https://sourceware.org/bugzilla/show_bug.cgi?id=32666
--- Comment #4 from Indu Bhagat ---
After some discussion on the RFC series
https://sourceware.org/pipermail/binutils/2025-April/140432.html, the patchset
to address this issue has been posted
https://inbox.sourceware.org/binutils/202505280540
https://sourceware.org/bugzilla/show_bug.cgi?id=32952
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=32666
Indu Bhagat changed:
What|Removed |Added
Status|NEW |ASSIGNED
--
You are receiving this mai
https://sourceware.org/bugzilla/show_bug.cgi?id=32952
Indu Bhagat changed:
What|Removed |Added
Status|NEW |ASSIGNED
--
You are receiving this mai
https://sourceware.org/bugzilla/show_bug.cgi?id=32953
--- Comment #2 from Indu Bhagat ---
V2 discussion on mailing list:
https://inbox.sourceware.org/binutils/20250515210619.3808179-5-indu.bha...@oracle.com/
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=32879
Indu Bhagat changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
Component: gas
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
Currently gas skips processing the .cfi_same_value for SFrame generation
altogether (with no warning either). This would be OK to do if the directive
was
Component: gas
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
In context of SFrame generation, it is incorrect to simply ignore all
.cfi_undefined directives. We may ignore only those .cfi_undefined which are
for
https://sourceware.org/bugzilla/show_bug.cgi?id=32879
Indu Bhagat changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
https://sourceware.org/bugzilla/show_bug.cgi?id=32879
--- Comment #2 from Indu Bhagat ---
As per the documentation, .cfi_def_cfa_register modifies a rule for computing
CFA; the register is updated, but the offset remains the same.
So I think the following suffices:
if (last_fre)
sframe_fre_se
https://sourceware.org/bugzilla/show_bug.cgi?id=32666
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=32789
Indu Bhagat changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
https://sourceware.org/bugzilla/show_bug.cgi?id=32789
--- Comment #3 from Indu Bhagat ---
(In reply to Jens Remus from comment #1)
> I missed to mention that this assumes linker mainline with Indu's "[RFC 0/4]
> Fix relocatable links with SFrame section" patch series on top:
> https://inbox.sourc
https://sourceware.org/bugzilla/show_bug.cgi?id=32789
--- Comment #4 from Indu Bhagat ---
>From Weinan's testing so far, it seems 32789 is _not_ affecting arm64 kernel
space stack tracer effort
https://lore.kernel.org/linux-toolchains/20250214182439.gpavslsvgw4xy7sf@jpoimboe/T/#m66a88cc3eb323b9fd
https://sourceware.org/bugzilla/show_bug.cgi?id=32789
--- Comment #2 from Indu Bhagat ---
Right. In the SFrame stack trace format, the FDEs (function descriptor Entry)
are in sorted order of the start PC of the functions in the ELF file, if
SFRAME_F_FDE_SORTED flag is set. It is true that if th
https://sourceware.org/bugzilla/show_bug.cgi?id=32789
Indu Bhagat changed:
What|Removed |Added
CC||jpoimboe at redhat dot com
--
You are
https://sourceware.org/bugzilla/show_bug.cgi?id=32789
Indu Bhagat changed:
What|Removed |Added
CC||wnliu at google dot com
--
You are rec
https://sourceware.org/bugzilla/show_bug.cgi?id=32589
Indu Bhagat changed:
What|Removed |Added
CC||wnliu at google dot com
--
You are rec
https://sourceware.org/bugzilla/show_bug.cgi?id=32666
--- Comment #3 from Indu Bhagat ---
Proposed fix for this issue together with PR libsframe/32589
https://sourceware.org/pipermail/binutils/2025-March/139853.html
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=32589
--- Comment #2 from Indu Bhagat ---
Proposed fix for this issue together with PR ld/32666
https://sourceware.org/pipermail/binutils/2025-March/139853.html
--
You are receiving this mail because:
You are on the CC list for the bug.
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
$ cat gc-section.s
.text
.globl foo
.type foo, @function
foo:
.cfi_startproc
ret
.cfi_endproc
.size
https://sourceware.org/bugzilla/show_bug.cgi?id=32297
Indu Bhagat changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://sourceware.org/bugzilla/show_bug.cgi?id=32298
Indu Bhagat changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://sourceware.org/bugzilla/show_bug.cgi?id=32296
Indu Bhagat changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://sourceware.org/bugzilla/show_bug.cgi?id=32298
--- Comment #2 from Indu Bhagat ---
Fixed in 2.44
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=32297
--- Comment #3 from Indu Bhagat ---
Fixed in 2.44
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=32613
Indu Bhagat changed:
What|Removed |Added
CC||indu.bhagat at oracle dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=32666
--- Comment #2 from Indu Bhagat ---
Confirmed.
