Re: [PATCH] ath10k: Wait until copy complete is actually done before completing

Douglas Anderson wrote: > On wcn3990 we have "per_ce_irq = true". That makes the > ath10k_ce_interrupt_summary() function always return 0xfff. The > ath10k_ce_per_engine_service_any() function will see this and think > that _all_ copy engines have an interrupt. Without checking, the > ath10k_ce

Re: [PATCH] ath10k: Wait until copy complete is actually done before completing

Hi, On Mon, Jun 15, 2020 at 7:56 AM Kalle Valo wrote: > > Doug Anderson writes: > > > On Mon, Jun 15, 2020 at 7:32 AM Kalle Valo wrote: > >> > >> Douglas Anderson wrote: > >> > >> > On wcn3990 we have "per_ce_irq = true". That makes the > >> > ath10k_ce_interrupt_summary() function always ret

Re: [PATCH] ath10k: Wait until copy complete is actually done before completing

Doug Anderson writes: > On Mon, Jun 15, 2020 at 7:32 AM Kalle Valo wrote: >> >> Douglas Anderson wrote: >> >> > On wcn3990 we have "per_ce_irq = true". That makes the >> > ath10k_ce_interrupt_summary() function always return 0xfff. The >> > ath10k_ce_per_engine_service_any() function will see

Re: [PATCH] ath10k: Wait until copy complete is actually done before completing

Hi, On Mon, Jun 15, 2020 at 7:32 AM Kalle Valo wrote: > > Douglas Anderson wrote: > > > On wcn3990 we have "per_ce_irq = true". That makes the > > ath10k_ce_interrupt_summary() function always return 0xfff. The > > ath10k_ce_per_engine_service_any() function will see this and think > > that _al

Re: [PATCH] ath10k: Wait until copy complete is actually done before completing

Douglas Anderson wrote: > On wcn3990 we have "per_ce_irq = true". That makes the > ath10k_ce_interrupt_summary() function always return 0xfff. The > ath10k_ce_per_engine_service_any() function will see this and think > that _all_ copy engines have an interrupt. Without checking, the > ath10k_ce

Re: [PATCH] ath10k: Wait until copy complete is actually done before completing

Hi Doug, The send callback for the CEs do check for hw_index/SRRI before trying to free the buffer. But adding a check for copy-complete (before calling the individual CE callbacks) seems to be the better approach. Hence I agree that this patch should be added. Thanks, Rakesh Pillai. On 202

[PATCH] ath10k: Wait until copy complete is actually done before completing

On wcn3990 we have "per_ce_irq = true". That makes the ath10k_ce_interrupt_summary() function always return 0xfff. The ath10k_ce_per_engine_service_any() function will see this and think that _all_ copy engines have an interrupt. Without checking, the ath10k_ce_per_engine_service() assumes that i