Sorry, I forgot something: I think the config line in the kernel compilation that fixed everything was `CONFIG_BCMA_BLOCKIO=y`. This stops BCMA from interfacing with devices, correct? Anyway, with SSB only, my card is working again at a normal speed. On Fri, Jan 24, 2014 at 8:45 PM, Wyatt Ward <wyattfward@xxxxxxxxx> wrote: > I have a bcm4311 chipset; it is identified in `lspci -vnn` as: > 01:00.0 Network controller [0280]: Broadcom Corporation BCM4311 > 802.11a/b/g [14e4:4312] (rev 01) > > Without compiling the kernel with some of BCMA disabled, I cannot make > my card work at a decent speed. > > This is in Linux Kernel 3.13, but I have had this problem with other > recent kernel versions. > My hardware is a late 2006, Intel Core 2 Duo iMac, and the linux > distro is Mint 16. > > The problem is that both SSB and BCMA are loading to support the b43 > module, where I only want SSB to load. I had this issue in 3.11, and > my custom build of 3.13 worked, but I don't see any changes to b43 or > BCMA that would effect this. I think that it was my tweaking of the > compile options for BCMA fixed this, but in any case > > I believe that the SSB and BCMA modules are interfering with each > other, killing my network speed seemingly at random. > I think it's because my card (*'0x4312') is being recognized in bcma > as a supported bcm4312 or 4313 chipset, when it actually can indicate > a 4311 chipset, which, as far as i know, is not supported by these, or > brcmsmac. I have to admit I am not clear on the differences between > these chipsets, but I do know the 4312 is a PHY device, and as far as > I know, the 4313 is an upgraded 4311. > > Attached are the `lspci` and `lsmod` outputs for when the 3.13 kernel > is loaded and when the 3.11 kernel is loaded. I believe that without > tweaking my compile settings, my system would still be behaving badly. > The 3.13 kernel works, and lsmod shows that bcma is not loaded. 3.11 > doesn't and shows it as loaded. I am having a hard time finding my > unmodified .config file for the kernel, but i have an intermediate > step's config file which also exhibited the problems. The `lsmod` and > `lspci` files also contain the outputs of `uname --all` so you can see > details of the kernel being used when I performed each test. > > I apologize if I am giving too much information, but I have never > reported a bug in the Linux Kernel before. I hope I did alright :) > -Wyatt -- 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
