Following the removal of gdbarch_sdb_reg_to_regnum,
i386_dbx_reg_to_regnum is only used by i386_svr4_dwarf_reg_to_regnum to
handle SSE and MMX registers. Remove it and inline the relevant bits in
i386_svr4_dwarf_reg_to_regnum.
Change-Id: Id74fad6ef6798c4cd061905f1c01eadd90e0a118
Approved-By: Andrew Burgess <aburgess@redhat.com>
Following the removal of the COFF debug info support,
gdbarch_sdb_reg_to_regnum is no longer used, remove it.
Change-Id: I2cb43465f1fdf74863edfa4dd00fd5f28a5a26bd
Approved-By: Andrew Burgess <aburgess@redhat.com>
These comments appear to be stale and no longer relevant.
Change-Id: I3470969b0c0f38d809fe074ffab93ac91202de18
Approved-By: Andrew Burgess <aburgess@redhat.com>
On s390x-linux (SLES 15 SP5), I'm running into:
...
FAIL: gdb.base/siginfo.exp: backtrace for nexti (pattern 2)
FAIL: gdb.base/siginfo.exp: step out of handler
...
The first FAIL is caused by a failure to unwind:
...
(gdb) bt^M
#0 handler (sig=26, info=0x3ffffffe428, context=0x3ffffffe4a8) at \
gdb.base/siginfo.c:31^M
Backtrace stopped: Cannot access memory at address 0x1a00000088^M
(gdb)
...
In contrast, on x86_64-linux I get instead:
...
(gdb) bt^M
#0 handler (sig=26, info=0x7fffffffc170, context=0x7fffffffc040) at \
gdb.base/siginfo.c:31^M
#1 <signal handler called>^M
#2 0x0000000000401201 in main () at gdb.base/siginfo.c:67^M
(gdb)
...
The memory access error is triggered here in s390_sigtramp_frame_unwind_cache:
...
/* Restore the previous frame's SP. */
prev_sp = read_memory_unsigned_integer (
info->saved_regs[S390_SP_REGNUM].addr (),
word_size, byte_order);
...
while trying to read an "Old-style RT frame" (for syscall sigreturn).
The problem is that we actually have a "New-style RT frame" (for syscall
rt_sigreturn).
[ See linux kernel source file arch/s390/kernel/signal.c for a detailed
explanation of the two. ]
The choice between the two is made earlier in that same function:
...
/* New-style RT frame:
retcode + alignment (8 bytes)
siginfo (128 bytes)
ucontext (contains sigregs at offset 5 words). */
if (next_ra == next_cfa)
{
...
}
/* Old-style RT frame and all non-RT frames:
old signal mask (8 bytes)
pointer to sigregs. */
else
...
I'm not sure why the check gives the wrong result, but I noticed that
s390_sigtramp_frame_sniffer is able to distinguish between the two, so fix
this by:
- factoring out new function s390_sigtramp_p out of
s390_sigtramp_frame_sniffer, and
- using s390_sigtramp_p in s390_sigtramp_frame_unwind_cache to distinguish
between the "Old-style RT frame" and "New-style RT frame".
This fixes the backtrace.
The second failure is:
...
(gdb) step^M
32 } /* handler */^M
1: x/i $pc^M
=> 0x1000772 <handler+50>: nopr^M
(gdb) step^M
0x000003fffdffe490 in __kernel_rt_sigreturn ()^M
1: x/i $pc^M
=> 0x3fffdffe490 <__kernel_rt_sigreturn>: svc 173^M
(gdb) FAIL: gdb.base/siginfo.exp: step out of handler
...
There is some code in process_event_stop_test that is supposed to trigger:
...
if (ecs->event_thread->control.step_range_end != 1
&& (ecs->event_thread->control.step_over_calls == STEP_OVER_UNDEBUGGABLE
|| ecs->event_thread->control.step_over_calls == STEP_OVER_ALL)
&& get_frame_type (frame) == SIGTRAMP_FRAME)
{
infrun_debug_printf ("stepped into signal trampoline");
/* The inferior, while doing a "step" or "next", has ended up in
a signal trampoline (either by a signal being delivered or by
the signal handler returning). Just single-step until the
inferior leaves the trampoline (either by calling the handler
or returning). */
keep_going (ecs);
return;
}
...
but it doesn't because frame is a NORMAL_FRAME instead of a SIGTRAMP_FRAME.
This is caused by the "dwarf2" unwinder triggering, which has higher priority
than the "s390 linux sigtramp" unwinder:
...
(gdb) maint info frame-unwinders
Name Type Class Enabled
dummy DUMMY_FRAME GDB Y
dwarf2 tailcall TAILCALL_FRAME DEBUGINFO Y
inline INLINE_FRAME GDB Y
jit NORMAL_FRAME EXTENSION Y
python NORMAL_FRAME EXTENSION Y
dwarf2 NORMAL_FRAME DEBUGINFO Y
dwarf2 signal SIGTRAMP_FRAME DEBUGINFO Y
s390 linux sigtramp SIGTRAMP_FRAME ARCH Y
s390 stub NORMAL_FRAME ARCH Y
s390 prologue NORMAL_FRAME ARCH Y
...
I found some code in dwarf2_frame_sniffer:
...
