Re: [edk2] csm issues

2013-04-10 Thread Rafael Machado
Hi Andrew. Your e-mail really clarified several things. We'll fix our allocation methods. Thanks a lot for the help. Att. Rafael R. Machado 2013/4/9 Andrew Fish > > On Apr 9, 2013, at 12:15 PM, Rafael Machado < > rafaelrodrigues.mach...@gmail.com> wrote: > > The only point I imagine is that t

Re: [edk2] csm issues

2013-04-09 Thread Andrew Fish
On Apr 9, 2013, at 12:15 PM, Rafael Machado wrote: > The only point I imagine is that there is a possibility of some application > fail if it alocates the first page returned by the GetMemoryMap function even > with this page being of type Desc->Type == EfiConventionalMemory And this > seems

Re: [edk2] csm issues

2013-04-09 Thread Rafael Machado
The only point I imagine is that there is a possibility of some application fail if it alocates the first page returned by the GetMemoryMap function even with this page being of type Desc->Type == EfiConventionalMemory And this seems to be possible when CSM is disabled. Should I disconsider the fi

Re: [edk2] csm issues

2013-04-09 Thread Andrew Fish
On Apr 9, 2013, at 11:42 AM, Rafael Machado wrote: > I think I was not clear. I didn't add all information > The first step at the application is to get the MemoryMap > (gBS->GetMemoryMap), and after that, alocate the pages of the first block I > can access Based on the description of that pa

Re: [edk2] csm issues

2013-04-09 Thread Rafael Machado
I think I was not clear. I didn't add all information The first step at the application is to get the MemoryMap (gBS->GetMemoryMap), and after that, alocate the pages of the first block I can access Based on the description of that page if Desc->Type == EfiConventionalMemory The problem seems tha

Re: [edk2] csm issues

2013-04-09 Thread Andrew Fish
On Apr 9, 2013, at 10:50 AM, Rafael Machado wrote: > Hi Andrew > > Thanks for the answer. > It's strange the the application allocated these addresses. > These seems to be at the BDA right ? They should not be allocated. > > At the application I use several calls to: > > gBS->AllocatePoo

Re: [edk2] csm issues

2013-04-09 Thread Rafael Machado
Hi Andrew Thanks for the answer. It's strange the the application allocated these addresses. These seems to be at the BDA right ? They should not be allocated. At the application I use several calls to: gBS->AllocatePool(EfiBootServicesData, sizeof(EFI_MEMORY_BLOCKS), (VOID**) &MemoryBlocks);

Re: [edk2] csm issues

2013-04-09 Thread Andrew Fish
On Apr 9, 2013, at 7:48 AM, Rafael Machado wrote: > After some debug here comes the question. > Why am I able to allocated more memory when csm is off ? Because the CSM allocates memory when it runs. > > Does uefi have some kind of virtual memory ? Not generally. For X64 and ARM paging is

Re: [edk2] csm issues

2013-04-09 Thread Rafael Machado
After some debug here comes the question. Why am I able to allocated more memory when csm is off ? Does uefi have some kind of virtual memory ? Thanks and Regards Rafael R. Machado 2013/4/9 Rafael Machado > Hi everyone > > > I'm having some problems with an uefi application that works perfec

[edk2] csm issues

2013-04-09 Thread Rafael Machado
Hi everyone I'm having some problems with an uefi application that works perfectly when CSM (Compatibility Support Module) is enabled but doesn't work when CSM is off. At the application I have a big array allocated to store some data. When CSM is on I have: current address: 00066FF8