This set addresses two issues one a contrived but real corner-case crash scenario and the other a simple typo. Debugging on a QCS405 which has an ath10k attached to PCIe its been found that a loop similar to the below [1] will cause. 1. A significant slow-down in the time it takes an individual ioread32() to complete. 2. A secure watchdog bite. This is as a result of the restart routine and the dump register routine running in parallel and a period of time during restart where dumping registers is unstable. The second patch is a simple fix to an apparent copy/paste error describing the behavior of a similar dump function. [1] Reset method while $1 do echo hw-restart > /sys/kernel/debug/ieee80211/phy0/ath10k/simulate_fw_crash echo hard > /sys/kernel/debug/ieee80211/phy0/ath10k/simulate_fw_crash done; Bryan O'Donoghue (2): ath10k: pci: Only dump ATH10K_MEM_REGION_TYPE_IOREG when safe ath10k: pci: Fix comment on ath10k_pci_dump_memory_sram drivers/net/wireless/ath/ath10k/pci.c | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) -- 2.24.0
