On Fri, 2013-02-01 at 03:57 -0800, Andrew Fish wrote:
> 
> Kevin,
> 
> This was all a long time ago, and my memory is fuzzy on things
> 16-bits ....
> 
> It looks like generic CSM code locks the 0xC0000 - 0xFFFFF region
> prior to boot, and I seem to remember that this was the common
> practice at the time. I'm not sure what old programs you are talking
> about, but I think they would not run on lots of PCs from the last
> decade+. 

I'm also hearing reports that some video BIOSes are failing when we use
the CSM on real hardware too; they also need the UMB region to be
writeable.
 
> All the OVMF does is abstract the 440 PAM registers via the
> EFI_LEGACY_REGION2_PROTOCOL. While it is not an ideal solution this
> code could be updated to lie and not lock a range that the SeaBIOS
> could use.

Ick. Let's just fix the spec to allow the CSM to tell us what range it
needs to be unlocked, really :)

It's an Intel in-house spec, so that shouldn't involve too much pain.
Hopefully. And there are other things which really *do* need fixing
(like the BBS_TABLE not having any field for LUN or anything like that,
so you can't distinguish between the various devices hanging off a
single PCI controller. Even IDE master/slave on primary/secondary bus is
only distinguishable by a nasty dirty evil implementation-specific
knowledge right now (knowing that it actually puts primary master in
BBS_TABLE[1], slave in BBS_TABLE[2] etc.).

I'd really like to fix things *properly* if we can, and resort to dirty
hacks only if we can't do that. I don't really want special-purpose
hacks on either side.

-- 
dwmw2

Attachment: smime.p7s
Description: S/MIME cryptographic signature

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_jan
_______________________________________________
edk2-devel mailing list
edk2-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/edk2-devel

Reply via email to