Fix broken error propagation in agp_amd64_probe() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 16/49] rose: fix dev_put() leak in rose_loopback_timer() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 17/49] rose: hold loopback neighbour reference across timer callback Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 18/49] rose: fix race between loopback timer and module removal Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 19/49] rose: clear neighbour pointer after rose_neigh_put() in state machines Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 20/49] rose: guard rose_neigh_put() against NULL in timer expiry Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 21/49] rose: fix netdev double-hold in rose_rx_call_request() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 22/49] rose: fix notifier unregistered too early in rose_exit() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 23/49] rose: set SOCK_DESTROY in rose_kill_by_device() for prompt cleanup Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 24/49] rose: disconnect orphaned STATE_2 sockets when device is gone Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 25/49] rose: fix netdev double-hold in rose_make_new() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 26/49] rose: release netdev ref and destroy orphaned incoming sockets Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 27/49] rose: drop CALL_REQUEST in loopback timer when device is not running Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 28/49] rose: cancel neighbour timers in rose_neigh_put() before freeing Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 29/49] rose: clear neighbour pointer in rose_kill_by_device() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 30/49] rose: dont free fd-owned sockets when reaping in the heartbeat Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 31/49] net: export netif_open for self_test usage Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 32/49] net: net_failover: Fix the deadlock in slave register Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 33/49] iio: light: veml6075: add bounds check to veml6075_it_ms index Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 34/49] iio: adc: ti-ads1298: add bounds check to pga_settings index Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 35/49] Input: rmi4 - fix register descriptor address calculation Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 36/49] Input: rmi4 - refactor register descriptor parsing Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 37/49] Input: rmi4 - fix type overflow in register counts Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 38/49] Input: rmi4 - fix num_subpackets overflow in register descriptor Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 39/49] Input: rmi4 - fix memory leak in rmi_set_attn_data() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 40/49] Input: rmi4 - iterative IRQ handler Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 41/49] Input: rmi4 - fix bit count in bitmap_copy() Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 42/49] crypto: qat - remove unused character device and IOCTLs Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 43/49] vc_screen: fix null-ptr-deref in vcs_notifier() during concurrent vcs_write Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 44/49] serial: qcom_geni: Fix RX DMA stall when SE_DMA_RX_LEN_IN is zero Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 45/49] serial: 8250_dw: unregister 8250 port if clk_notifier_register() fails Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 46/49] drivers/base/memory: set mem->altmap after successful device registration Greg Kroah-Hartman
2026-06-25 13:03 ` [PATCH 7.0 47/49] ksmbd: reject non-VALID session in compound request branch Greg Kroah-Hartman
2026-06-25 13:04 ` [PATCH 7.0 48/49] media: vidtv: fix NULL pointer dereference in vidtv_mux_push_si Greg Kroah-Hartman
2026-06-25 13:04 ` [PATCH 7.0 49/49] virtiofs: fix UAF on submount umount Greg Kroah-Hartman
2026-06-25 13:44 ` [PATCH 7.0 00/49] 7.0.14-rc1 review Florian Fainelli
2026-06-25 15:27 ` Brett A C Sheffield
2026-06-25 17:30 ` Justin Forbes
2026-06-25 18:06 ` Peter Schneider
2026-06-25 23:58 ` Shuah Khan
2026-06-26  5:05 ` Ron Economos

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=20260625125638.666531247@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=krzk@kernel.org \
    --cc=patches@lists.linux.dev \
    --cc=sashal@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=tudor.ambarus@linaro.org \
    /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