On 11/11/13 22:13, David Woodhouse wrote:
> On Mon, 2013-11-11 at 21:03 +0100, Laszlo Ersek wrote:
>>
>> You're looking at it the wrong way.
>>
>> You contribute to so many projects that even you can't keep
>> account! ;)
> 
> There's a reason I try to push things upstream as fast as possible
> rather than leaving them lying around where I have to remember them,
> certainly.
> 
> Speaking of which, there's about five commits in
> http://git.infradead.org/users/dwmw2/edk2.git which are still resolutely
> present even after a 'git pull --rebase', one of which is yours. Any
> hint/reminder as to what's going on with them?

(1) 6ba4bf8 OvmfPkg: Don't build in QemuVideoDxe when we have CSM

If I remember correctly, I thought this patch was good:
<http://thread.gmane.org/gmane.comp.bios.tianocore.devel/1954/focus=1977>

Gerd's RPM carries this patch out of tree. So do I, in my private clone.
I think it should go in.


(2) 9682aea GenericBdsLib: Detect VirtIO disks and make them appear in
            the boot menu

(a) It touches IntelFrameworkModulePkg. I think you'll have to justify
it very strongly to the maintainer to have it picked up (even though you
have the advantage of working for Intel).

This point has a technical side too -- you're putting virtio-specific
knowledge in a central BDS file. Virtio is not an industry standard yet.
(I'm just playing the devil's advocate here.)

(b) "Sub-Type WTF" is probably not the commenting style the maintainer
would welcome with open arms :)

BTW I also seem to have reviewed (more like, commented on) this patch
before:
<http://thread.gmane.org/gmane.comp.bios.tianocore.devel/1662/focus=1669>.

The patch is not strictly related to CSM (you might want to see
virtio-blk disks listed in the boot menu anyway). But, honestly, I
*never* boot from that menu, except the EFI shell (which is
memory-mapped). For me personally this patch is not very important (but
of course I won't argue against it either).

AFAIR Gerd's RPM doesn't carry this patch.


(3) 7ff0ee9 LegacyBios: set NumberBbsEntries to the size of BbsTable

Can you justify this as a standalone bugfix to the maintainer? (Sorry, I
can't recall if this was a bugfix or a generalization needed for the
next patch.)


(4) 6f34630 LegacyBbs: add boot entries for virtio-blk devices

(a) Again, IntelFrameworkModulePkg plus virtio...

(b) If I understand correctly, patch (4) -- and maybe patch (3) -- are
important if you want to legacy-boot OSes from virtio-blk disks, from
the boot menu. For me personally that functionality is even less
relevant than (2).

As I understand it, the motivation behind this patch (and LegacyBbs in
general) is the intent to make OVMF (or UEFI more generally) the default
(or only) boot firmware for virtual machines.

In practice however even libvirt should allow now the user to select a
firmware for his/her guest conveniently. So I think that most users
simply use pure SeaBIOS for guests that they'd have to boot via these
menu entries otherwise.

I'm not opposing this patch of course, it's just that I can't see any
real demand for it, and that makes it quite hard to argue in front of
the IntelFrameworkModulePkg maintainer. (From my POV.)

Gerd's RPM doesn't carry patches (3) and (4).


(5) 9607e51 LegacyBios: Update EFI_COMPATIBILITY16_TABLE to match 0.98
            CSM spec update

This patch is of course very important and it should go in (for which I
think there's a very good chance, since you managed to change the spec
-- from that point on it's just "make the code follow the spec").

Gerd's RPM carries an earlier version of this patch (also written by
you) which (IIRC) unlocks a fixed region used by SeaBIOS.


> I updated the final commit in that tree, as well as the only outstanding
> commit in http://git.infradead.org/users/dwmw2/seabios.git to handle the
> writable UMB stack according to the v0.98 spec update.

Gerd co-maintains SeaBIOS, I propose to take it up with him.

Thanks
Laszlo

------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most 
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
edk2-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/edk2-devel

Reply via email to