From: Ard Biesheuvel <ard.biesheuvel@linaro.org>
To: Michael Zimmermann <sigmaepsilon92@gmail.com>
Cc: "edk2-devel@lists.01.org" <edk2-devel@lists.01.org>,
Leif Lindholm <leif.lindholm@linaro.org>
Subject: Re: [PATCH 2/2] ArmPkg/DefaultExceptionHandlerLib AARCH64: add minimal backtrace to crash dump
Date: Wed, 7 Sep 2016 12:32:34 +0100 [thread overview]
Message-ID: <CAKv+Gu8o3D7AvmXRZMK7OiXjMOZ-3DdcvB78vtwEAvWgKt7v0Q@mail.gmail.com> (raw)
In-Reply-To: <CAN9vWDK4u4uwR60C1DbzmrtFSz2GsoQKYWc5OJ98wUjqdq=fTw@mail.gmail.com>
On 7 September 2016 at 12:25, Michael Zimmermann
<sigmaepsilon92@gmail.com> wrote:
>> However, looking at this
>> code, this is still not sufficient to find the *next* frame pointer on
>> the stack.
> are you sure about that? this code looks like it does just that:
> https://github.com/torvalds/linux/blob/master/arch/arm/kernel/stacktrace.c
>
First of all, that comment makes little sense:
* With framepointer enabled, a simple function prologue looks like this:
* mov ip, sp
* stmdb sp!, {fp, ip, lr, pc}
I assume this should be
stmdb sp!, {fp, ip, lr}
* sub fp, ip, #4
*
* A simple function epilogue looks like this:
* ldm sp, {fp, sp, pc}
But then, it looks like GCC in ARM mode happens to generate stack
frames that are easily unwound.
However, in EDK2 we use Thumb2 primarily, and the code I snipped is
actually from a -fno-omit-frame-pointer Thumb2 build, and it does not
look like whatever is pushed onto the stack can be decoded
unambiguously like that.
> On Wed, Sep 7, 2016 at 1:03 PM, Ard Biesheuvel <ard.biesheuvel@linaro.org>
> wrote:
>>
>> On 7 September 2016 at 10:48, Michael Zimmermann
>> <sigmaepsilon92@gmail.com> wrote:
>> > nice, can we do this for ARM too? I usually need to add DEBUG((...))'s
>> > all
>> > over the place for hours until I found the reason for a fault.
>> >
>>
>> This is going to be tricky. Unlike AARCH64, which unambiguously
>> describes in the AAPCS how to link the stack frames using x29, the
>> AAPCS for ARM does not standardise this, and so different compilers
>> may do different things, also depending on whether you are executing
>> in ARM or Thumb mode. For instance, the following prologue was snipped
>> from a BdsDxe.dll build using GCC5 in Thumb2 mode.
>>
>> 00000254 <_ModuleEntryPoint>:
>> 254: e96d 7e02 strd r7, lr, [sp, #-8]!
>> 258: b084 sub sp, #16
>> 25a: af00 add r7, sp, #0
>>
>> Googling around, i can indeed find sources that describe how GCC uses
>> r7 as the frame pointer for Thumb2 code. However, looking at this
>> code, this is still not sufficient to find the *next* frame pointer on
>> the stack. IOW, there is no standardised layout to find the next frame
>> pointer inside the stack frame.
>>
>> The only way to do this reliably is using unwind tables, but this is
>> *much* more complicated than the 20 line patch that enables it for
>> AArch64
>>
>> Ard.
>>
>>
>> > On Wed, Sep 7, 2016 at 10:21 AM, Ard Biesheuvel
>> > <ard.biesheuvel@linaro.org>
>> > wrote:
>> >>
>> >> When dumping the CPU state after an unhandled fault, walk the stack
>> >> frames and decode the return addresses so we can show a minimal
>> >> backtrace. Unfortunately, we do not have sufficient information to
>> >> show the function names, but at least we can see the modules and the
>> >> return addresses inside the modules.
>> >>
>> >> Contributed-under: TianoCore Contribution Agreement 1.0
>> >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>> >> ---
>> >>
>> >> Example output:
>> >>
>> >> Synchronous Exception at 0x00000000F5EA4C8C
>> >>
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe/DEBUG/ArmVeNorFlashDxe.dll
>> >> loaded at 0x00000000F5E90000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe/DEBUG/ArmVeNorFlashDxe.dll
>> >> (0x00000000F5EA4AE8) loaded at 0x00000000F5E90000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe/DEBUG/ArmVeNorFlashDxe.dll
>> >> (0x00000000F5EA1BFC) loaded at 0x00000000F5E90000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe/DEBUG/ArmVeNorFlashDxe.dll
>> >> (0x00000000F5EA23F4) loaded at 0x00000000F5E90000
>> >> called from
>> >>
>> >> /work/jenkins/workspace/ap-uefi-bin/EDK2_ARCH/ARM/EDK2_BINARY/FatPkg/label/sas-sw/Build/Fat/RELEASE_GCC49/AARCH64/FatPkg/EnhancedFatDxe/Fat/DEBUG/Fat.dll
>> >> (0x00000000FADC454C) loaded at 0x00000000FADC3000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/MdeModulePkg/Core/Dxe/DxeMain/DEBUG/DxeCore.dll
