By: Igor ((Not Given)), July 23, 2004 10:00 pm
Room: Moderated Discussions
tom vier (no@thanks.net) on 7/22/04 wrote:
---------------------------
>if you can write protect your mobo's flash, it's not as big a deal. i wonder how
>much ucode rom intel had to add to support the update decryption.
Write-protecting flash wouldn't help much, it would only prevent you from making the hack permanent by flashing it into the BIOS. Anyway, you can issue an update at any time provided that you run it in kernel mode because WRMSR is privileged instruction.
As far as Intel is concerned I doubt that their protection is much stronger. I have seen small repeating sequences in their updates for the processors that have the same family. Only when model number changes do they become completely different. Stepping doesn't matter much. What I think is that they change encryption/decription codes when they perform a mask shrinking (as you know, F1x was 0.18u P4 core, F2x was 0.13u and now F3x is 0.09u core).
---------------------------
>if you can write protect your mobo's flash, it's not as big a deal. i wonder how
>much ucode rom intel had to add to support the update decryption.
Write-protecting flash wouldn't help much, it would only prevent you from making the hack permanent by flashing it into the BIOS. Anyway, you can issue an update at any time provided that you run it in kernel mode because WRMSR is privileged instruction.
As far as Intel is concerned I doubt that their protection is much stronger. I have seen small repeating sequences in their updates for the processors that have the same family. Only when model number changes do they become completely different. Stepping doesn't matter much. What I think is that they change encryption/decription codes when they perform a mask shrinking (as you know, F1x was 0.18u P4 core, F2x was 0.13u and now F3x is 0.09u core).