Re: [PATCH edk2-platforms 0/3] drop GUIDs from NOR flash bank descriptors

[Ard Biesheuvel <ard.biesheuvel@linaro.org>]

On Wed, 21 Nov 2018 at 14:48, Thomas Abraham <thomas.abraham@arm.com> wrote:
>
> Hi Ard,
>
> On Wed, Nov 21, 2018 at 5:31 PM Ard Biesheuvel
> <ard.biesheuvel@linaro.org> wrote:
> >
> > Align edk2-platform with upcoming changes to EDK2 to get rid of per-bank
> > NOR flash GUIDs.
> >
> > Ard Biesheuvel (3):
> >   Platform/ARM: replace hardcoded VenHW() device paths referring to NOR
> >     flash
> >   Silicon/SynQuacer: drop per-bank NOR flash GUIDs
> >   Platform/ARM: drop per-bank NOR flash GUIDs
> >
> >  Platform/ARM/JunoPkg/ArmJuno.dec                   |  2 +-
> >  Platform/ARM/JunoPkg/ArmJuno.dsc                   |  2 +-
> >  .../JunoPkg/Library/NorFlashJunoLib/NorFlashJuno.c |  2 --
> >  .../ARM/SgiPkg/Library/NorFlashLib/NorFlashLib.c   |  2 --
> >  Platform/ARM/SgiPkg/SgiPlatform.dsc                |  2 +-
> >  Platform/ARM/VExpressPkg/ArmVExpress-CTA15-A7.dsc  |  2 +-
> >  Platform/ARM/VExpressPkg/ArmVExpress.dsc.inc       |  2 +-
> >  .../NorFlashArmVExpressLib/NorFlashArmVExpress.c   |  4 ----
> >  .../SynQuacer/Drivers/Fip006Dxe/NorFlashDxe.c      | 14 +++++++-------
> >  .../SynQuacer/Drivers/Fip006Dxe/NorFlashDxe.h      |  3 +++
> >  .../NorFlashSynQuacerLib/NorFlashSynQuacer.c       |  6 ------
> >  11 files changed, 15 insertions(+), 26 deletions(-)
>
> Tested this patch series and "[PATCH v2 0/5] ArmPlatformPkg,
> ArmVirtPkg: discover NOR flash banks from DTB" patch series on the
> Juno board. With these patches applied, the boot fails on Juno board
> with the following messages. I have not yet tried to debug the issue
> but wanted to let you know this.
>
> [...]
> Loading driver at 0x000F830C000 EntryPoint=0x000F831B2AC IScsiDxe.efi
> add-symbol-file
> /home/thopan01/devel/juno/uefi/uefi/edk2/Build/ArmJuno/DEBUG_GCC5/AARCH64/MdeModulePkg/Universal/Network/Udp4Dxe/Udp4Dxe/DEBUG/Udp4Dxe.dll
> 0xF8300000
> Loading driver at 0x000F82FF000 EntryPoint=0x000F8306DF0 Udp4Dxe.efi
> add-symbol-file
> /home/thopan01/devel/juno/uefi/uefi/edk2/Build/ArmJuno/DEBUG_GCC5/AARCH64/Platform/ARM/Drivers/FdtPlatformDxe/FdtPlatformDxe/DEBUG/FdtPlatformDxe.dll
> 0xF82EE000
> Loading driver at 0x000F82ED000 EntryPoint=0x000F82F76EC FdtPlatformDxe.efi
> Found image: fip in block 5.
> Found image: norkern in block 20.
> Found image: ramdisk.img in block 116.
> Found image: hdlcdclk in block 151.
> Found image: selftest in block 152.
> Found image: board.dtb in block 156.
> Found image: scp_bl1 in block 249.
> Found image: bl1 in block 251.
> Found image: startup.nsh in block 252.
> ASSERT [BootMonFs]
> /home/thopan01/devel/juno/uefi/uefi/edk2/MdePkg/Library/BaseLib/String.c(173):
> ((UINTN) String & 0x00000001) == 0
>

Hi Thomas,

Thanks for testing.

Could you please apply this patch and re-test:

--- a/Platform/ARM/VExpressPkg/ArmVExpress.dsc.inc
+++ b/Platform/ARM/VExpressPkg/ArmVExpress.dsc.inc
@@ -299,7 +299,7 @@
 !if $(TARGET) == RELEASE
   gEfiMdePkgTokenSpaceGuid.PcdDebugPropertyMask|0x21
 !else
-  gEfiMdePkgTokenSpaceGuid.PcdDebugPropertyMask|0x2f
+  gEfiMdePkgTokenSpaceGuid.PcdDebugPropertyMask|0x3f
 !endif

   #  DEBUG_INIT      0x00000001  // Initialization

This will produce a backtrace when an ASSERT() is triggered, and from
this we can figure out what the call stack looks like exactly (The
ASSERT() occurs in StrLen (), which is called with a misaligned
unicode string, but I cannot figure out where the call is made)

Please share the backtrace and all the .dlls listed in it if you can please.