From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: "Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
patches@lists.linux.dev, "Michal Rábek" <mrabek@redhat.com>,
"Tomas Henzl" <thenzl@redhat.com>,
"Changhui Zhong" <czhong@redhat.com>,
"Ewan D. Milne" <emilne@redhat.com>,
"John Meneghini" <jmeneghi@redhat.com>,
"Martin K. Petersen" <martin.petersen@oracle.com>,
"Sasha Levin" <sashal@kernel.org>
Subject: [PATCH 6.1 71/72] scsi: sg: Fix occasional bogus elapsed time that exceeds timeout
Date: Thu, 15 Jan 2026 17:49:21 +0100 [thread overview]
Message-ID: <20260115164146.078671547@linuxfoundation.org> (raw)
In-Reply-To: <20260115164143.482647486@linuxfoundation.org>
6.1-stable review patch. If anyone has any objections, please let me know.
------------------
From: Michal Rábek <mrabek@redhat.com>
[ Upstream commit 0e1677654259a2f3ccf728de1edde922a3c4ba57 ]
A race condition was found in sg_proc_debug_helper(). It was observed on
a system using an IBM LTO-9 SAS Tape Drive (ULTRIUM-TD9) and monitoring
/proc/scsi/sg/debug every second. A very large elapsed time would
sometimes appear. This is caused by two race conditions.
We reproduced the issue with an IBM ULTRIUM-HH9 tape drive on an x86_64
architecture. A patched kernel was built, and the race condition could
not be observed anymore after the application of this patch. A
reproducer C program utilising the scsi_debug module was also built by
Changhui Zhong and can be viewed here:
https://github.com/MichaelRabek/linux-tests/blob/master/drivers/scsi/sg/sg_race_trigger.c
The first race happens between the reading of hp->duration in
sg_proc_debug_helper() and request completion in sg_rq_end_io(). The
hp->duration member variable may hold either of two types of
information:
#1 - The start time of the request. This value is present while
the request is not yet finished.
#2 - The total execution time of the request (end_time - start_time).
If sg_proc_debug_helper() executes *after* the value of hp->duration was
changed from #1 to #2, but *before* srp->done is set to 1 in
sg_rq_end_io(), a fresh timestamp is taken in the else branch, and the
elapsed time (value type #2) is subtracted from a timestamp, which
cannot yield a valid elapsed time (which is a type #2 value as well).
To fix this issue, the value of hp->duration must change under the
protection of the sfp->rq_list_lock in sg_rq_end_io(). Since
sg_proc_debug_helper() takes this read lock, the change to srp->done and
srp->header.duration will happen atomically from the perspective of
sg_proc_debug_helper() and the race condition is thus eliminated.
The second race condition happens between sg_proc_debug_helper() and
sg_new_write(). Even though hp->duration is set to the current time
stamp in sg_add_request() under the write lock's protection, it gets
overwritten by a call to get_sg_io_hdr(), which calls copy_from_user()
to copy struct sg_io_hdr from userspace into kernel space. hp->duration
is set to the start time again in sg_common_write(). If
sg_proc_debug_helper() is called between these two calls, an arbitrary
value set by userspace (usually zero) is used to compute the elapsed
time.
To fix this issue, hp->duration must be set to the current timestamp
again after get_sg_io_hdr() returns successfully. A small race window
still exists between get_sg_io_hdr() and setting hp->duration, but this
window is only a few instructions wide and does not result in observable
issues in practice, as confirmed by testing.
Additionally, we fix the format specifier from %d to %u for printing
unsigned int values in sg_proc_debug_helper().
Signed-off-by: Michal Rábek <mrabek@redhat.com>
Suggested-by: Tomas Henzl <thenzl@redhat.com>
Tested-by: Changhui Zhong <czhong@redhat.com>
Reviewed-by: Ewan D. Milne <emilne@redhat.com>
Reviewed-by: John Meneghini <jmeneghi@redhat.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Link: https://patch.msgid.link/20251212160900.64924-1-mrabek@redhat.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/scsi/sg.c | 20 +++++++++++++-------
1 file changed, 13 insertions(+), 7 deletions(-)
diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index 7a5d73f89f459..b63f7c09c97a1 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -735,6 +735,8 @@ sg_new_write(Sg_fd *sfp, struct file *file, const char __user *buf,
sg_remove_request(sfp, srp);
return -EFAULT;
}
+ hp->duration = jiffies_to_msecs(jiffies);
+
if (hp->interface_id != 'S') {
sg_remove_request(sfp, srp);
return -ENOSYS;
@@ -819,7 +821,6 @@ sg_common_write(Sg_fd * sfp, Sg_request * srp,
return -ENODEV;
}
- hp->duration = jiffies_to_msecs(jiffies);
if (hp->interface_id != '\0' && /* v3 (or later) interface */
(SG_FLAG_Q_AT_TAIL & hp->flags))
at_head = 0;
@@ -1342,9 +1343,6 @@ sg_rq_end_io(struct request *rq, blk_status_t status)
"sg_cmd_done: pack_id=%d, res=0x%x\n",
srp->header.pack_id, result));
srp->header.resid = resid;
- ms = jiffies_to_msecs(jiffies);
- srp->header.duration = (ms > srp->header.duration) ?
