From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7a) Gecko/20040212 Description of problem: during bootup to install of Fedora Core2 Test1, the kernel detected my onboard SATA controller (MSI KT6 Delta) and the drive I have connected to it. However, it had trouble and eventually locked up. it printed the messages: hde: max request size 1024KiB hde: lost interrupt hde: lost interrupt hde: lost interrupt hde: 156301488 sectors (80026MB)... it had more messages, including complaining about DMA, but the CD media test kicked in before I could write it all down. Current FC1 kernel (2.4.22-1.2166.nptl) identifies the SATA controller as: RAID bus controller: PCI device 1106:3149 (VIA Technologies, Inc.) (rev 128). the drive is a Seagate Barracude 80GB drive, model ST380013AS after the media test, the GUI interface came up, but my computer locked up hard at that point. I'm guessing there might be a bootup parameter to pass to get SATA working, but I couldn't find it in the release notes (and the FC1 kernel doesn't need it).
In FC2 test 1 during bootup to install the kernel failed to detect my onboard SATA controller and naturallyt the drive. My FC1 kernel (2.4.22-1.2166.nptl) identifies the SATA controller as: 02:04.0 RAID bus controller: Promise Technology, Inc. PDC20376 (rev 02) Forget FC2 test 1 for now... even the work around in building a 2.6.2-1.79 kernel and making a driver disk to boot with fails for driver sata_promise.ko and scsi_mod.ko.
For those interested: The work around that has proven well in the past but no longer is found at: http://www.fys.ku.dk/~esban/promise376.html
I'm confused... are you having trouble with the VIA or Promise controller?
VIA. I don't have a Promise controller (raxet does). Also I tried the development kernel-2.6.3-1.116, and it failed in the same way.
Does booting with "noapic" or "nomce" fix things?
I grabbed kernel 2.6.3-2.1.253.2.1 "noapic", "nomce", both and neither all had the same behavior as I originally reported.
i have a system with a aopen AK77-600gn i have a duron 1.6ghz cpu in it motherboard with a via SATA controller i have a WD 80 gb hdd attached to it and i got those errors early on i swapped the mobo since i had only had it 3 days the system was unuasable it would randomly drop the sata drive and give a whole bunch of dma and irq errors once i replaced the board i upgraded bios to latest available and all worked great for awhile about 3-4 weeks latter the errors returned and the hdd was making intresting sounds so i returned the hdd and got a new one and havent had any problems since i got the HDD back just before easter. im not 1000% sure if it was kernel related but it seemed to be more hardware related. i never installed FC1 on it as the installer didnt recognise any hdd's as being present. it never detected my SATA controller and there is no other HDD's in the system
This works for me with VIA AMD64 boards, still an issue for others?
Is test3 any better?
No. In fact, I tried kernel-2.6.5-1.349 last night. comment 6 still holds.
Are you sure its not hardware related? is there a bios update at all? i had my system up for over a week with no problems running 2.6.5-1.339 Just rebooted today into 2.6.5-1.350 the relevent part of dmesg for my chipset shows VIA8237SATA: IDE controller at PCI slot 0000:00:0f.0 VIA8237SATA: chipset revision 128 VIA8237SATA: 100% native mode on irq 11 ide2: BM-DMA at 0xcc00-0xcc07, BIOS settings: hde:pio, hdf:pio ide3: BM-DMA at 0xcc08-0xcc0f, BIOS settings: hdg:pio, hdh:pio hde: ST380013AS, ATA DISK drive
I just realised i put some wrong information in my initial comment i have a seagate SATA drive not a WD i got a few WD PATA drives for my filesever at the same time and messed them up.
I upgraded my BIOS which included (among other things): - Updates VIA SATA ROM version 2.31 but there was no change in the observed behavior (still testing 2.6.5-1.349)
Just wondering how you are going with this? This week i had my sata drive fail. i ran seagates diagnostic utilities and it had a whole bunch of damage on the surface scan. I will be taking the drive back in the morning but it will probably take two weeks to get the drive back and i am moving OS in three weeks so i wont get to test it much until after i have moved. It seemed to be working great until the hdd failed. i was using 2.6.6-1.377
Part of the problem is that the kernel should be using sata_via not IDE for the SATA controller.
it is sing sata_via but the driver represents the hdd as an ide one not scsi. if i went to install FC1 or RHEL on the machine i couldnt since they dont know how to deal with the via sata chipset and tell me i have no HDD. with my replacement HDD all seems fine.
I reinstalled FC2 from scratch after accidently nuking my drive. With the current kernel (kernel-2.6.7-1.494.2.2), the sata_via is used and the drive comes up as sda, but it still doesn't work. It hangs with: sda:<3>ata2: command 0x25 timeout stat 0x50 host_stat 0x4 I tried noapic and nomce again, but they didn't help.
ok i have had no issues foe awhile now i replaced my sata data cable as it was the only thing that had not been replaced and that seemed to finally fix all errors
Having the same trouble as Andrew Schultz (comment #17). FC2 installer won't boot because /dev/hde loses an interrupt. Latest FC2 kernel (2.6.8-1.521) won't boot because /dev/sda times out. The only possible way to get FC2 to work on this machine is to disable the SATA controller. I tried one of the FC3 tests (test 1 I believe) and it crashed with garbled text when the installer loads the sata_via module. The SATA drive I'm using has FC1 running nicely on it for many months now, currently using kernel 2.4.22-1.2197.nptl and the sata modules in the initrd.
I am having a similar problems. FC2 will get all the way to the installation process then freezes. I get the lost interrupt message and the 'can't find the dma' at boot up, but the process continues thru the beginning installation process. It moves incrediably slow but still moves on. I have turned ACPI on and off, changed out hardware, tried linux noprobe, turned off the dma in boot and it happens the same way everytime. I get to the installation process and the machine locks up. I am using the latest version of FC2 downloaded from their site. I have the latest BIOS upgrade to the MSI KT6 Delta board, and can find no latest upgrades to the VIA chipset. What I would like to know is why Mandrake 10.0 Official kernel 2.6.3-7mdk will work and this won't. I would rather use FC2 but if I can't get it to install I am stuck. My system is: AMD 2600xp w/ barton core MSI KT Delta lsr w/ via sata raid 2xWD Raptor SATA drives 36G each in RAID1 array 256M 333Mhz DDR Matrox Millenium video Please forgive if I am posting in the wrong place, I am new to all this!!
I've got the exact same problem with FC2, I finaly got my K7 motherboard working with both IDE and SATA drrives with FC1 but everytime I've tried a FC2 system I get the Lost Interupt/handup problem. Turning off acpi/pcm makes no difference. However I did finaly have a problem with my FC1 system when I added a 250 gig SATA drive. Basicly writes would fail to the SATA drives until I discovered I could get rid of this problem by disabling the IDE DMA. Could this be related?
my hardware is working ok once again with 2.6.10-1.9_FC2 w00t