Page MenuHomeFreeBSD
Differential D59650

trap_cfi(9): introduce arm implementation
Needs ReviewPublic
Actions

Authored by mchoo on Sun, Sep 13, 6:19 PM.
Tags
None
Referenced Files
F173065286: D59650.id186803.diff
Wed, Sep 23, 7:43 AM
F172962785: D59650.diff
Tue, Sep 22, 11:07 AM
Unknown Object (File)
Tue, Sep 22, 1:38 AM
Unknown Object (File)
Tue, Sep 22, 12:37 AM
Unknown Object (File)
Mon, Sep 21, 9:16 PM
Unknown Object (File)
Mon, Sep 21, 5:06 PM
Unknown Object (File)
Mon, Sep 21, 4:55 PM
Unknown Object (File)
Sun, Sep 20, 8:43 PM
View All 27 Files
Subscribers
imp
mmel
oh

Details

Reviewers
jhb
andrew
Group Reviewers
ARM
Summary

MFC after: 2 weeks
MFC to: stable/14, stable/15
Sponsored by: FreeBSD Foundation

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Skipped
Unit
No Test Coverage
Build Status
Buildable 76941
Build 73824: arc lint + arc unit

Event Timeline

mchoo created this revision.Sun, Sep 13, 6:19 PM
Herald added a reviewer: andrew. · View Herald TranscriptSun, Sep 13, 6:19 PM
Herald added a subscriber: imp. · View Herald Transcript
mchoo requested review of this revision.Sun, Sep 13, 6:19 PM
Harbormaster completed remote builds in B76857: Diff 186620.Sun, Sep 13, 6:19 PM
mchoo added a parent revision: D59649: trap_cfi(9): introduce amd64 implementation.Sun, Sep 13, 6:19 PM
mchoo added a child revision: D59651: trap_cfi(9): introduce powerpc implemenation.
mchoo added a reviewer: ARM.Sun, Sep 13, 6:33 PM
mchoo edited the summary of this revision. (Show Details)Sun, Sep 13, 9:07 PM
mchoo updated this revision to Diff 186792.Tue, Sep 15, 5:35 PM

Add second part

Harbormaster completed remote builds in B76941: Diff 186792.Tue, Sep 15, 5:35 PM
mchoo added a comment.Tue, Sep 15, 5:36 PM

I wanted to test this through QEMU, but it seems like armv7 minidump is broken (it cannot locate allproc, stoppcbs, etc)

mchoo mentioned this in D59651: trap_cfi(9): introduce powerpc implemenation.Tue, Sep 15, 5:37 PM
mchoo retitled this revision from trap_cfi(9): add assembly interface for arm to trap_cfi(9): introduce arm implementation.
mchoo updated this revision to Diff 186803.Tue, Sep 15, 7:40 PM

cleanup

Harbormaster completed remote builds in B76950: Diff 186803.Tue, Sep 15, 7:40 PM
oh added a subscriber: oh.Tue, Sep 15, 9:12 PM
mchoo added a comment.Wed, Sep 16, 3:16 AM
In D59650#1370098, @mchoo wrote:

I wanted to test this through QEMU, but it seems like armv7 minidump is broken (it cannot locate allproc, stoppcbs, etc)

D59716 allowed me extract kernel dump from QEMU, and the frame unwinding works as expected. Ready for review.

mmel added a subscriber: mmel.Sun, Sep 20, 8:10 AM

Of course, I don't know anything about annotating the trap frame.
However, adding instructions to the absolutely hot path(s) in the system to fix the debugger doesn't seem like a good idea to me.

Why can't stop lying about the annotation of the stored registers and properly annotate the entire trapframe (including the SPSR), allowing the debugger to do its job?

sys/arm/arm/exception.S
349–350

Why has this change been made? It doesn't make sense to me.

mmel added a comment.Mon, Sep 21, 6:37 AM

I forgot about it yesterday. At worst, you can use .cfi_escape + DW_CFA_val_expression to select proper LR and SP directly in debugger.

Revision Contents
Changeset List

PathSize
sys/
arm/
arm/
174 lines
9 lines
5 lines
include/
1 line
2 lines
55 lines
conf/
1 line

Diff 186792

View Options

sys/arm/arm/exception.S

Loading...
View Options

sys/arm/arm/swtch.S

Loading...
View Options

sys/arm/arm/vm_machdep.c

Loading...
View Options

sys/arm/include/cpu.h

Loading...
View Options

sys/arm/include/frame.h

Loading...
View Options

sys/arm/include/trap_cfi.h

Loading...
View Options

sys/conf/ldscript.arm

Loading...