- (ms - srp->header.duration) : 0;
if (0 != result) {
struct scsi_sense_hdr sshdr;
@@ -1393,6 +1391,9 @@ sg_rq_end_io(struct request *rq, blk_status_t status)
done = 0;
}
srp->done = done;
+ ms = jiffies_to_msecs(jiffies);
+ srp->header.duration = (ms > srp->header.duration) ?
+ (ms - srp->header.duration) : 0;
write_unlock_irqrestore(&sfp->rq_list_lock, iflags);
if (likely(done)) {
@@ -2529,6 +2530,7 @@ static void sg_proc_debug_helper(struct seq_file *s, Sg_device * sdp)
const sg_io_hdr_t *hp;
const char * cp;
unsigned int ms;
+ unsigned int duration;
k = 0;
list_for_each_entry(fp, &sdp->sfds, sfd_siblings) {
@@ -2566,13 +2568,17 @@ static void sg_proc_debug_helper(struct seq_file *s, Sg_device * sdp)
seq_printf(s, " id=%d blen=%d",
srp->header.pack_id, blen);
if (srp->done)
- seq_printf(s, " dur=%d", hp->duration);
+ seq_printf(s, " dur=%u", hp->duration);
else {
ms = jiffies_to_msecs(jiffies);
- seq_printf(s, " t_o/elap=%d/%d",
+ duration = READ_ONCE(hp->duration);
+ if (duration)
+ duration = (ms > duration ?
+ ms - duration : 0);
+ seq_printf(s, " t_o/elap=%u/%u",
(new_interface ? hp->timeout :
jiffies_to_msecs(fp->timeout)),
- (ms > hp->duration ? ms - hp->duration : 0));
+ duration);
}
seq_printf(s, "ms sgat=%d op=0x%02x\n", usg,
(int) srp->data.cmd_opcode);
--
2.51.0
next prev parent reply other threads:[~2026-01-15 17:13 UTC|newest]
Thread overview: 85+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-01-15 16:48 [PATCH 6.1 00/72] 6.1.161-rc1 review Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 01/72] atm: Fix dma_free_coherent() size Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 02/72] net: 3com: 3c59x: fix possible null dereference in vortex_probe1() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 03/72] btrfs: always detect conflicting inodes when logging inode refs Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 04/72] mei: me: add nova lake point S DID Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 05/72] lib/crypto: aes: Fix missing MMU protection for AES S-box Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 06/72] counter: interrupt-cnt: Drop IRQF_NO_THREAD flag Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 07/72] drm/pl111: Fix error handling in pl111_amba_probe Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 08/72] gpio: rockchip: mark the GPIO controller as sleeping Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 09/72] wifi: avoid kernel-infoleak from struct iw_point Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 10/72] libceph: prevent potential out-of-bounds reads in handle_auth_done() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 11/72] libceph: replace overzealous BUG_ON in osdmap_apply_incremental() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 12/72] libceph: make free_choose_arg_map() resilient to partial allocation Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 13/72] libceph: return the handler error from mon_handle_auth_done() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 14/72] libceph: make calc_target() set t->paused, not just clear it Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 15/72] ext4: introduce ITAIL helper Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 16/72] ext4: fix out-of-bound read in ext4_xattr_inode_dec_ref_all() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 17/72] net: Add locking to protect skb->dev access in ip_output Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 18/72] tls: Use __sk_dst_get() and dst_dev_rcu() in get_netdev_for_sock() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 19/72] csky: fix csky_cmpxchg_fixup not working Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 20/72] ARM: 9461/1: Disable HIGHPTE on PREEMPT_RT kernels Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 21/72] alpha: dont reference obsolete termio struct for TC* constants Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 22/72] NFSv4: ensure the open stateid seqid doesnt go backwards Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 23/72] NFS: Fix up the automount fs_context to use the correct cred Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 24/72] smb/client: fix NT_STATUS_UNABLE_TO_FREE_VM value Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 25/72] smb/client: fix NT_STATUS_DEVICE_DOOR_OPEN value Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 26/72] smb/client: fix NT_STATUS_NO_DATA_DETECTED value Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 27/72] scsi: ipr: Enable/disable IRQD_NO_BALANCING during reset Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 28/72] scsi: ufs: core: Fix EH failure after W-LUN resume error Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 29/72] scsi: Revert "scsi: libsas: Fix exp-attached device scan after probe failure scanned in again after probe failed" Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 30/72] arm64: dts: add off-on-delay-us for usdhc2 regulator Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 31/72] ARM: dts: imx6q-ba16: fix RTC interrupt level Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 32/72] arm64: dts: imx8mp: Fix LAN8740Ai PHY reference clock on DH electronics i.MX8M Plus DHCOM Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 33/72] netfilter: nft_synproxy: avoid possible data-race on update operation Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 