Something changed between 2.6.12-1.1450_FC4 and 2.6.13-1.1558_FC5 in the IDE code. The disk of my old 770Z thinkpad no longer starts with dma enabled. It fails with dma_timer_expiry: dma status == 0x21, DMA timeout error. The workaround to boot with ide=nodma works for me. boot logs follow as attachment. [root@dhcp9 ~]# more /tmp/hdparm.2 /dev/hda: multcount = 16 (on) IO_support = 0 (default 16-bit) unmaskirq = 0 (off) using_dma = 0 (off) keepsettings = 0 (off) readonly = 0 (off) readahead = 256 (on) geometry = 29216/15/63, sectors = 27609120, start = 0 Model=IBM-DARA-215000, FwRev=GL5HA52A, SerialNo=GY0GYL10738 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs } RawCHS=16383/15/63, TrkSize=0, SectSize=0, ECCbytes=4 BuffType=DualPortCache, BuffSize=418kB, MaxMultSect=16, MultSect=16 CurCHS=16383/15/63, CurSects=15481935, LBA=yes, LBAsects=27609120 IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4 DMA modes: mdma0 mdma1 mdma2 UDMA modes: udma0 udma1 *udma2 udma3 udma4 AdvancedPM=yes: mode=0x9F (159) WriteCache=enabled Drive conforms to: ATA/ATAPI-4 T13 1153D revision 17: 1 2 3 4 * signifies the current active mode
Created attachment 118950 [details] Boot log of 2.6.12-1.1450_FC4
Created attachment 118951 [details] boot log of kernel 2.6.13-1.1558_FC5
booting with "acpi=force pci=noacpi" works around the problem. This laptop is rather old, so APM is used instead of ACPI by default. And pci=noacpi is also needed in this case to properly handle IRQs.
Should no longer be a problem because the ACPI cutoff date is enabled again Can you close the bug if current rawhide is ok ?
yes, DMA works again on rawhide, both with ACPI and with APM. thanks.