On 9 September 2015 at 18:49, Michael Zimmermann <[email protected]> wrote: > thx, the patches work just fine.(they apply and the runtime warnings are > gone).
OK, great. May I take that as a tested-by? Which GCC version are you using btw? > if edk2 patches aren't that usable, how the maintainers test it? because > usually people don't post git links. > > On Wed, Sep 9, 2015 at 6:31 PM, Ard Biesheuvel <[email protected]> > wrote: >> >> On 9 September 2015 at 18:21, Michael Zimmermann >> <[email protected]> wrote: >> > what I meant is that git am fails: >> > Applying: BaseTools/GenFw: remove ARM and RVCT references from ELF64 >> > code >> > error: patch failed: BaseTools/Source/C/GenFw/Elf64Convert.c:360 >> > error: BaseTools/Source/C/GenFw/Elf64Convert.c: patch does not apply >> > Patch failed at 0001 BaseTools/GenFw: remove ARM and RVCT references >> > from >> > ELF64 code >> > The copy of the patch that failed is found in: >> > >> > /media/Data/repositories/git/efidroid/uefi/edk2/.git/rebase-apply/patch >> > When you have resolved this problem, run "git am --continue". >> > If you prefer to skip this patch, run "git am --skip" instead. >> > To restore the original branch and stop patching, run "git am --abort". >> > >> > maybe I did sth. wrong?(I've never used a mailing list patch before). I >> > downloaded the mbox from gmane: >> > http://download.gmane.org/gmane.comp.bios.edk2.devel/2036/2037 >> > >> >> No, it is just painful, with the CR/LF conversion etc >> >> What I usually do is >> >> $ cat |todos |git am >> >> and then paste it, followed by Ctrl-D. That will usually work, unless >> the patch adds or removes files. >> >> I pushed the patches here: >> >> https://git.linaro.org/people/ard.biesheuvel/uefi-next.git/shortlog/refs/heads/arm-memprot >> >> So after applying you need to >> - git clean -dxf BaseTools/ ; make -C BaseTools >> - rm Conf/tools_def.txt ; source edksetup.sh >> >> to make sure everything is up to date after the changes >> >> -- >> Ard. >> >> > On Wed, Sep 9, 2015 at 6:17 PM, Ard Biesheuvel >> > <[email protected]> >> > wrote: >> >> >> >> >> >> >> >> On 9 sep. 2015, at 18:10, Michael Zimmermann <[email protected]> >> >> wrote: >> >> >> >> Yes I'm using 32bit ARM :) >> >> thx for the patches - unfortunatelythe patches fail for me. >> >> >> >> >> >> Did you regenerate Conf/tools_def.txt and rebuild the BaseTools/ ? >> >> >> >> On Wed, Sep 9, 2015 at 5:33 PM, Ard Biesheuvel >> >> <[email protected]> >> >> wrote: >> >>> >> >>> On 9 September 2015 at 17:26, Gao, Liming <[email protected]> >> >>> wrote: >> >>> > Michael: >> >>> > Do you use the linker script BaseTools/Scripts/GccBase.lds and -z >> >>> > common-page-size=4096? >> >>> > >> >>> >> >>> Are you building for 32-bit ARM by any chance? That does not have this >> >>> feature wired up yet. I posted a v2 of my series that addresses this >> >>> yesterday. >> >>> >> >>> >> >>> >> >>> > From: Michael Zimmermann [mailto:[email protected]] >> >>> > Sent: Wednesday, September 9, 2015 3:22 PM >> >>> > To: Yao, Jiewen >> >>> > Cc: Gao, Liming; [email protected] >> >>> > Subject: Re: [edk2] Section Alignment of elf binaries compiled with >> >>> > GCC(Linux) >> >>> > >> >>> > Unfortunately I missed the replies but I debugged this problem >> >>> > further >> >>> > and the Problem is that GenFw set's the alignment based on >> >>> > "sh_addralign" in >> >>> > the Elf header. >> >>> > >> >>> > The 'common-page-size' flag doesn't change the value of this field >> >>> > though. what it does change is the Alignment value of the Program >> >>> > Headers: >> >>> > Program Headers: >> >>> > Type Offset VirtAddr PhysAddr FileSiz MemSiz Flg >> >>> > Align >> >>> > LOAD 0x010000 0x00000000 0x00000000 0x042f8 0x042f8 R E >> >>> > 0x10000 >> >>> > LOAD 0x0142f8 0x000142f8 0x000142f8 0x000d4 0x000f4 RW >> >>> > 0x10000 >> >>> > GNU_STACK 0x000000 0x00000000 0x00000000 0x00000 0x00000 RWE >> >>> > 0x10 >> >>> > the section alignment still stays at 0/1/4. >> >>> > >> >>> > On Tue, Aug 18, 2015 at 4:49 AM, Yao, Jiewen >> >>> > <[email protected]<mailto:[email protected]>> wrote: >> >>> > OK >> >>> > >> >>> > -----Original Message----- >> >>> > From: Gao, Liming >> >>> > Sent: Tuesday, August 18, 2015 10:48 AM >> >>> > To: Yao, Jiewen; Michael Zimmermann; >> >>> > [email protected]<mailto:[email protected]> >> >>> > Subject: RE: [edk2] Section Alignment of elf binaries compiled with >> >>> > GCC(Linux) >> >>> > >> >>> > Jiewen: >> >>> > The updated message is useful. I suggest to change error level >> >>> > from >> >>> > EFI_D_ERROR