[PATCH] ath10k: correct target assert problem due to CE5 stuck

mpubbise at qti.qualcomm.com mpubbise at qti.qualcomm.com
Thu Apr 12 05:14:59 PDT 2018

From: Manikanta Pubbisetty <mpubbise at codeaurora.org>

Correct a minor bug in the commit 0628467f97b5 ("ath10k: fix
copy engine 5 destination ring stuck") which introduced a change to fix
firmware assert that happens when ring indices of copy engine 5 are stuck
for a specific duration, problem with this fix is that it did not use
ring arithmatic. As a result,firmware asserts did not go away entirely
athough the frequency of occurrence has reduced. Using ring arithmatic
to fix the issue.

Tested on QCA9984(fw version-10.4-3.4-00082).

Fixes: 0628467f97b5 ("ath10k: fix copy engine 5 destination ring stuck)
Signed-off-by: Manikanta Pubbisetty <mpubbise at codeaurora.org>
 drivers/net/wireless/ath/ath10k/ce.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/wireless/ath/ath10k/ce.c b/drivers/net/wireless/ath/ath10k/ce.c
index b9def7b..6efa69e 100644
--- a/drivers/net/wireless/ath/ath10k/ce.c
+++ b/drivers/net/wireless/ath/ath10k/ce.c
@@ -615,7 +615,7 @@ void ath10k_ce_rx_update_write_idx(struct ath10k_ce_pipe *pipe, u32 nentries)
 	/* Prevent CE ring stuck issue that will occur when ring is full.
 	 * Make sure that write index is 1 less than read index.
-	if ((cur_write_idx + nentries)  == dest_ring->sw_index)
+	if (((cur_write_idx + nentries) & nentries_mask) == dest_ring->sw_index)
 		nentries -= 1;
 	write_index = CE_RING_IDX_ADD(nentries_mask, write_index, nentries);

More information about the ath10k mailing list