In presence of bad offsets, relocating SFrame sections with bad .rela.sframe
will result in bad data in the SFrame section.
The issue is that with relocatable link, we need to manually adjust (i
|UNCONFIRMED |NEW
Ever confirmed|0 |1
CC||indu.bhagat at oracle dot com
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=32589
--- Comment #1 from Indu Bhagat ---
Currently the section contents are not relocated before textual dump. Hence
the value 0.
$ readelf -r file.o
Relocation section '.rela.sframe' at offset 0x8728 contains 3 entries:
Offset Info
https://sourceware.org/bugzilla/show_bug.cgi?id=32589
Indu Bhagat changed:
What|Removed |Added
Status|NEW |ASSIGNED
--
You are receiving this mai
https://sourceware.org/bugzilla/show_bug.cgi?id=31654
Indu Bhagat changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
Component: libsframe
Assignee: indu.bhagat at oracle dot com
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
It will be useful for readelf / objdump to display the appropriate function
start address (see pc = 0x0, and the PC of the first FRE) in the SFrame
https://sourceware.org/bugzilla/show_bug.cgi?id=32298
--- Comment #1 from Indu Bhagat ---
https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=47c88752f9adf3815566ae52731867a3bb87b044
Fri, 1 Nov 2024 22:36:35 + (15:36 -0700)
commit 47c88752f9adf3815566ae52731867a3bb87b044
ld: generate
https://sourceware.org/bugzilla/show_bug.cgi?id=32297
--- Comment #2 from Indu Bhagat ---
https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=1785837a25709721358dab9ee16bbfee74f5f4d1
Fri, 1 Nov 2024 22:26:13 + (15:26 -0700)
commit 1785837a25709721358dab9ee16bbfee74f5f4d1
ld: fix PR/3
https://sourceware.org/bugzilla/show_bug.cgi?id=32296
--- Comment #1 from Indu Bhagat ---
Committed fix
https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=938fb512184d90d24b86c42ef53ef33c943cbe8b
Fri, 1 Nov 2024 22:36:03 + (15:36 -0700)
commit 938fb512184d90d24b86c42ef53ef33c943cbe8b
Priority: P2
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
The linker currently does not generate any SFrame stack trace information for
the .plt.got sections. There is SFrame information for other
https://sourceware.org/bugzilla/show_bug.cgi?id=32298
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=32297
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=32296
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=32297
--- Comment #1 from Indu Bhagat ---
A hand-crafted reproducer:
$ cat bar.c
void foo();
unsigned long var;
void test () {
var = (unsigned long)foo;
foo ();
}
$ cat foo.s
.globl foo
foo:
.section.note.GNU-stack,"",@progbits
$ gc
Priority: P2
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
On an IBT enabled system, a user reported the following failure:
# cat test-noop.c
int main() {}
# gcc -O2 -Wa,-gsframe -o test-noop test
Priority: P2
Component: ld
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
$ cat foo.s
.text
.p2align 4
.globl bar
.type bar, @function
bar:
.cfi_startproc
https://sourceware.org/bugzilla/show_bug.cgi?id=32187
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
In aarch64, the following code pattern may be used for managing stack
statically:
mov x16, IMM
add sp, sp, x16
Currently, when using --scfi=experimental, the above
https://sourceware.org/bugzilla/show_bug.cgi?id=32161
Indu Bhagat changed:
What|Removed |Added
CC||indu.bhagat at oracle dot com
--
You
https://sourceware.org/bugzilla/show_bug.cgi?id=31284
Indu Bhagat changed:
What|Removed |Added
Resolution|--- |FIXED
Status|NEW
https://sourceware.org/bugzilla/show_bug.cgi?id=31213
Indu Bhagat changed:
What|Removed |Added
Resolution|--- |FIXED
Status|NEW
https://sourceware.org/bugzilla/show_bug.cgi?id=31213
--- Comment #2 from Indu Bhagat ---
Further improvements in the diagnostics issued when skipping SFrame FDE
generation were done recently:
https://sourceware.org/pipermail/binutils/2024-June/135056.html
--
You are receiving this mail because
https://sourceware.org/bugzilla/show_bug.cgi?id=31654
--- Comment #2 from Indu Bhagat ---
This is repeatable on other arches too using binutils-gdb master.
