* [PATCH] RDMA/rxe: validate access flags before swapping the MR's PD
@ 2026-08-27 17:18 Norbert Szetei
2026-08-27 22:26 ` Zhu Yanjun
0 siblings, 1 reply; 2+ messages in thread
From: Norbert Szetei @ 2026-08-27 17:18 UTC (permalink / raw)
To: linux-rdma; +Cc: Zhu Yanjun, Jason Gunthorpe, Leon Romanovsky, linux-kernel
rxe_rereg_user_mr() reassigns mr->ibmr.pd first and only then
validates the IB_MR_REREG_ACCESS argument:
if (flags & IB_MR_REREG_PD) {
rxe_put(old_pd);
rxe_get(pd);
mr->ibmr.pd = ibpd;
}
if (flags & IB_MR_REREG_ACCESS) {
if (access & ~RXE_ACCESS_SUPPORTED_MR)
return ERR_PTR(-EOPNOTSUPP);
mr->access = access;
}
Both flags pass the entry check because RXE_MR_REREG_SUPPORTED is
IB_MR_REREG_PD | IB_MR_REREG_ACCESS, so a caller can reach the access
check with mr->ibmr.pd already reassigned.
mr->ibmr.pd is owned by the core, which adjusts pd->usecnt only on the
success path: ib_uverbs_rereg_mr() jumps to put_new_uobj on a driver error
without undoing the reassignment, so mr->pd == new_pd while the usecnts
still charge the MR to orig_pd. ib_dereg_mr_user() then decrements
new_pd, whose count can reach zero while a memory window still references
it; uverbs_free_pd() frees the PD on that count alone and rxe_mw_cleanup()
writes to freed memory:
BUG: KASAN: slab-use-after-free in __rxe_put+0x31/0xa0
Write of size 4 at addr ffff8881301dd690 by task rxe_poc/591
__rxe_put+0x31/0xa0
rxe_mw_cleanup+0x42/0x200
__rxe_cleanup+0x115/0x370
rxe_dealloc_mw+0x4c/0x80
Allocated by task 591:
ib_uverbs_alloc_pd+0x258/0x540
Freed by task 591:
ib_dealloc_pd_user+0x174/0x210
uverbs_free_pd+0x8d/0xc0
ib_uverbs_dealloc_pd+0x18e/0x1d0
Validate the access flags before mutating any state so the callback either
applies every requested change or none.
Fixes: 544c7f62cf32 ("RDMA/rxe: Implement rereg_user_mr")
Cc: stable@vger.kernel.org
Signed-off-by: Norbert Szetei <norbert@doyensec.com>
---
drivers/infiniband/sw/rxe/rxe_verbs.c | 13 +++++++------
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c
index 96c7716057fe..3864284522eb 100644
--- a/drivers/infiniband/sw/rxe/rxe_verbs.c
+++ b/drivers/infiniband/sw/rxe/rxe_verbs.c
@@ -1331,19 +1331,20 @@ static struct ib_mr *rxe_rereg_user_mr(struct ib_mr *ibmr, int flags,
if (err)
return ERR_PTR(err);
+ if ((flags & IB_MR_REREG_ACCESS) &&
+ (access & ~RXE_ACCESS_SUPPORTED_MR)) {
+ rxe_err_mr(mr, "access = %#x not supported\n", access);
+ return ERR_PTR(-EOPNOTSUPP);
+ }
+
if (flags & IB_MR_REREG_PD) {
rxe_put(old_pd);
rxe_get(pd);
mr->ibmr.pd = ibpd;
}
- if (flags & IB_MR_REREG_ACCESS) {
- if (access & ~RXE_ACCESS_SUPPORTED_MR) {
- rxe_err_mr(mr, "access = %#x not supported\n", access);
- return ERR_PTR(-EOPNOTSUPP);
- }
+ if (flags & IB_MR_REREG_ACCESS)
mr->access = access;
- }
return NULL;
}
--
2.55.0
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] RDMA/rxe: validate access flags before swapping the MR's PD
2026-08-27 17:18 [PATCH] RDMA/rxe: validate access flags before swapping the MR's PD Norbert Szetei
@ 2026-08-27 22:26 ` Zhu Yanjun
0 siblings, 0 replies; 2+ messages in thread
From: Zhu Yanjun @ 2026-08-27 22:26 UTC (permalink / raw)
To: Norbert Szetei, linux-rdma, yanjun.zhu
Cc: Zhu Yanjun, Jason Gunthorpe, Leon Romanovsky, linux-kernel
在 2026/8/27 10:18, Norbert Szetei 写道:
> rxe_rereg_user_mr() reassigns mr->ibmr.pd first and only then
> validates the IB_MR_REREG_ACCESS argument:
>
> if (flags & IB_MR_REREG_PD) {
> rxe_put(old_pd);
> rxe_get(pd);
> mr->ibmr.pd = ibpd;
> }
>
> if (flags & IB_MR_REREG_ACCESS) {
> if (access & ~RXE_ACCESS_SUPPORTED_MR)
> return ERR_PTR(-EOPNOTSUPP);
> mr->access = access;
> }
>
> Both flags pass the entry check because RXE_MR_REREG_SUPPORTED is
> IB_MR_REREG_PD | IB_MR_REREG_ACCESS, so a caller can reach the access
> check with mr->ibmr.pd already reassigned.