to EFI_D_INFO. >> >>> > >> >>> > Thanks >> >>> > Liming >> >>> > -----Original Message----- >> >>> > From: Yao, Jiewen >> >>> > Sent: Tuesday, August 18, 2015 10:46 AM >> >>> > To: Gao, Liming; Michael Zimmermann; >> >>> > [email protected]<mailto:[email protected]> >> >>> > Subject: RE: [edk2] Section Alignment of elf binaries compiled with >> >>> > GCC(Linux) >> >>> > >> >>> > Hi >> >>> > How about we update debug message like below: >> >>> > >> >>> > Index: PropertiesTable.c >> >>> > =================================================================== >> >>> > --- PropertiesTable.c (revision 18191) >> >>> > +++ PropertiesTable.c (working copy) >> >>> > @@ -1120,7 +1120,7 @@ >> >>> > >> >>> > SetPropertiesTableSectionAlignment (SectionAlignment); >> >>> > if ((SectionAlignment & >> >>> > (EFI_ACPI_RUNTIME_PAGE_ALLOCATION_ALIGNMENT >> >>> > - 1)) != 0) { >> >>> > - DEBUG ((EFI_D_ERROR, "!!!!!!!! InsertImageRecord - Section >> >>> > Alignment(0x%x) is not %dK !!!!!!!!\n", >> >>> > + DEBUG ((EFI_D_ERROR, "!!!!!!!! UEFI2.5 PropertiesTable - >> >>> > Runtime >> >>> > Driver Section Alignment(0x%x) is not %dK !!!!!!!!\n", >> >>> > SectionAlignment, EFI_ACPI_RUNTIME_PAGE_ALLOCATION_ALIGNMENT >> >>> > >> >> >>> > 10)); >> >>> > PdbPointer = PeCoffLoaderGetPdbPointer ((VOID*) (UINTN) >> >>> > ImageAddress); >> >>> > if (PdbPointer != NULL) { >> >>> > >> >>> > >> >>> > >> >>> > >> >>> > -----Original Message----- >> >>> > From: edk2-devel >> >>> > >> >>> > [mailto:[email protected]<mailto:[email protected]>] >> >>> > On Behalf Of Gao, Liming >> >>> > Sent: Tuesday, August 18, 2015 10:39 AM >> >>> > To: Michael Zimmermann; >> >>> > [email protected]<mailto:[email protected]> >> >>> > Subject: Re: [edk2] Section Alignment of elf binaries compiled with >> >>> > GCC(Linux) >> >>> > >> >>> > Hi, >> >>> > This is a warning message that describes the runtime driver >> >>> > alignment >> >>> > is not 4K. >> >>> > >> >>> > UEFI PropertiesTable table feature expects all runtime driver >> >>> > alignment is 4K. When DxeCore loads Runtime driver, it will check >> >>> > its >> >>> > alignment and report such warning message if it doesn't meet with >> >>> > the >> >>> > alignment. If you want to enable this feature, you need to make sure >> >>> > all >> >>> > runtime driver at 4K. If you don't enable it, you can just ignore >> >>> > this >> >>> > message. >> >>> > >> >>> > To configure runtime driver with 4K alignment, you can modify DSC >> >>> > file to add the following section. >> >>> > [BuildOptions.common.EDKII.DXE_RUNTIME_DRIVER] >> >>> > GCC: *_*_*_DLINK_FLAGS = -z common-page-size=0x1000 >> >>> > MSFT: *_*_*_DLINK_FLAGS = /ALIGN:4096 >> >>> > >> >>> > Thanks >> >>> > Liming >> >>> > -----Original Message----- >> >>> > From: edk2-devel >> >>> > >> >>> > [mailto:[email protected]<mailto:[email protected]>] >> >>> > On Behalf Of Michael Zimmermann >> >>> > Sent: Sunday, August 16, 2015 12:32 PM >> >>> > To: [email protected]<mailto:[email protected]> >> >>> > Subject: [edk2] Section Alignment of elf binaries compiled with >> >>> > GCC(Linux) >> >>> > >> >>> > When booting(a new device I'm currently working on) I get these >> >>> > warnings: >> >>> > !!!!!!!! InsertImageRecord - Section Alignment(0x20) is not 4K >> >>> > !!!!!!!! >> >>> > >> >>> > the warning is raised by >> >>> > "MdeModulePkg/Core/Dxe/Misc/PropertiesTable.c" >> >>> > and when compiling using GCC from Linux, the SectionAlignment is set >> >>> > by >> >>> > "BaseTools/Source/C/GenFw/Elf32Convert.c". >> >>> > >> >>> > I checked the resulting binaries using "readpe" and indeed they have >> >>> > a >> >>> > SectionAlignment of 0x20, while the precompiled binaries like >> >>> > Shell.efi have >> >>> > a Alignment of 0x1000(4K). >> >>> > >> >>> > So, is this a bug of my GCC compiler or of EDK2? >> >>> > _______________________________________________ >> >>> > edk2-devel mailing list >> >>> > [email protected]<mailto:[email protected]> >> >>> > https://lists.01.org/mailman/listinfo/edk2-devel >> >>> > _______________________________________________ >> >>> > edk2-devel mailing list >> >>> > [email protected]<mailto:[email protected]> >> >>> > https://lists.01.org/mailman/listinfo/edk2-devel >> >>> > >> >>> > _______________________________________________ >> >>> > edk2-devel mailing list >> >>> > [email protected] >> >>> > https://lists.01.org/mailman/listinfo/edk2-devel >> >> >> >> >> > > > _______________________________________________ edk2-devel mailing list [email protected] https://lists.01.org/mailman/listinfo/edk2-devel