34/72] netfilter: nf_tables: fix memory leak in nf_tables_newrule() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 35/72] netfilter: nf_conncount: update last_gc only when GC has been performed Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 36/72] net: marvell: prestera: fix NULL dereference on devlink_alloc() failure Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 37/72] bridge: fix C-VLAN preservation in 802.1ad vlan_tunnel egress Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 38/72] net: mscc: ocelot: Fix crash when adding interface under a lag Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 39/72] inet: ping: Fix icmp out counting Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 40/72] net: sock: fix hardened usercopy panic in sock_recv_errqueue Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 41/72] netdev: preserve NETIF_F_ALL_FOR_ALL across TSO updates Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 42/72] net/mlx5e: Dont print error message due to invalid module Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 43/72] net: wwan: iosm: Fix memory leak in ipc_mux_deinit() Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 44/72] eth: bnxt: move and rename reset helpers Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 45/72] bnxt_en: Fix potential data corruption with HW GRO/LRO Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 46/72] net: fix memory leak in skb_segment_list for GRO packets Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 47/72] HID: quirks: work around VID/PID conflict for appledisplay Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 48/72] net/sched: sch_qfq: Fix NULL deref when deactivating inactive aggregate in qfq_reset Greg Kroah-Hartman
2026-01-15 16:48 ` [PATCH 6.1 49/72] net: usb: pegasus: fix memory leak in update_eth_regs_async() Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 50/72] net: enetc: fix build warning when PAGE_SIZE is greater than 128K Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 51/72] arp: do not assume dev_hard_header() does not change skb->head Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 52/72] pinctrl: qcom: lpass-lpi: mark the GPIO controller as sleeping Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 53/72] mm/pagewalk: add walk_page_range_vma() Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 54/72] ksm: use range-walk function to jump over holes in scan_get_next_rmap_item Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 55/72] ALSA: ac97bus: Use guard() for mutex locks Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 56/72] ALSA: ac97: fix a double free in snd_ac97_controller_register() Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 57/72] nfsd: provide locking for v4_end_grace Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 58/72] NFS: trace: show TIMEDOUT instead of 0x6e Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 59/72] nfs_common: factor out nfs_errtbl and nfs_stat_to_errno Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 60/72] NFSD: Remove NFSERR_EAGAIN Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 61/72] bpf: Fix an issue in bpf_prog_test_run_xdp when page size greater than 4K Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 62/72] bpf: Make variables in bpf_prog_test_run_xdp less confusing Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 63/72] bpf: Support specifying linear xdp packet data size for BPF_PROG_TEST_RUN Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 64/72] bpf, test_run: Subtract size of xdp_frame from allowed metadata size Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 65/72] bpf: Fix reference count leak in bpf_prog_test_run_xdp() Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 66/72] powercap: fix race condition in register_control_type() Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 67/72] powercap: fix sscanf() error return value handling Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 68/72] can: j1939: make j1939_session_activate() fail if device is no longer registered Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 69/72] ASoC: amd: yc: Add quirk for Honor MagicBook X16 2025 Greg Kroah-Hartman
2026-01-15 16:49 ` [PATCH 6.1 70/72] ASoC: fsl_sai: Add missing registers to cache default Greg Kroah-Hartman
2026-01-15 16:49 ` Greg Kroah-Hartman [this message]
2026-01-15 16:49 ` [PATCH 6.1 72/72] bpf: test_run: Fix ctx leak in bpf_prog_test_run_xdp error path Greg Kroah-Hartman
2026-01-15 19:15 ` [PATCH 6.1 00/72] 6.1.161-rc1 review Brett A C Sheffield
2026-01-15 19:36 ` Slade Watkins
2026-01-15 22:14 ` Florian Fainelli
2026-01-16 7:55 ` Francesco Dolcini
2026-01-16 9:14 ` Peter Schneider
2026-01-16 10:22 ` Ron Economos
2026-01-16 10:33 ` Jon Hunter
2026-01-16 15:39 ` Mark Brown
2026-01-16 17:44 ` Hardik Garg
2026-01-16 19:28 ` Shuah Khan
2026-01-17 14:35 ` Miguel Ojeda
2026-01-19 11:10 ` Jeffrin Thalakkottoor
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=20260115164146.078671547@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=czhong@redhat.com \
--cc=emilne@redhat.com \
--cc=jmeneghi@redhat.com \
--cc=martin.petersen@oracle.com \
--cc=mrabek@redhat.com \
--cc=patches@lists.linux.dev \
--cc=sashal@kernel.org \
--cc=stable@vger.kernel.org \
--cc=thenzl@redhat.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.