[ipxe-devel] [romprefix] Regression within AMI BIOS

Michael Brown mbrown at fensystems.co.uk
Wed Jun 27 18:13:38 UTC 2012


On Monday 25 Jun 2012 23:09:15 Anton D. Kachalov wrote:
> Recently I've faced a problem with BIOS Option ROM embedding. From the
> bisecting I've found a problem commit:
> 
> $ git bisect bad
>  9e8d431a0d58caba3c5f65d4e30fd259d2f782e1 is the first bad commit
>  commit 9e8d431a0d58caba3c5f65d4e30fd259d2f782e1
>  Date:   Sun Jun 10 18:25:26 2012 +0100
>      [romprefix] Add a dummy ROM header to cover the .mrom payload

Thanks for doing the bisect!

Could you describe the problem?  Specifically, what do you see on-screen?

> Also, I have a problem on AMI that doesn't support PMM, with ROM bigger
> than ~58400 bytes. BIOS complains about:
> 
>     "Insufficient Runtime space for MPS data. System may operate in PIC or
> Non-MPS mode"
> 
> and then
> 
>     "Unknown BIOS error. Error code =004B".
> 
> And asks for F1 every time. However, PXE works.
> 
> PS. Mobo is Asus KGP(M)E-D16.

I don't think there's anything we can do about that.  If the ROM were attached 
to a NIC (rather than embedded within the BIOS) then using a .mrom image might 
help.  As it is, if the BIOS doesn't support PMM then iPXE can't do anything 
to reduce its consumption of option ROM space.

Michael



More information about the ipxe-devel mailing list