On 07/14/2012 11:48 AM, Arend van Spriel wrote: > On 07/11/2012 04:32 PM, Rafał Miłecki wrote: >> 2012/7/11 Arend van Spriel <arend@xxxxxxxxxxxx>: >>> Hi Rafal >>> >>> I am testing with BCM43224 in PCIe slot: >>> $ lspci -vns 03:00.0 >>> 03:00.0 0280: 14e4:4353 (rev 01) >>> Subsystem: 1028:000e >>> Flags: bus master, fast devsel, latency 0, IRQ 17 >>> Memory at e6e00000 (64-bit, non-prefetchable) [size=16K] >>> Capabilities: <access denied> >>> Kernel driver in use: bcma-pci-bridge >>> >>> In the log the irq field in the bcma_device structure is printed in >>> brcmsmac probe function and the value seems invalid hence the >>> request_irq() failure. Any clues? brcmsmac prints: irq: -421527552 this is the reason request_irq() fails. Does this problem also occurs on other wireless cards? >>> >>> Gr. AvS >>> >>> [86629.110723] bcma: bus0: Found chip with id 0xA8D8, rev 0x01 and >>> package 0x0A >>> [86629.110783] bcma: bus0: Core 0 found: ChipCommon (manuf 0x4BF, id >>> 0x800, rev 0x22, class 0x0) >>> [86629.110828] bcma: bus0: Core 1 found: IEEE 802.11 (manuf 0x4BF, id >>> 0x812, rev 0x17, class 0x0) >>> [86629.110939] bcma: bus0: Core 2 found: PCIe (manuf 0x4BF, id 0x820, >>> rev 0x0F, class 0x0) >>> [86629.144286] bcma: bus0: Bus registered >>> [86672.144812] brcmsmac bcma0:0: mfg 4bf core 812 rev 23 class 0 irq >>> -421527552 This number is wrong. This number is a unsigned int, and %d should only be used for signed ints. Have you checked if irq is set correctly in drivers/bcma/main.c or if the pci_dev structure contains a wrong value? -421527552 is 0xe6e00000 in hex, this looks like it reads out a wrong memory region when accession a structure, have you tried to to a complete recompile of the kernel? >>> [86672.150745] brcmsmac: Applying 43224B0+ WARs >>> [86672.153155] ieee80211 phy0: wl0: request_irq() failed >>> [86672.153169] brcmsmac: brcms_bcma_probe: brcms_attach failed! >> >> No idea at the moment. I can setup my machine and give it a try. I'll >> need a day or rather two. >> >> Are you using wireless-testing? >> > > Additional info: the issue does not show on a x86_64 platform, which has > the same PCIe chip. This is strange but it increases my thoughts that the error is not directly related to brcmsmac and bcma. -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