/* On some targets, signal trampolines may have unwind information.
We need to recognize them so that we set the frame type
correctly. */
if (fde->cie->signal_frame
|| dwarf2_frame_signal_frame_p (get_frame_arch (this_frame),
this_frame))
return self->type () == SIGTRAMP_FRAME;
...
and an example implementation i386_linux_dwarf_signal_frame_p, and after
copying this approach, indeed the stepping failure was fixed, but the
backtrace broken again.
Instead, fix this by giving the "s390 linux sigtramp" unwinder a higher
priority:
...
(gdb) maint info frame-unwinders
Name Type Class Enabled
dummy DUMMY_FRAME GDB Y
dwarf2 tailcall TAILCALL_FRAME DEBUGINFO Y
inline INLINE_FRAME GDB Y
jit NORMAL_FRAME EXTENSION Y
python NORMAL_FRAME EXTENSION Y
s390 linux sigtramp SIGTRAMP_FRAME ARCH Y
dwarf2 NORMAL_FRAME DEBUGINFO Y
dwarf2 signal SIGTRAMP_FRAME DEBUGINFO Y
s390 stub NORMAL_FRAME ARCH Y
s390 prologue NORMAL_FRAME ARCH Y
...
Also fixes test-case gdb.base/sigaltstack.exp and gdb.base/sigbpt.exp.
Tested on s390x-linux.
Reviewed-By: Keith Seitz <keiths@redhat.com>
PR tdep/33708
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=33708
aarch64-pe +FAIL: gas/aarch64/fix-adj
arm-pe +FAIL: MVE vmlas instructions
arm-wince-pe +FAIL: MVE vmlas instructions
These new tests use ELF directives.
* testsuite/gas/aarch64/fix-adj.d: Only run on ELF targets.
* testsuite/gas/arm/mve-vmlas.d: Likewise.
so that the memory can be reclaimed easily when reset_resolved_wilds
throws it all away.
* ldlang.c (matching_obstack): New static var.
(add_matching_section): Rewrite to use matching_obstack.
(lang_init): Init matching_obstack.
(reset_resolved_wilds): Free matching_obstack.
These are small, and making them inline lets me call
lang_statement_append without adding a forward declaration in the
next patch.
* ldlang.c (lang_for_each_statement, lang_list_init),
(lang_statement_append): Move to..
* ldlang.h: ..here making them inline functions.
(bfd_input_just_syms): Group with other inlines.
Replace bfd_map_over_sections with iteration over gdb_bfd_sections.
Change-Id: I92ba6b5ef9e9ab3d2ebe364373aeaf459fd6e34c
Approved-By: Tom Tromey <tom@tromey.com>
Replace bfd_map_over_sections with iteration over gdb_bfd_sections.
Change core_process_module_section's signature to take a cpms_data
reference instead of a void pointer.
Change-Id: Ic4f9a12c9c1479799ec87658fd88490106b61836
Approved-By: Tom Tromey <tom@tromey.com>
Replace bfd_map_over_sections with iteration over gdb_bfd_sections.
Rename copy_sections to copy_section and change its signature to take
explicit parameters instead of the callback-style void pointer.
Change-Id: I30f1c3c051415722f2220b7fba48103101b640e8
Approved-By: Tom Tromey <tom@tromey.com>
Replace bfd_map_over_sections with iteration over gdb_bfd_sections.
Change-Id: If2e1141f2e9345fe50e2a63ce8f8682e82a60f21
Approved-By: Tom Tromey <tom@tromey.com>
Replace bfd_map_over_sections with iteration over gdb_bfd_sections.
Change-Id: I34182a67ee52da479c1868a5880709c401cbf34e
Approved-By: Tom Tromey <tom@tromey.com>
Replace bfd_map_over_sections with iteration over gdb_bfd_sections.
Change-Id: Ib236b364dc81f43cab8d27271d413eb61d6c703e
Approved-By: Tom Tromey <tom@tromey.com>
On aarch64-linux, I run into:
...
(gdb) info auxv^M
...
26 AT_HWCAP2 Extension of AT_HWCAP 0x181^M
29 ??? 0x0^M
31 AT_EXECFN File name of executable 0xffffffffffb9 ...
...
28 AT_RSEQ_ALIGN rseq allocation alignment 32^M
0 AT_NULL End of vector 0x0^M
(gdb) WARNING: Unrecognized tag value: 29 ... ??? ... 0x0^M
FAIL: gdb.base/auxv.exp: info auxv on live process
...
Fix this by handling AT_HWCAP3 in default_print_auxv_entry. Likewise for
AT_HWCAP4.
Tested on aarch64-linux.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=32590
Since glibc 2.39, containing commit 3ab9b88 ("powerpc: Add HWCAP3/HWCAP4 data
to TCB for Power Architecture."), glibc's elf/elf.h contains constants
AT_HWCAP3/AT_HWCAP4.
Add these in elf/common.h.
Approved-By: Simon Marchi <simon.marchi@efficios.com>
This is a test-case to make sure that
8eb1701823
("gdb: fix loading compressed scripts from `.debug_gdb_scripts`-section")
doesn't regress.
