Andy Green wrote: > Just a thought, did you test the reset fix against Atheros stack before > now? Maybe the problem is simply that the Atheros stack, known to be > unable to find a device powered on SDIO after the stack initializes, is > initialized before the pm-wlan thing.
Good idea ... but no, the reset happens well before the rest. Here's the module_init order in stable-tracking: [21474537.300000] @@@ gta02_wlan_power(1) [21474538.290000] @@@ stack/busdriver/sdio_bus_os.c:sdio_busdriver_init [21474538.300000] @@@ stack/lib/sdio_lib_os.c:sdio_lib_init [21474538.305000] @@@ drivers/sdio/hcd/s3c24xx/s3c24xx_hcd.c:s3c24xx_hcd_init [21474538.310000] @@@ function/wlan/ar6000/ar6000/ar6000_drv.c:ar6000_init_module and here in stable: [ 3.420000] @@@ stack/busdriver/sdio_bus_os.c:sdio_busdriver_init [ 3.430000] @@@ stack/lib/sdio_lib_os.c:sdio_lib_init [ 3.435000] @@@ drivers/sdio/hcd/s3c24xx/s3c24xx_hcd.c:s3c24xx_hcd_init [ 3.580000] @@@ function/wlan/ar6000/ar6000/ar6000_drv.c:ar6000_init_module So, no changes there. (Pheew, four initializations that are probably highly sensitive to order, when just one would do nicely :) But the card detection gets kicked off later. Let's see what happens there ... - Werner
