In some usecases transaction ids are dynamically allocated inside
the controller driver after sending the messages which have generic
acknowledge responses. So check for this before refcounting pm_runtime.
Without this we would end up imbalancing runtime pm count by
doing pm_runtime_put() in both slim_do_transfer() and slim_msg_response()
for a single pm_runtime_get() in slim_do_transfer()
Fixes: d3062a210930 ("slimbus: messaging: add slim_alloc/free_txn_tid()")
Cc: <stable(a)vger.kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org>
---
drivers/slimbus/messaging.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/slimbus/messaging.c b/drivers/slimbus/messaging.c
index 6097ddc43a35..e5ae26227bdb 100644
--- a/drivers/slimbus/messaging.c
+++ b/drivers/slimbus/messaging.c
@@ -131,7 +131,8 @@ int slim_do_transfer(struct slim_controller *ctrl, struct slim_msg_txn *txn)
goto slim_xfer_err;
}
}
-
+ /* Initialize tid to invalid value */
+ txn->tid = 0;
need_tid = slim_tid_txn(txn->mt, txn->mc);
if (need_tid) {
@@ -163,7 +164,7 @@ int slim_do_transfer(struct slim_controller *ctrl, struct slim_msg_txn *txn)
txn->mt, txn->mc, txn->la, ret);
slim_xfer_err:
- if (!clk_pause_msg && (!need_tid || ret == -ETIMEDOUT)) {
+ if (!clk_pause_msg && (txn->tid == 0 || ret == -ETIMEDOUT)) {
/*
* remove runtime-pm vote if this was TX only, or
* if there was error during this transaction
--
2.21.0
As tid is unsigned its hard to figure out if the tid is valid or
invalid. So Start the transaction ids from 1 instead of zero
so that we could differentiate between a valid tid and invalid tids
This is useful in cases where controller would add a tid for controller
specific transfers.
Fixes: d3062a210930 ("slimbus: messaging: add slim_alloc/free_txn_tid()")
Cc: <stable(a)vger.kernel.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org>
---
drivers/slimbus/messaging.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/slimbus/messaging.c b/drivers/slimbus/messaging.c
index f2b5d347d227..6097ddc43a35 100644
--- a/drivers/slimbus/messaging.c
+++ b/drivers/slimbus/messaging.c
@@ -66,7 +66,7 @@ int slim_alloc_txn_tid(struct slim_controller *ctrl, struct slim_msg_txn *txn)
int ret = 0;
spin_lock_irqsave(&ctrl->txn_lock, flags);
- ret = idr_alloc_cyclic(&ctrl->tid_idr, txn, 0,
+ ret = idr_alloc_cyclic(&ctrl->tid_idr, txn, 1,
SLIM_MAX_TIDS, GFP_ATOMIC);
if (ret < 0) {
spin_unlock_irqrestore(&ctrl->txn_lock, flags);
--
2.21.0
Specifically, commit c4824ae7db418aee6f50f308a20b832e58e997fd triggers the problem. Reverting this change restores functionality.
The device is an RME Multiface II, using the snd-hdsp driver.
Expected behavior: Device plays sound normally
Exhibited behavior: When a program attempts to open the device, the following ALSA lib error happens:
ALSA lib pcm_direct.c:1169:(snd1_pcm_direct_initialize_slave) slave plugin does not support mmap interleaved or mmap noninterleaved access
This change hasn't affected my other computers with less esoteric hardware, so probably the problem lies with the snd-hdsp driver, but the device is unusable without reverting that commit.
I am available to test any patches for this issue.
Thanks,
J Woods