The test was mostly implemented by Tom Tromey, with the exception of the
check to make sure that the sections are actually compressed and
corrected compiler flags (-Wl,--compress-debug-sections=zlib-gabi was
missing).
This is an individual patch since the actual fix already landed in the
repository.
Approved-By: Tom Tromey <tom@tromey.com>
Nothing seems to use this anymore. However, the private field is still
used internally by buildsym, so keep it.
Change-Id: Ie6fbd96110a3c5603359a483855bbecc4008e5b4
Approved-By: Tom Tromey <tom@tromey.com>
With the removal of the COFF debug info support, nothing uses
buildsym-legacy anymore, remove it.
Change-Id: I46612509889b6868f5889a4347ad8d0a28d29127
Approved-By: Tom Tromey <tom@tromey.com>
This patch removes support for reading in the COFF debug info.
I am not a specialist of the COFF format, but my understanding is that
debug info in the COFF format consisted of extra symbol table entries
describing the standard symbol table entries. For instance, when
compiling without debug info, a function would get a simple symbol
entry, akin to an ELF symbol. We would create a minimal symbol out of
that. When compiling with debug info, that symbol table entry would be
followed by special additional entries, describing things like
parameters and local variables (what would now be in DWARF). We would
create some full symbols out of that.
This patch removes everything that reads this extra information, on the
basis that any target still used today would not use it anymore, having
switched to DWARF instead. We still read the standard symbol table
entries and create minimal symbols out of that. This can be seen when
loading a Windows executable (the only COFF executable kind I know how
to produce):
$ ./gdb -nx -q --data-directory=data-directory -ex "set debug symtab-create 2" -ex "file a.exe"
...
[symtab-create] record_full: recording minsym: mst_file_bss 0x140008018 6 argv
[symtab-create] record_full: recording minsym: mst_file_bss 0x140008008 6 managedapp
[symtab-create] record_full: recording minsym: mst_text 0x1400013e0 0 WinMainCRTStartup
[symtab-create] record_full: recording minsym: mst_text 0x140001400 0 mainCRTStartup
[symtab-create] record_full: recording minsym: mst_text 0x140001420 0 atexit
[symtab-create] record_full: recording minsym: mst_file_data 0x140005000 3 __EH_FRAME_BEGIN__
[symtab-create] record_full: recording minsym: mst_file_bss 0x140008060 6 obj
...
[symtab-create] install: installing 299 minimal symbols of objfile /home/smarchi/src/binutils-gdb/gdb/a.exe
I did not and can't easily test more than that. At least, the remaining
code is understandable enough that I think we would be able to fix any
bug that comes up.
This change should not alter the debugging experience on Windows.
For more info on the COFF symbol table format:
https://www.delorie.com/djgpp/doc/coff/symtab.html
Change-Id: I83220589b8e5b242a4ac42a842e504a4aa47aada
Approved-By: Tom Tromey <tom@tromey.com>
Replace two uses of bfd_map_over_sections with an iteration over
gdb_bfd_sections.
Re-use cs_to_bfd_section in cs_section_address to simplify it.
Change-Id: I2b8f70cc1deba151b7b286affe78a43ac1a26375
Approved-By: Tom Tromey <tom@tromey.com>
This changes the tracepoint code to use make_unique_xstrdup (and
make_unique_xstrndup). This meant changing the types of some members
of uploaded_tp -- but it seems to me that using array types there did
not add any value.
Approved-By: Andrew Burgess <aburgess@redhat.com>
This replaces a number of uses of 'ptr.reset (xstrdup ())'
with 'ptr = make_unique_xstrdup ()'.
The main motivation for this is that, IMO, it's better to avoid the
reset method when possible.
Approved-By: Andrew Burgess <aburgess@redhat.com>
Continuation of fix to VMLA
Bit 12 of the first halfword in the VMLAS instruction is listed as (0)
in the ARMARM (document DDI0553B.w, version ID07072023).
This means that the instruction does not discriminate between signed
and unsigned types and processing elements do not use the bit.
The encoding used by gas was based on an older version
of the document that made the sign important.
This change makes it possible to use vmlas.i8 (16,32) in addition to
vmlas.u8 and vmlas.s8 mnemonics, with the i8, i16 and i32 aliases becoming
the default when disassembling.
The generated encoding sets bit 12 to 0, compatibly with other
assembler implementations.
This patch adds support for MLB invalidate (MLBI) instruction.
Syntax: MLBI <mlbi_op>{, <Xt>}
This instruction is an alias to "SYS #4, C7, C0, #<op2>{, <Xt>}"
and MLBI being the preferred disassembly.
The following list of MLBI operations are supported in this patch for the
MLBI instructions enabled by "+mpamv2"
* alle1
* vmalle1
* vpide1
* vpmge1
On s390 64-bit (s390x) compilers may save the SP, FP, and RA registers,
which are of interest in SFrame, in other registers, such as floating-
point registers, for instance when in a leaf function.
SFrame does not explicitly track the SP. Instead SFrame relies on the
architecture-specific CFA definition to recover the SP. The s390x ELF
ABI [1] defines the CFA as SP at call site + 160, which results in the
implicit SP recovery rule SP = CFA - 160.
Assuming that CFI on s390 64-bit (s390x) adheres to the CFA definition,
it is safe to ignore any CFI directives, that specify the SP register at
entry to be saved either on the stack or in another register, as the SP
can then always be recovered using the implicit SP recovery rule.
