Tweet
First, I want to compliment you on the interactive TiVo Upgrade. It's well written. I just finished upgrading a 649080 DT to a 300 GB Seagate DB35 drive and by all I can tell thus far it's working very well. The problem was weaknees_lba_boot_cd.iso, not the interactive instructions (which were informative, even though I ultimately had to do the Hinsdale method).
The weaknees ISO burned just fine and would start to boot, and then dump out completely (no bash) with a "panic" error. The screen scrolled past so fast with messages it was impossible to read them all. It appeared that Linux had problems finding or perhaps using the IDE interrupts.
After several attempts, including trying to force using bios IRQs (with a command switch added at the boot prompt), I gave up, downloaded the LBA48 ISO for use with the Hinsdale method, and burned it. Got the same panic error - but noted that there was an option to boot with no DMA. That finally worked and Linux booted (I presume drive access was now in PIO Block Mode instead of (U)DMA). Following the Hinsdale instructions I was able to completely copy the old 80 GB drive to the new 300 GB drive, including preserving all the setup options, season passes, and recordings. Took a l-o-n-g time -- about 12 hours -- presumably because UDMA was disabled -- PIO Block Mode *is* quite slow with read/write compared to UDMA. My normal configuration of a machine is hard drive on primary EIDE channel and optical drive on secondary EIDE channel. I pulled the WinXP drive completely and put the TiVo drives on the primary EIDE channel as master and slave (leaving the CD/DVD on the secondary as master). Had to translate the MFSTools command line(s) to the hda and hdb for the two TiVo drives, but that wasn't a problem.
Unfortunately I could not boot the weaknees lba ISO with DMA switched off. The motherboard is a Soyo K8USA DRAGON Ultra Black Label, with socket 754 AMD Athlon 64 3400+. Supporting chipset (Northbridge/Southbridge) are an AMD 8151 (aka ALi/ULi M1687) and ALi/ULi M1563. This is not an ancient motherboard (e.g. resurrected Socket 7 box), although it's one of the earlier Socket 754 board architectures for the Athlon 64, but am wondering if that was part of the problem. However, that Northbridge/Southbridge is fully supported by at least SuSe Linux at least from somewhere around v9.x onward (haven't checked Ubuntu or Red Hat) which leaves me a little puzzled.
After completing the Hinsdale method with DMA turned off, it also occurred to me that motherboards in this class support APIC - which allows newer O/S to number interrupts beyond 16 in the O/S layer (even though they're still shared at the hardware layer) - and that was turned on in the BIOS (Win9x/ME doesn't support it, don't know about Win2k, but WinXP does). I'm not savvy enough about Linux to know if it supports APIC.
Searched the fora here and although I found a reboot loop problem, I didn't see any mention of a kernel "panic error" that failed to boot, completely dumping out of Linux. At some point when you do other revisions, you might consider adding a switch to turn DMA off with the Weaknees bootable CD. It's what saved me with the bootable CD I burned to use the Hinsdale method.
Any clues as to what I might have done differently to boot the Weaknees CD would be most welcome. We have three TiVos. The other two are older, single tuner 240080 models and know at some point I will upgrade at least one of them to a bigger drive - one of which is beginning to show early signs of drive failure - and would like to use the Weaknees bootable CD if possible - most hopefully with UDMA support to make file transfer much faster.
Postscript about the 649080 DT TiVo:
The drive I found inside was a Western Digital Caviar SE, model WD0800BB which is a general purpose desktop computer EIDE drive. Within the WD Caviar SE line, the "BB" models have the smaller cache (the "JB" have the larger cache).
Thanks,
-- John Lind
The weaknees ISO burned just fine and would start to boot, and then dump out completely (no bash) with a "panic" error. The screen scrolled past so fast with messages it was impossible to read them all. It appeared that Linux had problems finding or perhaps using the IDE interrupts.
After several attempts, including trying to force using bios IRQs (with a command switch added at the boot prompt), I gave up, downloaded the LBA48 ISO for use with the Hinsdale method, and burned it. Got the same panic error - but noted that there was an option to boot with no DMA. That finally worked and Linux booted (I presume drive access was now in PIO Block Mode instead of (U)DMA). Following the Hinsdale instructions I was able to completely copy the old 80 GB drive to the new 300 GB drive, including preserving all the setup options, season passes, and recordings. Took a l-o-n-g time -- about 12 hours -- presumably because UDMA was disabled -- PIO Block Mode *is* quite slow with read/write compared to UDMA. My normal configuration of a machine is hard drive on primary EIDE channel and optical drive on secondary EIDE channel. I pulled the WinXP drive completely and put the TiVo drives on the primary EIDE channel as master and slave (leaving the CD/DVD on the secondary as master). Had to translate the MFSTools command line(s) to the hda and hdb for the two TiVo drives, but that wasn't a problem.
Unfortunately I could not boot the weaknees lba ISO with DMA switched off. The motherboard is a Soyo K8USA DRAGON Ultra Black Label, with socket 754 AMD Athlon 64 3400+. Supporting chipset (Northbridge/Southbridge) are an AMD 8151 (aka ALi/ULi M1687) and ALi/ULi M1563. This is not an ancient motherboard (e.g. resurrected Socket 7 box), although it's one of the earlier Socket 754 board architectures for the Athlon 64, but am wondering if that was part of the problem. However, that Northbridge/Southbridge is fully supported by at least SuSe Linux at least from somewhere around v9.x onward (haven't checked Ubuntu or Red Hat) which leaves me a little puzzled.
After completing the Hinsdale method with DMA turned off, it also occurred to me that motherboards in this class support APIC - which allows newer O/S to number interrupts beyond 16 in the O/S layer (even though they're still shared at the hardware layer) - and that was turned on in the BIOS (Win9x/ME doesn't support it, don't know about Win2k, but WinXP does). I'm not savvy enough about Linux to know if it supports APIC.
Searched the fora here and although I found a reboot loop problem, I didn't see any mention of a kernel "panic error" that failed to boot, completely dumping out of Linux. At some point when you do other revisions, you might consider adding a switch to turn DMA off with the Weaknees bootable CD. It's what saved me with the bootable CD I burned to use the Hinsdale method.
Any clues as to what I might have done differently to boot the Weaknees CD would be most welcome. We have three TiVos. The other two are older, single tuner 240080 models and know at some point I will upgrade at least one of them to a bigger drive - one of which is beginning to show early signs of drive failure - and would like to use the Weaknees bootable CD if possible - most hopefully with UDMA support to make file transfer much faster.
Postscript about the 649080 DT TiVo:
The drive I found inside was a Western Digital Caviar SE, model WD0800BB which is a general purpose desktop computer EIDE drive. Within the WD Caviar SE line, the "BB" models have the smaller cache (the "JB" have the larger cache).
Thanks,
-- John Lind
Comment