On 03/18/16 15:38, David Woodhouse wrote:
> On Tue, 2016-03-08 at 15:20 +0100, Laszlo Ersek wrote:
>> +UINT64 PciExBarBase;
> ...
>> if (mHostBridgeDevId == INTEL_Q35_MCH_DEVICE_ID) {
> ...
>> + PciExBarBase = FixedPcdGet64 (PcdPciExpressBaseAddress);
> ...
>> }
> ...
>> if (mHostBridgeDevId == INTEL_Q35_MCH_DEVICE_ID) {
> ...
>> + AddReservedMemoryBaseSizeHob (PciExBarBase, SIZE_256MB, FALSE);
Ugh, please don't snip context from patches. It's much harder to find
the context without the quote.
> The mHostBridgeDevId variable is global. Therefore as far as the
> compiler knows, it can change between those two if() statements.
>
> VS2008 complains thus:
>
> e:\edk2\ovmfpkg\platformpei\platform.c(275) : warning C4701: potentially
> uninitialized local variable 'PciExBarBase' used
Sure, VS2008 is wrong. :)
> If only we had some kind of code submission process where your
> submission could be automatically subjected to build tests on all
> platforms, and the status of those builds could be automatically
> displayed right there with your request... :)
I would already be using the Microsoft toolchains for test building, if
they were available at zero cost for the purposes I need them for. :)
You don't have to convince me about the utility of a build farm. I think
I (co-)proposed it years ago.
Who will build it and administer it? Who will foot the electricity &
cooling bills? Etc etc etc. If github can cross-build (or, heck,
natively build) ArmVirtPkg for aarch64, and also build OVMF with all
Microsoft toolchains we care about, now *that* would be "value add".
(And, if such a facility existed by now, I wouldn't depend on github to
test-build my stuff for me; I'd use the facility myself.)
Thanks
Laszlo
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel