On 02/20/14 06:53, Sergey Isakov wrote:
> Hi Laszlo,
> I think there may be a bug: Variable.c
> Look
> -------------------
> NewLoadContext->ForceReconnect = (BOOLEAN) (NewLoadContext->Attributes &
> LOAD_OPTION_FORCE_RECONNECT);
> -------------------
> But LOAD_OPTION_FORCE_RECONNECT == 2 and BOOLEAN == UINT8 
> so ForceReconnect will be 2. Then
> --------------
> *((UINT32 *) Ptr) = LOAD_OPTION_ACTIVE | (ForceReconnect << 1);
> --------------
> I am not sure it is a bug but I think it should be corrected as
> -------------
> NewLoadContext->ForceReconnect = (BOOLEAN) ((NewLoadContext->Attributes
> & LOAD_OPTION_FORCE_RECONNECT)?1:0);
> -------------
> Wishes,
> Sergey

Thanks for the tip!

I'll look into this as soon as I can, because the bug is for real,
wherever it is -- while working on my OVMF vidmode series, I rebooted
one of my VMs so many times that it had *completely* run out flash
storage, hit an ASSERT() *before* reaching the setup screens, and that
was it.

Of course I could find out about this only because my nvvarstore is
split off into a separate file, and when I test a new OVMF build, the
varstore remains intact (as it should do)... but with ever less room :)
There were warning signs: after some point, the nvvar reclaim code ran
at each boot, but in vain: the free workspace continued to decrease.

The ASSERT() wasn't very hard to deal with, using a virtual machine
(restore pristine nvvarstore file from fresh build), but I suspect the
bug (wherever it is) might have bricked a real HW box :(

So we need to exterminate this insect.

Laszlo


------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
edk2-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/edk2-devel

Reply via email to