From: Arun Iyengar aiyengar1990@gmail.com
optee_supp_thrd_req() checks supp->ctx without supp->mutex. If the supplicant is released between the check and queueing, a supp_nowait request is queued after optee_supp_release() aborted the queue and waits until a new supplicant opens. For in-kernel clients that can be forever.
Do the check under supp->mutex, together with queueing the request.
Fixes: 42bf4152d8a7 ("tee: add supp_nowait flag in tee_context struct") Cc: stable@vger.kernel.org Signed-off-by: Arun Iyengar aiyengar1990@gmail.com --- Found by Sashiko in its review of [1]; the race is older than that patch. Tested on an Orange Pi 5 (7.3-based, KASAN and lockdep) with a test-only 2 ms delay between the check and queueing; 1242 of 2000 supp_nowait requests stranded without this patch, none with it. A real tee-supplicant was not tested.
Developed with Claude (Anthropic), which wrote the code and the commit message. I reviewed it.
[1] https://lore.kernel.org/r/20261115-optee-supp-close-v1-1-4e0b14f3a97c@gmail.... --- drivers/tee/optee/supp.c | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-)
diff --git a/drivers/tee/optee/supp.c b/drivers/tee/optee/supp.c index 06747e90c230..fd8d1f363f2f 100644 --- a/drivers/tee/optee/supp.c +++ b/drivers/tee/optee/supp.c @@ -91,13 +91,6 @@ u32 optee_supp_thrd_req(struct tee_context *ctx, u32 func, size_t num_params, struct optee_supp_req *req; u32 ret;
- /* - * Return in case there is no supplicant available and - * non-blocking request. - */ - if (!supp->ctx && ctx->supp_nowait) - return TEEC_ERROR_COMMUNICATION; - req = kzalloc_obj(*req); if (!req) return TEEC_ERROR_OUT_OF_MEMORY; @@ -109,6 +102,16 @@ u32 optee_supp_thrd_req(struct tee_context *ctx, u32 func, size_t num_params,
/* Insert the request in the request list */ mutex_lock(&supp->mutex); + /* + * Return in case there is no supplicant available and + * non-blocking request. + */ + if (!supp->ctx && ctx->supp_nowait) { + mutex_unlock(&supp->mutex); + kfree(req); + return TEEC_ERROR_COMMUNICATION; + } + req->id = idr_alloc(&supp->idr, req, 1, 0, GFP_KERNEL); if (req->id < 0) { mutex_unlock(&supp->mutex);
--- base-commit: e21989d08e0d6abacfff71387d0cb36dd96a4079 change-id: 20261119-optee-supp-toctou-da2c5f097e31
Best regards, -- Arun Iyengar aiyengar1990@gmail.com
op-tee@lists.trustedfirmware.org