>
> mr->ibmr.pd is owned by the core, which adjusts pd->usecnt only on the
> success path: ib_uverbs_rereg_mr() jumps to put_new_uobj on a driver error
> without undoing the reassignment, so mr->pd == new_pd while the usecnts
> still charge the MR to orig_pd. ib_dereg_mr_user() then decrements
> new_pd, whose count can reach zero while a memory window still references
> it; uverbs_free_pd() frees the PD on that count alone and rxe_mw_cleanup()
> writes to freed memory:
>
> BUG: KASAN: slab-use-after-free in __rxe_put+0x31/0xa0
> Write of size 4 at addr ffff8881301dd690 by task rxe_poc/591
> __rxe_put+0x31/0xa0
> rxe_mw_cleanup+0x42/0x200
> __rxe_cleanup+0x115/0x370
> rxe_dealloc_mw+0x4c/0x80
> Allocated by task 591:
> ib_uverbs_alloc_pd+0x258/0x540
> Freed by task 591:
> ib_dealloc_pd_user+0x174/0x210
> uverbs_free_pd+0x8d/0xc0
> ib_uverbs_dealloc_pd+0x18e/0x1d0
>
> Validate the access flags before mutating any state so the callback either
> applies every requested change or none.
>
> Fixes: 544c7f62cf32 ("RDMA/rxe: Implement rereg_user_mr")> Cc: stable@vger.kernel.org
> Signed-off-by: Norbert Szetei <norbert@doyensec.com>
Thanks a lot. I am fine with this commit.
Reviewed-by: Zhu Yanjun <yanjun.zhu@linux.dev>
Zhu Yanjun
> ---
> drivers/infiniband/sw/rxe/rxe_verbs.c | 13 +++++++------
> 1 file changed, 7 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c
> index 96c7716057fe..3864284522eb 100644
> --- a/drivers/infiniband/sw/rxe/rxe_verbs.c
> +++ b/drivers/infiniband/sw/rxe/rxe_verbs.c
> @@ -1331,19 +1331,20 @@ static struct ib_mr *rxe_rereg_user_mr(struct ib_mr *ibmr, int flags,
> if (err)
> return ERR_PTR(err);
>
> + if ((flags & IB_MR_REREG_ACCESS) &&
> + (access & ~RXE_ACCESS_SUPPORTED_MR)) {
> + rxe_err_mr(mr, "access = %#x not supported\n", access);
> + return ERR_PTR(-EOPNOTSUPP);
> + }
> +
> if (flags & IB_MR_REREG_PD) {
> rxe_put(old_pd);
> rxe_get(pd);
> mr->ibmr.pd = ibpd;
> }
>
> - if (flags & IB_MR_REREG_ACCESS) {
> - if (access & ~RXE_ACCESS_SUPPORTED_MR) {
> - rxe_err_mr(mr, "access = %#x not supported\n", access);
> - return ERR_PTR(-EOPNOTSUPP);
> - }
> + if (flags & IB_MR_REREG_ACCESS)
> mr->access = access;
> - }
>
> return NULL;
> }
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-08-27 22:26 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-08-27 17:18 [PATCH] RDMA/rxe: validate access flags before swapping the MR's PD Norbert Szetei
2026-08-27 22:26 ` Zhu Yanjun
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®