mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Weibin Liu <liuwb@xiaopeng.com>
To: linux-scsi@vger.kernel.org,
	suganath-prabu.subramani@broadcom.com,
	sreekanth.reddy@broadcom.com, ranjan.kumar@broadcom.com,
	sathya.prakash@broadcom.com
Cc: linux-kernel@vger.kernel.org,
	James.Bottomley@HansenPartnership.com,
	MPT-FusionLinux.pdl@broadcom.com, mkp@kernel.org,
	stable@vger.kernel.org, zhugl3@xiaopeng.com
Subject: [PATCH v2] scsi: mpt3sas: validate data_out size before SMP link-reset check
Date: Thu, 29 Oct 2026 15:33:16 +0800	[thread overview]
Message-ID: <20261029073316.4188569-1-liuwb@xiaopeng.com> (raw)

_ctl_do_mpt_command() handles MPI2_FUNCTION_SMP_PASSTHROUGH by reading
data[1] and data[10] of the data_out buffer to detect SMP link-reset
requests. The non-IMMEDIATE path sets data = data_out after checking only
for a NULL pointer, so a payload shorter than 11 bytes can read past the
end of the buffer, and out-of-bounds bytes can unexpectedly set
ioc_link_reset_in_progress and ignore_loginfos.

However, a short SMP request is not invalid: SMP_REPORT_GENERAL, for
example, can use an 8-byte payload. Rejecting every data_out smaller than
11 bytes would break valid requests.

Instead, require enough data only when checking for a link reset, and
keep sending the request normally otherwise. The IMMEDIATE path reads
from the request frame rather than data_out, so data_out_sz does not
apply to it.

Fixes: f92363d12359 ("[SCSI] mpt3sas: add new driver supporting 12GB SAS")
Cc: stable@vger.kernel.org # 3.8+
Assisted-by: LLM
Reviewed-by: Guanglei Zhu <zhugl3@xiaopeng.com>
Signed-off-by: Weibin Liu <liuwb@xiaopeng.com>
---
Changes in v2:
- Do not reject valid short SMP requests. Apply the data_out size check
  only to link-reset detection.
- Keep IMMEDIATE handling independent of data_out_sz and use the SAS
  definitions instead of magic numbers.

Link: https://lore.kernel.org/all/20261015-mpt3sas-dataout-sz-v1-1-0781d24bb7ad@xiaopeng.com/

 drivers/scsi/mpt3sas/mpt3sas_ctl.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_ctl.c b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
index 8bb947004885..b069dd384ed7 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_ctl.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_ctl.c
@@ -57,9 +57,13 @@
 #include <linux/io.h>
 #include <linux/uaccess.h>
 
+#include <scsi/sas.h>
+
 #include "mpt3sas_base.h"
 #include "mpt3sas_ctl.h"
 
+/* PHY operation byte in an SMP PHY Control request */
+#define SMP_PHY_CONTROL_OP_OFFSET	10
 
 static struct fasync_struct *async_queue;
 static DECLARE_WAIT_QUEUE_HEAD(ctl_poll_wait);
@@ -959,13 +963,17 @@ _ctl_do_mpt_command(struct MPT3SAS_ADAPTER *ioc, struct mpt3_ioctl_command karg,
 		Mpi2SmpPassthroughRequest_t *smp_request =
 		    (Mpi2SmpPassthroughRequest_t *)mpi_request;
 		u8 *data;
+		bool immediate;
 
 		if (!ioc->multipath_on_hba) {
 			/* ioc determines which port to use */
 			smp_request->PhysicalPort = 0xFF;
 		}
-		if (smp_request->PassthroughFlags &
-		    MPI2_SMP_PT_REQ_PT_FLAGS_IMMEDIATE)
+
+		immediate = smp_request->PassthroughFlags &
+			    MPI2_SMP_PT_REQ_PT_FLAGS_IMMEDIATE;
+
+		if (immediate)
 			data = (u8 *)&smp_request->SGL;
 		else {
 			if (unlikely(data_out == NULL)) {
@@ -978,7 +986,11 @@ _ctl_do_mpt_command(struct MPT3SAS_ADAPTER *ioc, struct mpt3_ioctl_command karg,
 			data = data_out;
 		}
 
-		if (data[1] == 0x91 && (data[10] == 1 || data[10] == 2)) {
+		if ((immediate ||
+		     data_out_sz >= SMP_PHY_CONTROL_OP_OFFSET + 1) &&
+		    data[1] == SMP_PHY_CONTROL &&
+		    (data[10] == PHY_FUNC_LINK_RESET ||
+		     data[10] == PHY_FUNC_HARD_RESET)) {
 			ioc->ioc_link_reset_in_progress = 1;
 			ioc->ignore_loginfos = 1;
 		}
-- 
2.50.1


                 reply	other threads:[~2026-10-29  7:33 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20261029073316.4188569-1-liuwb@xiaopeng.com \
    --to=liuwb@xiaopeng.com \
    --cc=James.Bottomley@HansenPartnership.com \
    --cc=MPT-FusionLinux.pdl@broadcom.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=mkp@kernel.org \
    --cc=ranjan.kumar@broadcom.com \
    --cc=sathya.prakash@broadcom.com \
    --cc=sreekanth.reddy@broadcom.com \
    --cc=stable@vger.kernel.org \
    --cc=suganath-prabu.subramani@broadcom.com \
    --cc=zhugl3@xiaopeng.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

all inboxes | Powered by JetHome®