One aarch64:
$ cat test.s
.cfi_startproc
.long 0
.cfi_def_cfa_offset 8
.long 0
.cfi_rel_offset 29,
https://sourceware.org/bugzilla/show_bug.cgi?id=31108
Indu Bhagat changed:
What|Removed |Added
Resolution|--- |INVALID
Status|NEW
https://sourceware.org/bugzilla/show_bug.cgi?id=30014
Indu Bhagat changed:
What|Removed |Added
Resolution|--- |FIXED
Status|NEW
https://sourceware.org/bugzilla/show_bug.cgi?id=31326
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
Priority: P2
Component: gas
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
As per the x86 ISA manual:
- 32-bit operands generate a 32-bit result, zero-extended to a 64-bit result
in the destination
https://sourceware.org/bugzilla/show_bug.cgi?id=31284
Indu Bhagat changed:
What|Removed |Added
Priority|P2 |P3
--
You are receiving this mail beca
Component: gas
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
Currently, if an indirect jump is seen, GCFG (a CFG of ginsns) cannot be
created. GCFG creation is necessary for SCFI (invoked via --scfi=experimental
on
https://sourceware.org/bugzilla/show_bug.cgi?id=31284
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=31108
--- Comment #5 from Indu Bhagat ---
I proposed a backport at gcc-patches mailing list
https://gcc.gnu.org/pipermail/gcc-patches/2023-December/639444.html
--
You are receiving this mail because:
You are on the CC list for the bug.
https://sourceware.org/bugzilla/show_bug.cgi?id=30914
Indu Bhagat changed:
What|Removed |Added
CC||indu.bhagat at oracle dot com
https://sourceware.org/bugzilla/show_bug.cgi?id=31213
Indu Bhagat changed:
What|Removed |Added
Priority|P2 |P3
--
You are receiving this mail beca
https://sourceware.org/bugzilla/show_bug.cgi?id=31213
Indu Bhagat changed:
What|Removed |Added
Assignee|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
: normal
Priority: P2
Component: gas
Assignee: unassigned at sourceware dot org
Reporter: indu.bhagat at oracle dot com
Target Milestone: ---
When the user specifies --gsframe, GAS consumes the .cfi_* directives to
generate the SFrame stack trace information
https://sourceware.org/bugzilla/show_bug.cgi?id=31108
--- Comment #4 from Indu Bhagat ---
Yes, the toplevel Makefile.* in GCC will need the install-* dependencies in GCC
13.2.0.
The original commit
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=3de627ffe4b51b3d82acdb5fb04c189978697832
chose to sk
https://sourceware.org/bugzilla/show_bug.cgi?id=30588
--- Comment #3 from Indu Bhagat ---
(In reply to Tom Tromey from comment #2)
> (In reply to Indu Bhagat from comment #1)
> > What is the incentive to support --disable-libsframe ?
>
> I have no need for it.
>
> libsframe has a source depende
https://sourceware.org/bugzilla/show_bug.cgi?id=30588
--- Comment #1 from Indu Bhagat ---
What is the incentive to support --disable-libsframe ?
Currently, the GNU ld, objdump, readelf use libsframe to read/write/dump SFrame
data.
Thinking aloud about this would imply. The SFrame section (.sfra
https://sourceware.org/bugzilla/show_bug.cgi?id=30014
Indu Bhagat changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
https://sourceware.org/bugzilla/show_bug.cgi?id=30014
--- Comment #2 from Indu Bhagat ---
Thanks for reporting this issue and providing all the details. Very helpful.
The issue seems to be that the applicable dependency (install-strip-bfd:
maybe-install-strip-libsframe) was missing.
As for libt
https://sourceware.org/bugzilla/show_bug.cgi?id=30014
--- Comment #1 from Indu Bhagat ---
Created attachment 14605
--> https://sourceware.org/bugzilla/attachment.cgi?id=14605&action=edit
Proposed fix for pr 30014
Proposed fix. Will test this patch on my end.
--
You are receiving this mail be
https://sourceware.org/bugzilla/show_bug.cgi?id=29856
--- Comment #1 from Indu Bhagat ---
Thanks for the report. The misaligned load here is for access to an element in
the in-memory abstraction provided by libsframe (namely struct
sframe_frame_row_entry). Now there are ways to resolve that (e.g.
|unassigned at sourceware dot org |indu.bhagat at oracle
dot com
--
You are receiving this mail because:
You are on the CC list for the bug.
70 matches
Mail list logo