[1]: s390x ELF ABI, https://github.com/IBM/s390x-abi/releases
Committed-by: Indu Bhagat <indu.bhagat@oracle.com>
gas/
* gen-sframe.c (sframe_xlate_do_register): Ignore .cfi_register SP
on s390x.
gas/testsuite/
* gas/cfi-sframe/cfi-sframe.exp (cfi-sframe-s390x-err-4): Rename
test to cfi-sframe-s390x-sp-register.
* gas/cfi-sframe/cfi-sframe-s390x-err-4.d: Rename to ...
* gas/cfi-sframe/cfi-sframe-s390x-err-4.s: Likewise.
* gas/cfi-sframe/cfi-sframe-s390x-sp-register.d: This. Test
that .cfi_register SP is ignored.
* gas/cfi-sframe/cfi-sframe-s390x-sp-register.s: Likewise. Add
minimal assembler sample.
This patch adds two new tests for SFrame V3 changes, focusing on the
newly added flexible FDE TYPE SFRAME_FDE_TYPE_FLEX.
Following tests are added:
- be-flipping-v3.c: Validates that big-endian SFrame V3 data is
correctly endian flipped when run on little-endian hosts. It
verifies the decoding of CFA offsets and the new V3 register/offset
metadata bitfields using the SFRAME_V3_FLEX_FDE_OFFSET_REG_* macros.
- findfre-flex-1.c: Tests a variety of sframe_find_fre lookup
scenarios, apart from checking the basic encoder/decoder APIs.
Documentation for the binary test data DATA-BE-V3 is provided in
README-be-flipping-v3 to ensure reproducibility.
libsframe/
* Makefile.in: Regenerate.
* testsuite/libsframe.decode/DATA-BE-V3: New test data.
* testsuite/libsframe.decode/README-be-flipping-v3: New file.
* testsuite/libsframe.decode/be-flipping-v3.c: New test.
* testsuite/libsframe.decode/decode.exp: Run be-flipping-v3.
* testsuite/libsframe.decode/local.mk: Add be-flipping-v3.
* testsuite/libsframe.find/find.exp: Run findfre-flex-1.
* testsuite/libsframe.find/findfre-flex-1.c: New test.
* testsuite/libsframe.find/local.mk: Add findfre-flex-1.
This commit amalgamates a patch set proposed by Jens Remus to enable the
SFrame Version 3 Flexible FDE Type (SFRAME_FDE_TYPE_FLEX) generation for
the s390x ABI.
Previously, s390x relied on architecture-specific encoding (shifting register
numbers into offset fields) to represent register recovery rules. This limited
the complexity of CFI that could be supported. With Flex FDE enabled:
- s390x can now represent .cfi_def_cfa using non-SP/FP registers.
- The architecture-specific function s390_sframe_xlate_do_register () in GAS
is replaced by the generic Flex FDE generation path.
- The SFrame V3 specific macros for s390x register encoding are removed
from libsframe/include, as the generic Flex FDE format handles explicit
register columns natively.
The testsuite is updated to replace negative tests (which asserted
warnings or empty SFrame generation for these patterns) with positive
tests verifying valid Flex FDE generation.
Co-authored-by: Jens Remus <jremus@linux.ibm.com>
gas/
* config/tc-s390.c (s390_support_flex_fde_p): Return true to
enable Flex FDE generation.
* gen-sframe.c (s390_sframe_xlate_do_register): Disable s390x
specific implementation.
(sframe_xlate_do_register): Invoke generic Flex FDE path now
that flex FDE generation is supported.
gas/testsuite/
* gas/cfi-sframe/cfi-sframe-s390x-err-1.d: Removed.
* gas/cfi-sframe/cfi-sframe-s390x-err-1.s: Moved to...
* gas/cfi-sframe/cfi-sframe-s390x-non-spfp-cfa-1.s: ...here.
* gas/cfi-sframe/cfi-sframe-s390x-err-2.d: Removed.
* gas/cfi-sframe/cfi-sframe-s390x-err-2.s: Moved to...
* gas/cfi-sframe/cfi-sframe-s390x-non-spfp-cfa-2.s: ...here.
* gas/cfi-sframe/cfi-sframe-s390x-fpra-register-1.d: Update to
expect Flex FDE output.
* gas/cfi-sframe/cfi-sframe-s390x-fpra-register-2.d: Likewise.
* gas/cfi-sframe/cfi-sframe.exp: Run renamed tests.
include/
* sframe.h (SFRAME_V3_S390X_OFFSET_IS_REGNUM): Remove.
(SFRAME_V3_S390X_OFFSET_ENCODE_REGNUM): Remove.
(SFRAME_V3_S390X_OFFSET_DECODE_REGNUM): Remove.
libsframe/
* sframe-dump.c (sframe_s390x_offset_regnum_p): Return false
for SFrame V3.
(sframe_s390x_offset_decode_regnum): Remove V3 support.
Update the SFrame specification to document Version 3. This version
introduces fundamental changes to support additional scenarios (e.g.,
s390x register-based recovery, x86_64 DRAP) using a 'Flexible FDE'
definition while maintaining compactness for standard frames.
Key changes documented include:
- The SFrame Function Descriptor Entry (FDE) is split into two
distinct structures:
[sframe_func_desc_idx] Fixed-size index for binary search.
[sframe_func_desc_attr] Variable-location attributes including new
info bytes.
- Flexible FDEs (SFRAME_FDE_TYPE_FLEX)
A new FDE type that interprets FRE bytes not as simple stack
offsets, but as pairs of "Control Data" and "Offset". This
allows encoding complex recovery rules (e.g., "CFA = *(RBP - 8)")
without bloating the format for standard cases.
- Provision for defining new FDE types in future. A total of 5 bits
are reserved for this purpose.
- Make explicit distinction between FDE Types vs. PC Type
[FDE Type] Defines how to interpret stack trace data (Default vs. Flex).
[PC Type] Defines how PCs are advanced for an FDE (Increment vs. Mask).
- Other renames like sfde_func_start_address is renamed to
sfdi_func_start_offset to accurately reflect that it is a relative
offset, not an absolute address.
- Remove SFRAME_F_FRAME_POINTER from SFrame V3. The corresponding bit
is now unused in SFrame V3.
libsframe/doc/
* sframe-spec.texi: Update text for SFrame Version 3.
(Changes from Version 2 to Version 3): New section.
(The SFrame FDE Index): New section documenting sframe_func_desc_idx.
(The SFrame FDE Attribute): New section documenting
sframe_func_desc_attr.
(The SFrame FDE Info Bytes): Expanded to document sfda_func_info and
sfda_func_info2 split.
(The SFrame FDE PC Types): Document SFRAME_V3_FDE_PCTYPE_INC and
SFRAME_V3_FDE_PCTYPE_MASK.
(The SFrame FDE Types): Document SFRAME_FDE_TYPE_DEFAULT and
SFRAME_FDE_TYPE_FLEX.
(Interpretation of SFrame FREs): Split into Default and Flexible
interpretation.
(Flexible FDE Type Interpretation): Document the Control
Data/Offset pair encoding.
SFrame V3 has 8 precious flag bits, two of which are being used. More
flag byte (s) can be added to the auxiliary header when it comes to
that. But for now, it may be worthwhile to use the 8-bits frugally.
SFRAME_F_FRAME_POINTER flag bit was added with the intention of marking
binaries built with frame-pointer preserved. A stack tracer could then
use this information, to unambiguously ascertain whether frame-pointer
based stack traces will be precise. But such a marking of binary will
ideally be done by the linker, and at the moment adding such a framework
is not justified for such small gain. The outcome of this is that
SFRAME_F_FRAME_POINTER is never set in SFrame V2 binaries.
Remove the definition SFRAME_F_FRAME_POINTER for SFrame V3. The
relinquished bit can be used (in future format incarnations) when
reading/dumping SFrame V2 sections are no longer supported by consumers.
Changing the values of existing flags, e.g.,
SFRAME_F_FDE_FUNC_START_PCREL, is not being done for V3, to avoid
version-specific flag bit reading (albeit doable) in consumers.
Related changes to the specification are done in a subsequent commit.
include/
* sframe.h (SFRAME_V3_F_ALL_FLAGS): Remove
SFRAME_F_FRAME_POINTER from the set of V3 flags.
libsframe/
* sframe-dump.c (dump_sframe_header_flags): Add a comment for
clarity.
In this test, two "special" case FDEs are linked:
- Signal frame where the SFrame stack trace data is not
representable.
- _start like outermost frame function.
This is useful test for sframe_encoder's merge input sections
functionality and its associated write code-paths (sframe_encoder_write)
too.
ld/testsuite/
* ld-x86-64/sframe-link-1.d: New test.
* ld-x86-64/sframe-signal.s: New input file.
* ld-x86-64/sframe-start.s: Likewise.
* ld-x86-64/x86-64.exp: Add new test.
Add a new command line option --discard-sframe to the linker.
This option allows users to prevent the linker from generating an output
.sframe section.
The rationale for this option is: Consider the case when say, the distro
is shipped with SFrame sections in the installed binaries/libraries. A
user application using these installed libraries, but not enabling
.sframe for itself just yet, will see an output .sframe corresponding to
the pulled in libraries. This is "partial" .sframe information for the
application. Adding such an option to the linker, gives user a way to
turn off the .sframe section completely without relying on a linker
script.
Previously, the existing --no-ld-generated-unwind-info option
controlled whether (not just .eh_frame for PLT entries, but also) SFrame
for PLT entries. The new command line option,
--discard-sframe now decouples SFrame from other unwind
formats (like .eh_frame), allowing for more control over the output
binary's SFrame data.
The option is added for architectures that currently support SFrame:
AArch64, s390x, and x86_64.
bfd/
* elf-sframe.c (_bfd_elf_parse_sframe): Mark with SEC_EXCLUDE if
--discard-sframe is in effect.
* elf64-s390.c (elf_s390_create_dynamic_sections): Use
discard_sframe to guard .sframe section creation.
* elfxx-x86.c (_bfd_x86_elf_link_setup_gnu_properties): Likewise.
include/
* bfdlink.h (struct bfd_link_info): Add discard_sframe bitfield.
ld/
* ldlex.h (enum option_values): Add OPTION_NO_LD_SFRAME_INFO.
