[ath9k-devel] [PATCH v2 01/10] ath10k: decouple pci init/deinit logic

2013-06-12 Thread Michal Kazior
Split logic that prepares the device for BMI phase/cleans up related resources. This is necessary for ath10k to be able to restart hw on the fly without reloading the module. Signed-off-by: Michal Kazior --- drivers/net/wireless/ath/ath10k/hif.h | 20 ++ drivers/net/wireless/ath/ath10k/pc

Re: [ath9k-devel] [PATCH v2 01/10] ath10k: decouple pci init/deinit logic

2013-06-13 Thread Kalle Valo
Michal Kazior writes: > Split logic that prepares the device for BMI > phase/cleans up related resources. > > This is necessary for ath10k to be able to restart > hw on the fly without reloading the module. > > Signed-off-by: Michal Kazior Few comments: > --- a/drivers/net/wireless/ath/ath10k/

Re: [ath9k-devel] [PATCH v2 01/10] ath10k: decouple pci init/deinit logic

2013-06-14 Thread Michal Kazior
On 13/06/13 20:09, Kalle Valo wrote: > So when booting the firwmware we call: > > hif_power_up() > hif_start() > > and when we want to kill the firmware we do: > > hif_stop() > hif_power_down() > > [...] You're right. I'll fix that. > I noticed ath10k_pci_ce_init() also allocs host memory etc. I

Re: [ath9k-devel] [PATCH v2 01/10] ath10k: decouple pci init/deinit logic

2013-06-14 Thread Kalle Valo
Michal Kazior writes: >> I noticed ath10k_pci_ce_init() also allocs host memory etc. If possible, >> in the future we might want to refactor the function into two: >> ath10k_pci_ce_init() and ath10k_pci_ce_start(). And the former would be >> called only from ath10k_pci_probe(). That way we would