On Saturday 14 April 2001 17:21, you wrote:
I aploogize if this is documented, but I haven't been able to find
it... are the new betas, beta3 and such, ok to use with the Via KT133
chipset? I remember Beta1 and KT133 was bad news, but I can't find whether
or not it has been fixed. Thanks
Terrence
The beta 1 and VIA AND large WD drives were a problem
Now it turns out that there is another bug--affecting cross-channel data
transfers of large amounts of data under DMA--a hardware race condition for
those using the 686B Southbridge chip which goes in the KT133 KT133A and
Apollo Pro chipsets. It is a HARDWARE bug which I understand VIA is now
working on by helping mainboard manufacturers issue new BIOS versions to set
up the PCI-IDE differently. The defense is to cripple the kernel at install
time when a VIA chipset is seen.
BTW, this bug while subtle ALSO corrupts data freely on Windows systems under
the given circumstances, and that race condition may also occur
intermittently (not reproducibly) in other two-channel DMA roles.
See www.theregister.com April 13th edition for story and links to the actual
tests.
Civileme