[ipxe-devel] [Realtek]: Possible bug - cannot load ipxe when intel-sata controller is set to RAID

[Matthew Helton]

Joshua,

What Michael has said is likely your problem: you are out of option
ROM space: To clarify; what Michael is talking about is the RAM the
BIOS requires at POST when enumerating BIOS devices (addon/integrated
option ROMs). The Patch will allow this to be shifted somewhat in memory;
it may help a bit. You only have about 64k of this "low" RAM to work with.


On Mon, Apr 15, 2013 at 9:37 AM, Joshua C. <joshuacov at gmail.com> wrote:

> Let me put more light into the situation: this is a msi p67a-gd65
> motherboard with am 8MB winbond bios-chip. The bios file has about 2MB
> "free" space so I can put some more files on the bios-chip itself. The
> motherboard uses a cheap realtek trl8111e chip which is soldgered on the
> motherboard and taps one of the pci-e lanes on the chipset. PCI-E 3.0 is
> not supported by the p67 chipset. The stock bios uses the propriatery
> realtek pxe stack which can be downloaded from the realtek official site
> (current version as of today: 2.54). The pxe stack is integrated into the
> bios file and I know that the realtek stack is less than 64kb in size.
>
> I have my custom ipxe rom which comes at about 81kb in size. As I said the
> bios flashes successfully and the ipxe works correctly as long as the
> intel-sata-controller is not set to raid-mode. The intel-orom is about
> 120kb and I know that some older versions on the intel chipset (i think p57
> was it) cannot work with newer oroms because of their size.
>
> I have disabled all other controllers on the motherboard (I don't need
> them anyway). The ones being active are usb, network and intel-sata.
>
> I can't say to the support service that I'm using a custom orom image
> because of the obvious reasons...
>
> I can compile the mrom image but since both the orom and mrom are in the
> same bios file I'm not sure if this can solve the problem. As I said this
> is not an expansion card with its own memory/extension bars. I can give it
> a try... Should I just compile both as stated here:
> http://ipxe.org/download and put them in the bios file? If so how can the
> mrom find the orom in this case??
>
>
>
>
> 2013/4/15 Michael Brown <mcb30 at ipxe.org>
>
>> On 15/04/13 13:10, Jarrod Johnson wrote:
>>
>>> On Mon, Apr 15, 2013 at 7:47 AM, Joshua C. <joshuacov at gmail.com
>>> <mailto:joshuacov at gmail.com>> wrote:
>>> > I have a strange problem with my bin/10ec8168.rom which I
>>> > successfully flashed in my bios. The orom loads and works normally
>>> > when the intel-sata controller is set to either
>>> > AHCI/IDE/Disabled.When I set it to RAID-mode then the orom doesn't
>>> load.
>>>
>>> The thing I think most likely is that you ran out of option rom space.
>>>   In x86 BIOS, it is a very limited area, and quite possibly switching
>>> the intel to raid bloated its footprint enough to leave insufficient
>>> space for your .rom.
>>>
>>> Have you tried make bin/10ec8168.mrom or does that not work with the
>>> device in question?  That would be my next step.  If there are more
>>> non-boot adapters that have option roms, I'd try disabling their option
>>> rom support in BIOS (at least IBM BIOS allows suppresion of slot option
>>> roms, I'm guessing most everyone does_.
>>>
>>
>> I'm pretty sure that this is correct; you are running out of option ROM
>> space when you enable RAID support.  Unfortunately, I don't think RTL8169s
>> don't have a sufficiently large memory BAR for .mrom to work.
>>
>> If your BIOS supports PCI3.00 (look at the PCI version shown in iPXE's
>> startup banner), and if the PMM allocations are working (check the startup
>> banner for non-zero values in "PMM+xxxxxxxx+yyyyyyyy"), then you could try
>> the attached patch.  This patch changes the "runtime size" field in the PCI
>> ROM header to the optimistic value (i.e. the value that will be achieved
>> only if the PMM allocations both work).
>>
>> Michael
>>
>> _______________________________________________
>> ipxe-devel mailing list
>> ipxe-devel at lists.ipxe.org
>> https://lists.ipxe.org/mailman/listinfo.cgi/ipxe-devel
>>
>>
>
>
> --
> --joshua
>
> _______________________________________________
> ipxe-devel mailing list
> ipxe-devel at lists.ipxe.org
> https://lists.ipxe.org/mailman/listinfo.cgi/ipxe-devel
>
>


-- 
There is never time enough to do it right, but there always seems to be
enough time to do it again.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ipxe.org/pipermail/ipxe-devel/attachments/20130415/62c0e49f/attachment.htm>