On Fri, 20 Apr 2001, Alan Cox wrote: > > Followup, this problem (IRQ routing conflict) continues to exist in > > 2.4.3-ac9. I booted back to 2.2.19 (same version of bttv) and > > everything is fine once again. Is this perhaps related to the VIA > > modifications that Alan reported in the 2.4.3-acx series? > > No. Its a real IRQ routing conflict according to the motherboard tables. > > HMM, i've wasted some time with this and I am shure that there is no real conflict. I have this on 2.4.x (no error in 2.2.x) (..) kernel: VP_IDE: IDE controller on PCI bus 00 dev 39 kernel: VP_IDE: chipset revision 6 kernel: VP_IDE: not 100% native mode: will probe irqs later kernel: ide0: BM-DMA at 0xe400-0xe407, BIOS settings: hda:DMA, hdb:DMA kernel: ide1: BM-DMA at 0xe408-0xe40f, BIOS settings: hdc:DMA, hdd:DMA (..) kernel: PCI: Found IRQ 10 for device 00:0a.0 kernel: IRQ routing conflict in pirq table for device 00:0a.0 kernel: eth0: RealTek RTL-8029 found at 0xec00, IRQ 12, 00:C0:DF:E1:E7:24. (..) My trial and error included a change in the order of cards on the PCI bus. No problems found. There is no real IRQ conflict, IRQs are properly allocated and I can also re-assign priorites with no problems. DMA modes on the VIA chipset are really fine, not a single problem. The error is related (triggered) by the VIA chipset and the video IRQ allocation but it is reported in the networking code upon searching/allocating the netw. IRQ. No problems when this IRQ is "fixed" (CMOS). When PnP is used this always leads to a hang. PnP is hard to workaround once you have to carefully take care of the modules loading order (PnP fisrt, IRQs allocation later...) To workaround the hang I am using "fixed" allocation. This way is fine. Abracos PauloCastro -- hom: http://www.momentus.com.br/users/hook/ GPG: http://www.momentus.com.br/users/hook/GPG-PauloCastro.asc