* lexsup.c (elf_sframe_list_options): New function.
(ld_list_options): Add sframe_info argument. Update callers.
* ld.texi: Update documentation.
* emulparams/sframe-info.sh: New file.
* emultempl/aarch64elf.em: Add --discard-sframe option
listing and handling.
* emulparams/elf64_s390.sh: Likewise.
* emulparams/elf_x86_64.sh: Likewise.
* Makefile.am: Update to handle sframe-info.sh and new list options.
* configure.ac: Handle SFRAME_INFO target variable.
* Makefile.in: Regenerate.
* configure: Regenerate.
ld/testsuite/
* ld-x86-64/x86-64.exp: New test.
* ld-x86-64/sframe-command-line-1.d: New test.
* ld-aarch64/aarch64-elf.exp: New test.
* ld-s390/s390.exp: New test.
* ld-x86-64/x86-64.exp: New test.
* ld-aarch64/sframe-command-line-1.d: New test.
* ld-s390/sframe-command-line-1.d: New test.
* ld-x86-64/sframe-command-line-1.d: New test.
This option will allow users to select emission of SFrame stack trace
information as per the SFrame version 3 specification. Currenly, SFrame
version 3 is also the default.
In future, as SFrame evolves, similar command line args may be added for
future versions.
gas/
* as.c (enum gen_sframe_version): New definition.
(parse_args): Add option processing for --gsframe-3.
* as.h (enum gen_sframe_version): New declaration.
* doc/as.texi: Document the new option.
* gen-sframe.c (sframe_set_version): Use enum gen_sframe_version
as version.
(output_sframe): Likewise.
gas/testsuite/
* gas/cfi-sframe/cfi-sframe.exp: New test.
* gas/cfi-sframe/cfi-sframe-common-1d.d: Test new command line
option --gsframe-3.
* gas/cfi-sframe/cfi-sframe-common-1d.s: Likewise.
This patch introduces a structural change to the SFrame V3 format. It
shifts the SFrame Function Descriptor Entry (FDE) (a physical entity in
SFrame V2) into a conceptual one in SFrame V3, such that an FDE is now
split into two distinct parts to optimize the binary search table and
data organization:
- FDE Index (sframe_func_desc_idx_v3): This structure contains the
essential indexing information: the function start address offset,
function size in bytes, and the offset to the SFrame FDE
attribute/Frame Row Entries (FREs) area for the function.
- FDE Attributes (sframe_func_desc_attr_v3): The metadata regarding the
function (number and size of FREs, FDE type, and repetition block
size etc.) is moved to a new structure.
On-Disk Layout: In V3, the "Attributes" are now stored immediately
preceding the SFrame FREs for that function. The sfde_func_start_fre_off
now points to the attr structure, and the actual FREs follow immediately
after. IOW, the "Attributes" are now moved to the FRE sub-section,
located immediately preceding the FREs for the respective function.
The above layout has the advantage that:
a) its cleaner with separation between the index elements vs other data
b) the index has better cache locality (by virtue of it being smaller
than the layout in SFrame V2).
c) As the format evolves, the guarantees of alignment for FDE index
are easier to maintain. FDE attr, being in the SFrame FRE
sub-section, carry no guarantees of alignment.
This had been previously suggested and communicated in an earlier
discussion on binutils mailling list
https://inbox.sourceware.org/binutils/29b1f7b0-61ea-410c-8aca-d5dd6115e668@oracle.com/
The read/write paths in sframe.c are updated to account for this split.
sframe_fde_tbl_init now requires access to the FRE buffer to populate
the internal FDE table, as the attributes are no longer resident in the
FDE section.
flip_sframe is refactored into version-specific handlers (_v2 and _v3)
because the endian-swapping logic now differs significantly. In V3, the
iterator must jump from the FDE table to the FRE section to swap the
attributes.
Lastly, the two entities generating SFrame sections (GAS and GNU ld)
both now must _not_ set the sfde_func_start_fre_off to zero, when the
number of FREs is zero. This is because now there will be some valid
attr data at that location.
Backward Compatibility: Due to the need to support readelf/objdump for
SFrame V2 sections, the patch explicitly maintains V2 support via
separate code paths (e.g., flip_sframe_fdes_with_fres_v2)
Note about alignment: Now that the sframe_func_desc_idx_v3 is refactored
out of the conceptual SFrame FDE, SFrame FDE index member elements are
at aligned boundaries again. The alignment property for SFrame FDE was
broken from an ealier patch "[08/36] sframe: gas: libsframe: use
uint16_t for num_fres of FDE" up until this one.
include/
* sframe.h (sframe_func_desc_entry_v3): Remove sfde_func_num_fres,
sfde_func_info, sfde_func_info2, and sfde_func_rep_size. Rename
to sframe_func_desc_idx_v3.
(sframe_func_desc_attr_v3): New SFrame FDE attribute structure.
libsframe/
* sframe.c (sframe_fde_tbl_init): Add argument for FRE buffer.
Read attributes from the FRE section for V3.
(flip_fde_desc): Rename from flip_fde. Check size against
sframe_func_desc_entry_v3.
(flip_fde_attr_v3): New function.
