qcacmn: do not process duplicate descriptor frame in RX

once in a while the HW is sending a descriptor which
is already processed by host. This can be a potential HW
issue, as a WAR we are not processing such duplicate descriptors
instead increment a counter and continue with next descriptor.

Change-Id: I6c9bc6a9fb4705b42284171a32855411aa5dd73f
CRs-Fixed: 2338543
This commit is contained in:
Tallapragada Kalyan
2019-01-09 19:13:19 +05:30
committed by nshrivas
parent e88c0e8ffa
commit eff377a1d3
7 changed files with 119 additions and 4 deletions

View File

@@ -1253,6 +1253,19 @@ dp_rx_wbm_err_process(struct dp_soc *soc, void *hal_ring, uint32_t quota)
continue;
}
/*
* this is a unlikely scenario where the host is reaping
* a descriptor which it already reaped just a while ago
* but is yet to replenish it back to HW.
* In this case host will dump the last 128 descriptors
* including the software descriptor rx_desc and assert.
*/
if (qdf_unlikely(!rx_desc->in_use)) {
DP_STATS_INC(soc, rx.err.hal_wbm_rel_dup, 1);
dp_rx_dump_info_and_assert(soc, hal_ring,
ring_desc, rx_desc);
}
nbuf = rx_desc->nbuf;
qdf_nbuf_unmap_single(soc->osdev, nbuf, QDF_DMA_BIDIRECTIONAL);