From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030703 Description of problem: on-board RTL-8139C fails to work (transmit timeouts) unless pci=noacpi specified on boot command line. Tried with both 2.4.21 and 2.6.0-test1 same result. Output of dmidecode and acpidmp attached. Version-Release number of selected component (if applicable): 2.4.21-20.1.2024.2.1.nptl, 2.6.0-0.test1.1.26 How reproducible: Always Steps to Reproduce: 1. Boot system without pci=noacpi. Actual Results: 8139too module loads, appears to work but error on dmesg and syslog indicates transmit timeout errors. No routes to network. Can ping interface itself but cannot ping any other systems on network. Expected Results: interface should have come up and run properly. Additional info: System board is Soyo P4X400 Dragon Ultra Platinum.
Created attachment 93251 [details] output of acpidmp under 2.6.0-test1
Created attachment 93252 [details] output of dmidecode
Please add the dmesg and /proc/interrupts output for the success and failing cases.
Created attachment 93289 [details] dmesg output, failure case (2.4 kernel)
Created attachment 93290 [details] dmesg output, success case (2.4 kernel)
Created attachment 93291 [details] /proc/interrupts, failure case (2.4 kernel)
Created attachment 93292 [details] /proc/interrupts, success case (2.4 kernel)
According to /proc/interrupts, both acpi and pci=noacpi came up with the same XT-PIC interrupt config -- but the acpi case gets no interrupts on eth0. In both cases, eth0 shares IRQ10 with USB, and according to dmesg the same USB controller in both cases. There are lots of interrupts on the other unshared USB controller. What USB devices are actually present -- can you attach the info from /proc/bus/usb? It would be interesting to see if the ethernet works when there are no USB devices in the system -- seems like an interrupt sharing issue.
OK, tried several things. Disconnected all USB devices (one ActionStar USB switch, and one RailDriver Modern Desktop, neither supported under Linux) and tried booting without PCI=noacpi, failed as before. Disabled all USB controllers (system has a total of 6 USB ports, 1.1 and 2.0), without PCI=noacpi, failed as before. Switched BIOS from PnP=no to PnP=yes, failed as before (USB disabled and enabled).
Created attachment 93425 [details] content of /proc/bus/usb/devices
Looks like the kernel is UP/PIC mode, can you try an SMP kernel. I think we should see better results becaue _PRT for APIC is much simpler. I think this system is DP or HT system.
2.4 smp kernel works. However, this is a UP system (with an HT-aware BIOS). attaching dmesg and /proc/interrupts outputs.
Created attachment 93697 [details] dmesg output, 2.4 smp kernel
Created attachment 93698 [details] /proc/interrupts from 2.4 smp kernel
We've got this machine, and updated the BIOS (http://www.soyousa.com/downloads/selectresults.php? language=&col1=Proc_Intel+Pentium+4&col2=159&col3=BIOS+Upgrade) to the latest . We were able to reproduce the problem with the RH kernel, but 2.6 with full ACPI worked fine with UP/PIC mode. So I think the problem is specific to that kernel. We are trying the stock 2.4.22 kernel.
Am also running the latest BIOS release (2AA7). The unofficial RedHat 2.6-test kernels have also failed for me in UP mode. SMP works. http://people.redhat.com/arjanv/2.5/RPMS.kernel/
Created attachment 93815 [details] fixed compile time bug reported by iasl Would you please try overriding DSDT using dsdtfixed.dsl. There are some compile-time bug caught by iasl.
Created attachment 93816 [details] Try to disable IRQ 0 for PCI Interrupt Link at PIC mode. Would you please try this patch. I also need output of lspci -vv for smp kernel. Ps. This patch is just for testing , I need to narrow down the problem.
Can you test 2.6.0-test3 with UP/ACPI? Ours worked fine on this machine, and we found different behavior with PCI Link device settings. We are root-causing it.
Sorry, please stop tesing 2.6.0-test3. Our tree has changes in BK tree.
Created attachment 93824 [details] Patch to resolve interrupt problems with ACPI With this patch, Serven kernel correctly sets up PCI interrrupt routing, and the network works fine. Encourage to try this on other machines.
Created attachment 93828 [details] output of lcpci -vv (smp kernel)
can someone point me to how to override the dsdt?
Please try the kernel patch that Jun attached earlier today, and you may not have to tinker with your DSDT.
Applied patch, unable to get kernel to build... ld -m elf_i386 -T /usr/src/linux-2.4.21-20.1.2024.2.1.nptl/arch/i386/vmlinux.lds -e stext arch/i386/kernel/head.o arch/i386/kernel/init_task.o init/main.o init/version.o init/do_mounts.o --start-group arch/i386/kernel/kernel.o arch/i386/mm/mm.o kernel/kernel.o mm/mm.o fs/fs.o ipc/ipc.o drivers/acpi/acpi.o drivers/char/char.o drivers/block/block.o drivers/misc/misc.o drivers/net/net.o drivers/char/drm/drm.o drivers/net/fc/fc.o drivers/net/appletalk/appletalk.o drivers/net/tokenring/tr.o drivers/net/wan/wan.o drivers/atm/atm.o drivers/ide/idedriver.o drivers/cdrom/driver.o drivers/pci/driver.o drivers/net/pcmcia/pcmcia_net.o drivers/net/wireless/wireless_net.o drivers/pnp/pnp.o drivers/video/video.o drivers/media/media.o drivers/md/mddev.o drivers/isdn/vmlinux-obj.o drivers/sensors/sensor.o net/network.o /usr/src/linux-2.4.21-20.1.2024.2.1.nptl/arch/i386/lib/lib.a /usr/src/linux-2.4.21-20.1.2024.2.1.nptl/lib/lib.a /usr/src/linux-2.4.21-20.1.2024.2.1.nptl/arch/i386/lib/lib.a --end-group -o .tmp_vmlinux1 fs/fs.o(.text+0x372e2): In function `zisofs_readpage': : undefined reference to `zlib_inflateInit_' fs/fs.o(.text+0x373e4): In function `zisofs_readpage': : undefined reference to `zlib_inflate' fs/fs.o(.text+0x37496): In function `zisofs_readpage': : undefined reference to `zlib_inflateEnd' fs/fs.o(.text.init+0xacc): In function `mnt_init': : undefined reference to `init_rootfs' fs/fs.o(.text.init+0xff1): In function `zisofs_init': : undefined reference to `zlib_inflate_workspacesize' make[1]: *** [kallsyms] Error 1 make[1]: Leaving directory `/usr/src/linux-2.4.21-20.1.2024.2.1.nptl' make: *** [vmlinux] Error 2 using the kernel-source rpm and the configs/kernel-2.4.21-i686.config I did double check and gcc32 is being used to build.
Created attachment 93853 [details] my config file Have my.config a try. It works for me. I also need the dmesg, if you resolve your problem using jun's patch Thanks, Luming
The patch does appear to work correctly. Don't suppose it will apply against 2.6.0-test3??? I usually run 2.6 to get AGP3 support... dmesg output attached below...
Created attachment 93877 [details] dmesg output, success w/patch
thanks for trying it out and supplying the dmesg. this patch is already in the latest 2.6.0 bk tree.
btw, confirmed working correctly under linus 2.6.0-test4