(sframe_decode_fde_desc_v2): New function extracted from
sframe_decode_fde.
(sframe_decode_fde_desc_v3): New function.
(sframe_decode_fde_attr_v3): New function.
(flip_sframe_fdes_with_fres_v2): New function for V2 flipping.
(flip_sframe_fdes_with_fres_v3): New function for V3 flipping.
(flip_sframe): Dispatch to version-specific flip functions.
(sframe_decode): Pass FRE buffer to sframe_fde_tbl_init.
(sframe_decoder_get_offsetof_fde_start_addr): Adjust for subset
of sframe_func_desc_entry_v3 restructured into
sframe_func_desc_idx_v3.
(sframe_encoder_get_offsetof_fde_start_addr): Likewise.
(sframe_find_fre): Skip attribute size to find FREs in V3.
(sframe_decoder_get_fre): Likewise.
(sframe_decoder_get_fres_buf): Likewise.
(sframe_encoder_add_fre): Add attribute size to byte count.
(sframe_encoder_add_fres_buf): Read attributes from buffer.
(sframe_encoder_write_fde): Write only FDE index fields.
(sframe_encoder_write_func_attr): New function.
(sframe_encoder_write_sframe): Write FDE attributes before FREs.
gas/
* gen-sframe.c (output_sframe_funcdesc): Do not reset
sfde_func_start_fre_off to zero when zero num FREs.
(output_sframe_func_desc_attr): New refactored out function.
(output_sframe_internal): Invoke output_sframe_func_desc_attr.
libsframe/testsuite/
* libsframe.decode/DATA2: Regenerate data file.
At link time, in _bfd_elf_merge_section_sframe (), it suffices to bring
over the all per-function stack trace metadata (all FREs) as a blob into
the SFrame encoder object. There is no need to "decode" each SFrame
FRE, only to add them in a serial fashion.
This is an optimization, and not directly related to any SFrame V3
related changes to the specification. This should also bring us a step
closer to supporting SFrame for targets which use linker relaxations.
Removing the need to decode the input FREs can allow the linker to
simply use the available set of FREs from (relocated) contents. To
support targets using linker relaxations in SFrame, other changes in the
SFrame parse functionality time may also be necessary, but this brings
us just a step closer.
Add two new APIs to accomplish this:
- sframe_decoder_get_fres_buf, and
- sframe_encoder_add_fres_buf
bfd/
* elf-sframe.c (_bfd_elf_merge_section_sframe): Get all FRE data
and add it all in bulk.
libsframe/
* libsframe.ver: Add new APIs.
* sframe.c (sframe_buf_fre_entry_size): New internal API to get
size of one SFrame FRE at the indicated buffer location, without
fully "decoding" the SFrame FRE.
(sframe_decoder_get_fres_buf): New definition.
(sframe_encoder_add_fres_buf): New definition.
include/
* sframe-api.h (sframe_decoder_get_fres_buf): New declaration.
(sframe_encoder_add_fres_buf): New declaration.
With the introduction of flex FDE type, handling .cfi_offset for
RA needs adjustment.
On architectures like x86_64, the return address (RA) is typically saved
at a fixed offset from the CFA. Previous versions of the SFrame format
assumed this fixed offset was invariant for the entire function on such
architectures. Consequently, GAS would warn and suppress SFrame
generation if it encountered a .cfi_offset directive for the RA
register that deviated from this fixed default.
SFrame V3 introduces "Flex FDEs" which allows tracking the RA location
explicitly even on architectures where it is usually fixed.
This patch updates sframe_xlate_do_offset () to leverage Flex FDEs. When
processing a .cfi_offset for the RA register:
- The check for non-representable RA offsets is relaxed. If the ABI
supports Flex FDEs (SFrame V3), GAS proceeds instead of issuing a
warning.
- For ABIs without explicit RA tracking (like AMD64), if the RA
offset differs from the default fixed offset, the FDE is marked as a
Flex FDE, and the new stack location is recorded.
- Logic is added to detect when the RA is restored to its standard
fixed offset. In this case, the tracking state is reset (ra_loc set
to SFRAME_FRE_ELEM_LOC_NONE), deferring to the ABI's default fixed RA
offset behavior.
gas/
* gen-sframe.c (sframe_xlate_do_offset): Support .cfi_offset for RA
by switching to Flex FDEs when necessary.
gas/testsuite/
* gas/cfi-sframe/cfi-sframe-x86_64-6.d: New test.
* gas/cfi-sframe/cfi-sframe-x86_64-6.s: New test to check
transition of location of REG_RA from register to CFA-8 (default
location on AMD64). Flex FDE in effect.
* gas/cfi-sframe/cfi-sframe.exp: Add new test.
SFrame does not track the SP. For recovery of the SP, SFrame relies on
the architecture/ABI's CFA definition:
CFA = SP [+ offset // on s390x]
Which results in the following implicit CFA value offset rule for SP:
SP = CFA [- offset // on s390x]
Where offset is zero for most architectures/ABIs, except s390x.
Therefore .cfi_register SP, reg directives must be rejected, as such
semantics cannot be represented in SFrame yet.
gas/
* gas/gen-sframe.c (s390_sframe_xlate_do_register): Check for
REG_SP and reject while generating a warning.
(sframe_xlate_do_register): Likewise.
gas/testsuite/
* gas/cfi-sframe/cfi-sframe-s390x-err-4.d: New test.
* gas/cfi-sframe/cfi-sframe-s390x-err-4.s: Likewise.
* gas/cfi-sframe/cfi-sframe-x86_64-empty-5.d: Likewise.
* gas/cfi-sframe/cfi-sframe-x86_64-empty-5.s: Likewise.
* gas/cfi-sframe/cfi-sframe.exp: Add new tests.
Use SFrame FDE of type SFRAME_FDE_TYPE_FLEX_TOPMOST_FRAME.
When FP, RA were moved to a general-purpose register, the SFrame
generation previously warned and skipped the FDE (except on s390x).
This patch updates the translator to detect .cfi_register for RA (and
FP), tracks the destination register in the SFrame row entry,
and emits the register in the relevant FRE offsets in SFrame FDE type
SFRAME_FDE_TYPE_FLEX.
gas/
* gen-sframe.c (sframe_row_entry_initialize): Propagate ra_reg
and ra_deref_p.
(sframe_xlate_do_register): Handle .cfi_register for RA/FP on
AMD64 by setting flex_p and recording the register.
gas/testsuite/gas/
* cfi-sframe/cfi-sframe.exp: Run new test.
* cfi-sframe/cfi-sframe-x86_64-5.d: New test.
* cfi-sframe/cfi-sframe-x86_64-5.s: Simple test for checking
FLEX FDE generation for `.cfi_register REG_RA, XX`.
* cfi-sframe/cfi-sframe-x86_64-esc-expr-3.d: New test.
* cfi-sframe/cfi-sframe-x86_64-esc-expr-3.s: New test with DWARF
expression for REG_FP, followed by .cfi_register and .cfi_offset
for REG_FP.
* cfi-sframe/cfi-sframe-x86_64-ra-undefined-flex-1.d: New test.
* cfi-sframe/cfi-sframe-x86_64-ra-undefined-flex-1.s: New test
for FLEX FDE with undefined RA.
gas/testsuite/
* gas/cfi-sframe/cfi-sframe-x86_64-4.d: New test.
* gas/cfi-sframe/cfi-sframe-x86_64-4.s: Non SP/FP based CFA.
* gas/cfi-sframe/cfi-sframe-x86_64-esc-expr-1.d: New test.
* gas/cfi-sframe/cfi-sframe-x86_64-esc-expr-1.s: DRAP pattern
with both CFA expression and FP expression.
* gas/cfi-sframe/cfi-sframe-x86_64-esc-expr-2.d: New test.
* gas/cfi-sframe/cfi-sframe-x86_64-esc-expr-2.s: Test CFA
expression with sleb128.
* gas/cfi-sframe/cfi-sframe.exp: Add new tests.
This patch updates the SFrame generation in GAS to translate specific
.cfi_escape directives into the new SFrame V3 FDE type
SFRAME_FDE_TYPE_FLEX.
The primary goal is to support code patterns where:
- the Canonical Frame Address (CFA) is not defined by a simple offset
from the Stack Pointer (SP) or Frame Pointer (FP), or where the CFA
rule involves a dereference. Such patterns are generated by compilers
for stack realignment (e.g., DRAP on x86_64, or when mixing legacy
codes that keep 4-byte stack alignment with modern codes that keep
16-byte stack alignment for SSE compatibility).
- the Frame Pointer is not defined by a simple offset from the CFA,
but may even involve another register and/or dereferencing.
- the Return Address is not defined by a simple offset from the CFA,
but may even involve another register and/or dereferencing.
Support for non-SP/FP based CFA: Update sframe_xlate_do_def_cfa () and
sframe_xlate_do_def_cfa_register () to detect when a non-SP/FP register
is used for the CFA.
Support for CFA expressions: A vital part of supporting the
above-mentioned cases on AMD64 is support for CFA expressions. Add
sframe_xlate_do_escape_cfa_expr () to parse simple
DW_CFA_def_cfa_expression sequence in .cfi_escape.
Support for FP expressions: Update sframe_xlate_do_escape_expr () to
handle DW_CFA_expression involving DW_OP_breg6 (rbp) on AMD64,
allowing for tracking of the Frame Pointer when it is saved with a
dereference rule in the DRAP pattern.
The "support" for both CFA expressions and FP expressions is quite
minimal, and is tailored to the most commonly seen occurrences generated
by GCC for AMD64.
gas/
* gas/gen-sframe.c (output_sframe_row_entry_offsets):
(sframe_xlate_do_def_cfa): Handle non-SP/FP CFA registers by setting
flex_p for AMD64.
(sframe_xlate_do_def_cfa_register): Likewise.
(sframe_xlate_escape_sleb128_to_int64): New definition.
(sframe_xlate_do_escape_cfa_expr): New function to handle
DW_CFA_def_cfa_expression of specific shapes.
(sframe_xlate_do_escape_expr): Update to handle dereferenced FP rules.
(sframe_xlate_do_cfi_escape): Invoke sframe_xlate_do_escape_cfa_expr ().
(create_sframe_all): In case of error, but when signal frame is
also true, there cannot be a flex FDE.
