On Thu, 07 Apr 2005, Daniel Ritz wrote: > On Tuesday 05 April 2005 08:56, Jefferson Cowart wrote: > > Sorry for the delay in sending this. I ended up getting back later than > > expected. In any case the hexdumps are below. Please let me know if you need > > anything else. > > also sorry for the delay :) > > > > > Under 2.6.5-bk1 > > =============== > > (/proc/bus/pci/00/07.0) > > 0000000 104c ac17 0007 0210 0002 0607 a820 0082 > > 0000010 8000 000c 00a0 0200 0100 b004 0000 1000 > > 0000020 f000 103f 0000 1040 f000 107f 4400 0000 > > 0000030 44fc 0000 4800 0000 48fc 0000 010a 0540 > > 0000040 0000 0000 0001 0000 0000 0000 0000 0000 > > 0000050 0000 0000 0000 0000 0000 0000 0000 0000 > > * > > 0000080 7020 2800 0000 0000 0000 0000 3822 0ba4 > > ^ > there you go. the intrtie bit is set. this means that both functions > use INTA to signal the interrupts. but the BIOS is broken and assigned > a different irq to function 1 compared to function 0. > so the reason it broke when the irq-routing patch was merged is that > irqs are actually probed to see if they work or if the routing needs > some tuning... > > the attached patch should fix it. it's untested 'cos i don't have a > laptop with TI bridge anymore...but it's simple > > rgds > -daniel
thanks a lot for your explenations and the provided patch. i need some sleep right now, but i'll build a kernel tommorow for jefferson - bug reporter - to test. a++ maks -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]