* [PATCH net] net/smc: fix limit_smc_hs sysctl on big-endian architecture
@ 2026-10-28 22:53 Fernando Fernandez Mancera
2026-10-28 22:57 ` sashiko-bot
0 siblings, 1 reply; 2+ messages in thread
From: Fernando Fernandez Mancera @ 2026-10-28 22:53 UTC (permalink / raw)
To: netdev
Cc: linux-s390, linux-rdma, jaka, horms, pabeni, kuba, edumazet,
davem, guwen, tonylu, mjambigi, sidraya, dust.li, alibuda,
Fernando Fernandez Mancera
net->smc.limit_smc_hs is a bool, but its sysctl entry uses
proc_dointvec_minmax() with .maxlen = sizeof(int), so the handler
accesses four bytes at the address of a one-byte field.
On little-endian this works with the extra bytes being written in struct
padding. On big-endian (tested on s390x) the bool sits at the most
significant byte of the int therefore the bool stays false and the "1"
goes into padding. The sysctl reads back 1 but SMC sockets created
afterwards see limit_smc_hs as disabled.
Fix this by using proc_dobool() handler and adjust the maxlen to the
right value.
Fixes: f8406a2fd279 ("net/smc: add sysctl for smc_limit_hs")
Signed-off-by: Fernando Fernandez Mancera <fmancera@suse.de>
---
Note: After reviewing another patch fixing sysctls, I decided to
review more of them and found this. I have tested the fix on a s390x
system.
---
net/smc/smc_sysctl.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/net/smc/smc_sysctl.c b/net/smc/smc_sysctl.c
index 09dad48337f6..f58cf256f61a 100644
--- a/net/smc/smc_sysctl.c
+++ b/net/smc/smc_sysctl.c
@@ -160,11 +160,9 @@ static const struct ctl_table smc_table[] = {
{
.procname = "limit_smc_hs",
.data = &init_net.smc.limit_smc_hs,
- .maxlen = sizeof(int),
+ .maxlen = sizeof(bool),
.mode = 0644,
- .proc_handler = proc_dointvec_minmax,
- .extra1 = SYSCTL_ZERO,
- .extra2 = SYSCTL_ONE,
+ .proc_handler = proc_dobool,
},
{
.procname = "smcr_max_send_wr",
--
2.56.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH net] net/smc: fix limit_smc_hs sysctl on big-endian architecture
2026-10-28 22:53 [PATCH net] net/smc: fix limit_smc_hs sysctl on big-endian architecture Fernando Fernandez Mancera
@ 2026-10-28 22:57 ` sashiko-bot
0 siblings, 0 replies; 2+ messages in thread
From: sashiko-bot @ 2026-10-28 22:57 UTC (permalink / raw)
To: Fernando Fernandez Mancera
Cc: linux-s390, linux-rdma, Heiko Carstens, Vasily Gorbik,
Alexander Gordeev, Christian Borntraeger
> net->smc.limit_smc_hs is a bool, but its sysctl entry uses
> proc_dointvec_minmax() with .maxlen = sizeof(int), so the handler
> accesses four bytes at the address of a one-byte field.
>
> On little-endian this works with the extra bytes being written in struct
> padding. On big-endian (tested on s390x) the bool sits at the most
> significant byte of the int therefore the bool stays false and the "1"
> goes into padding. The sysctl reads back 1 but SMC sockets created
> afterwards see limit_smc_hs as disabled.
>
> Fix this by using proc_dobool() handler and adjust the maxlen to the
> right value.
>
> Fixes: f8406a2fd279 ("net/smc: add sysctl for smc_limit_hs")
> Signed-off-by: Fernando Fernandez Mancera <fmancera@suse.de>
Sashiko has reviewed this patch and found no issues.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261028225350.3753-1-fmancera@suse.de?part=1
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-10-28 22:57 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-28 22:53 [PATCH net] net/smc: fix limit_smc_hs sysctl on big-endian architecture Fernando Fernandez Mancera
2026-10-28 22:57 ` sashiko-bot
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox