Bob Heise wrote: > Hello, > > I am using 2.6.19.1 + wireless-dev git kernel, bcm43xx driver with v.3 > firmware on an HP zv5000z amd64 with 2GB RAM. > > lspci -n > 02:02.0 Class 0280: 14e4:4320 (rev 03) > > lspci -vv > 02:02.0 Network controller: Broadcom Corporation BCM4306 802.11b/g Wireless > LAN Controller (rev 03) > Subsystem: Hewlett-Packard Company NX9500 Built-in Wireless > Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- > Stepping- SERR+ FastB2B- > Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- > <TAbort- <MAbort- >SERR- <PERR- > Latency: 64 > Interrupt: pin A routed to IRQ 17 > Region 0: Memory at e0104000 (32-bit, non-prefetchable) [size=8K] > > > This setup works if I limit the system RAM with mem=1024M on the boot prompt, > but with all the RAM it let's me modprobe but if I try to set the essid or do > a iwlist eth4 scanning etc., it deadlocks. I tried Larry's dma_over_1GB patch > from 01 December > <https://lists.berlios.de/pipermail/bcm43xx-dev/2006-December/003244.html>, > but the system still hangs. I could not get netconsole to work, but attached > is most of the dmesg + tail -f /var/messages which I captured via ssh. Also > included is a picture of the kernel panic, which did not make it across ssh.
Thanks for your report. It seems that the dma_over_1GB patch works on some systems, but not on others. I do not know why. I just got a new AMD 64 X2 computer with a motherboard capable of holding 2GB RAM. It currently has only 1GB with a Dell 1390 (BCM4311) that is capable of full 32-bit DMA. I'm trying to decide if I could do any debugging of the problem if I got more memory and tricked the program into thinking I had only 30-bit DMA capability. There is one other problem in that TX DMA doesn't work with the PCI-E hardware with 2.6.20 or wireless-2.6 versions. I need to fix that problem first. Larry _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