>> >> (0x00000000FE47C868) loaded at 0x00000000FE471000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/IntelFrameworkModulePkg/Universal/BdsDxe/BdsDxe/DEBUG/BdsDxe.dll
>> >> (0x00000000FAE2C50C) loaded at 0x00000000FAE01000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/IntelFrameworkModulePkg/Universal/BdsDxe/BdsDxe/DEBUG/BdsDxe.dll
>> >> (0x00000000FAE2C674) loaded at 0x00000000FAE01000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/IntelFrameworkModulePkg/Universal/BdsDxe/BdsDxe/DEBUG/BdsDxe.dll
>> >> (0x00000000FAE2C264) loaded at 0x00000000FAE01000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/IntelFrameworkModulePkg/Universal/BdsDxe/BdsDxe/DEBUG/BdsDxe.dll
>> >> (0x00000000FAE36998) loaded at 0x00000000FAE01000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/IntelFrameworkModulePkg/Universal/BdsDxe/BdsDxe/DEBUG/BdsDxe.dll
>> >> (0x00000000FAE03668) loaded at 0x00000000FAE01000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/MdeModulePkg/Core/Dxe/DxeMain/DEBUG/DxeCore.dll
>> >> (0x00000000FE4733F0) loaded at 0x00000000FE471000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/MdeModulePkg/Core/Dxe/DxeMain/DEBUG/DxeCore.dll
>> >> (0x00000000FE4724A0) loaded at 0x00000000FE471000
>> >> called from
>> >>
>> >> /home/ard/build/edk2/Build/ArmVExpress-FVP-AArch64/DEBUG_GCC5/AARCH64/MdeModulePkg/Core/Dxe/DxeMain/DEBUG/DxeCore.dll
>> >> (0x00000000FE472024) loaded at 0x00000000FE471000
>> >> ---
>> >> .../AArch64/DefaultExceptionHandler.c | 21
>> >> +++++++++++++++++++++
>> >> 1 file changed, 21 insertions(+)
>> >>
>> >> diff --git
>> >>
>> >> a/ArmPkg/Library/DefaultExceptionHandlerLib/AArch64/DefaultExceptionHandler.c
>> >>
>> >> b/ArmPkg/Library/DefaultExceptionHandlerLib/AArch64/DefaultExceptionHandler.c
>> >> index 31fc936b21ff..84b442f2b6f4 100644
>> >> ---
>> >>
>> >> a/ArmPkg/Library/DefaultExceptionHandlerLib/AArch64/DefaultExceptionHandler.c
>> >> +++
>> >>
>> >> b/ArmPkg/Library/DefaultExceptionHandlerLib/AArch64/DefaultExceptionHandler.c
>> >> @@ -152,9 +152,30 @@ DefaultExceptionHandler (
>> >> CHAR8 *Pdb;
>> >> UINTN ImageBase;
>> >> UINTN PeCoffSizeOfHeader;
>> >> + UINT64 *Fp;
>> >> +
>> >> Pdb = GetImageName (SystemContext.SystemContextAArch64->ELR,
>> >> &ImageBase, &PeCoffSizeOfHeader);
>> >> if (Pdb != NULL) {
>> >> DEBUG ((EFI_D_ERROR, "%a loaded at 0x%016lx \n", Pdb,
>> >> ImageBase));
>> >> +
>> >> + Pdb = GetImageName (SystemContext.SystemContextAArch64->LR,
>> >> &ImageBase,
>> >> + &PeCoffSizeOfHeader);
>> >> + if (Pdb != NULL) {
>> >> + DEBUG ((EFI_D_ERROR, "called from %a (0x%016lx) loaded at
>> >> 0x%016lx \n",
>> >> + Pdb, SystemContext.SystemContextAArch64->LR, ImageBase));
>> >> + }
>> >> + for (Fp = (UINT64 *)SystemContext.SystemContextAArch64->FP;
>> >> + *Fp != 0;
>> >> + Fp = (UINT64 *)Fp[0]) {
>> >> + if (Fp[1] == SystemContext.SystemContextAArch64->LR) {
>> >> + continue;
>> >> + }
>> >> + Pdb = GetImageName (Fp[1], &ImageBase, &PeCoffSizeOfHeader);
>> >> + if (Pdb != NULL) {
>> >> + DEBUG ((EFI_D_ERROR, "called from %a (0x%016lx) loaded at
>> >> 0x%016lx \n",
>> >> + Pdb, Fp[1], ImageBase));
>> >> + }
>> >> + }
>> >> }
>> >> DEBUG_CODE_END ();
>> >>
>> >> --
>> >> 2.7.4
>> >>
>> >> _______________________________________________
>> >> edk2-devel mailing list
>> >> edk2-devel@lists.01.org
>> >> https://lists.01.org/mailman/listinfo/edk2-devel
>> >
>> >
>
>
next prev parent reply other threads:[~2016-09-07 11:32 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-07 8:21 [PATCH 1/2] BaseTools/tools_def AARCH64: enable frame pointers for DEBUG builds Ard Biesheuvel
2016-09-07 8:21 ` [PATCH 2/2] ArmPkg/DefaultExceptionHandlerLib AARCH64: add minimal backtrace to crash dump Ard Biesheuvel
2016-09-07 9:48 ` Michael Zimmermann
2016-09-07 11:03 ` Ard Biesheuvel
2016-09-07 11:25 ` Michael Zimmermann
2016-09-07 11:32 ` Ard Biesheuvel [this message]
2016-09-09 13:08 ` Ard Biesheuvel
2016-09-09 18:03 ` Michael Zimmermann
2016-09-09 16:03 ` Andrew Fish
2016-09-07 15:00 ` [PATCH 1/2] BaseTools/tools_def AARCH64: enable frame pointers for DEBUG builds Leif Lindholm
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-list from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAKv+Gu8o3D7AvmXRZMK7OiXjMOZ-3DdcvB78vtwEAvWgKt7v0Q@mail.gmail.com \
--to=devel@edk2.groups.io \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox