| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Fix these little typos from commit 5fe70629ceaf ("Change file
initialization to use INIT_GDB_FILE macro").
Change-Id: Ib9ae29988dfda1165de47467087f154624916629
|
|
|
|
Commit:
commit b23903836007d1acaf7f8c059ab000ee83fcebfa
Date: Tue Mar 21 13:01:26 2023 +0100
gdb: linux-namespaces: enter user namespace when appropriate
added a new test gdb.base/user-namespace-attach.exp. It has been
reported that this test will sometimes fail, like this:
(gdb) attach 184732
Attaching to process 184732
warning: process 184732 is a zombie - the process has already terminated
ptrace: Operation not permitted.
(gdb) FAIL: gdb.base/user-namespace-attach.exp: flags=--mount --map-root-user: attach to inferior
the test tries to run the 'unshare' application. Sometimes though,
the application is present, but the set of flags used is not
supported (maybe due to restrictions on the local machine), so we see
behaviour like this:
$ unshare --mount --map-root-user /bin/true; echo $?
unshare: unshare failed: Operation not permitted
1
Handle this case by first running 'unshare' with the same flags, but
using '/bin/true', if this fails then assume the flags are not
supported, and skip the test.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=33108
|
|
This patch introduces a new macro, INIT_GDB_FILE. This is used to
replace the current "_initialize_" idiom when introducing a per-file
initialization function. That is, rather than write:
void _initialize_something ();
void
_initialize_something ()
{
...
}
... now you would write:
INIT_GDB_FILE (something)
{
...
}
The macro handles both the declaration and definition of the function.
The point of this approach is that it makes it harder to accidentally
cause an initializer to be omitted; see commit 2711e475 ("Ensure
cooked_index_entry self-tests are run"). Specifically, the regexp now
used by make-init-c seems harder to trick.
New in v2: un-did some erroneous changes made by the script.
The bulk of this patch was written by script.
Regression tested on x86-64 Fedora 41.
|
|
|
|
This patch add support for FEAT_PoPS feature which can be enabled
through +pops command line flag.
This patch also adds support for following DC instructions and the
spec can be found here [1].
1. "dc cigdvaps" enabled on passing +memtag+pops command line flags.
2. "dc civaps" enabled on passing +pops command line flag.
[1]: https://developer.arm.com/documentation/ddi0601/2025-03/AArch64-Instructions?lang=en
|
|
After the last commit there were still a couple of calls to
m_stream->emit_style_escape in the pager_file class. As discussed in
the last commit, these are likely wrong, but I'd not been able to
produce any bugs because of them.
The reason why there are no bugs is that these calls are, I think,
entirely redundant. Consider this block:
if (m_wrap_column)
{
/* We are about to insert a newline at an historic
location in the WRAP_BUFFER. Before we do we want to
restore the default style. To know if we actually
need to insert an escape sequence we must restore the
current applied style to how it was at the WRAP_COLUMN
location. */
m_applied_style = m_wrap_style;
m_stream->emit_style_escape (ui_file_style ());
/* If we aren't actually wrapping, don't output
newline -- if chars_per_line is right, we
probably just overflowed anyway; if it's wrong,
let us keep going. */
m_stream->puts ("\n");
}
What we know (see previous commit) is that the call:
m_stream->emit_style_escape (ui_file_style ());
is dangerous as m_stream->m_applied_style is going to be out of sync
with its current state. Actually, m_stream->m_applied_style is likely
to be the default style as it is not updated elsewhere. So why does
this not cause problems?
Well, GDB's style output is always done in tightly scoped regions.
That means if we want to print some styled output, and then apply a
wrap point the code might look like this:
fprintf_styled (gdb_stdout, file_name_style, "some text");
gdb_stdout->wrap_here (4);
But, after printing 'some text', the style of gdb_stdout will have
returned to the default style.
My claim is that, whenever we encounter a wrap_here call, the stream
in question will _always_ have been returned to the default style.
This means that, in the block above, the call:
m_stream->emit_style_escape (ui_file_style ());
will never emit anything because it depends on a check against
m_stream->m_applied_style, which will always mean that the above call
does nothing. But that's OK. By chance, we'll have always placed the
stream into a default style state anyway, so no harm done.
Similarly, the other call:
/* Having finished inserting the wrapping we should
restore the style as it was at the WRAP_COLUMN. */
m_stream->emit_style_escape (m_wrap_style);
Tries to return m_stream to the state it was in at the point of the
wrap_here call. But, as described above, this will always be the
default style, so the above call will do nothing, but that just
happens to be exactly what we want!
So what does this commit do?
Well, I "fix" the above code by removing the
m_stream->emit_style_escape calls and replacing them with calls to
puts, passing in the escape sequence for the required style, but only
if the m_stream style as tracked by pager_file::m_stream_style
indicates this is needed.
Got the reasons given above, this should mean there is no change after
this patch. We still shouldn't be emitting any extra escape
sequences. But, should we ever manage to get into a state where we
call wrap_here with a stream in a style other than the default, then
this should mean things work as expected.
There should be no user visible changes after this commit.
Approved-By: Tom Tromey <tom@tromey.com>
|
|
This commit fixes a couple of issues relating to the pagination
prompt and styling. The pagination prompt is this one:
--Type <RET> for more, q to quit, c to continue without paging--
I did try to split this into multiple patches, based on the three
issues I describe below, but in the end, the fixes were all too
interconnected, so it ended up as one patch that makes two related,
but slightly different changes:
1. Within the pager_file class, relying on the m_applied_style
attribute of the wrapped m_stream, as is done when calling
m_stream->emit_style_escape, is not correct, so stop doing that, and
2. Failing to update m_applied_style within the pager_file class can
leave that attribute out of date, which can then lead to styling
errors later on, so ensure m_applied_style is always updated.
The problems I have seen are:
1. After quitting from a pagination prompt, the next command can
incorrectly style its output. This was reported as bug PR
gdb/31033, and is fixed by this commit.
2. The pagination prompt itself could be styled. The pagination
prompt should always be shown in the default style.
3. After continuing the output at a pagination prompt, GDB can fail
to restore the default style the next time the output (within the
same command) switches back to the default style.
There are tests for all these issues as part of this patch.
The pager_file class is a sub-class of wrapped_file, this means that a
pager_file is itself a ui_file, while it also manages a pointer to a
ui_file object (called m_stream). An instance of pager_file can be
installed as the gdb_stdout ui_file object.
Output sent to a pager_file is stored within an internal
buffer (called m_wrap_buffer) until we have a complete line, when the
content is flushed to the wrapped m_stream. If sufficient lines have
been written out then the pager_file will present the pagination
prompt and allow the user to continue viewing output, or quit the
current command.
As a pager_file is a ui_file, it has an m_applied_style member
variable.
The managed stream (m_stream) is also a ui_file, and so also has an
m_applied_style member variable.
In some places within the pager_file class we attempt to change the
current style of the m_stream using calls like this:
m_stream->emit_style_escape (style);
See pager_file::emit_style_escape, pager_file::prompt_for_continue,
and pager_file::puts. These calls will end up in
ui_file::emit_style_escape, which tries to skip emitting unnecessary
style escapes by checking if the requested style matches the current
m_applied_style value.
The m_applied_style value is updated by calls to the emit_style_escape
function.
The problem here is that most of the time pager_file doesn't change
the style of m_stream by calling m_stream->emit_style_escape. Most of
the time, style changes are performed by pager_file writing the escape
sequence into m_wrap_buffer, and then later flushing this buffer to
m_stream by calling m_stream->puts.
It has to be done this way. Calling m_stream->emit_style_escape
would, if it actually changed the style, immediately change the style
by emitting an escape sequence. But pager_file doesn't want that, it
wants the style change to happen later, when m_wrap_buffer is
flushed.
To avoid excessive style escape sequences being written into
m_wrap_buffer, the pager_file::m_applied_style performs a function
similar to the m_applied_style within m_stream, it tracks the current
style for the end of m_wrap_buffer, and only allows style escape
sequences to be emitted if the style is actually changing.
However, a consequence of this is the m_applied_style within m_stream,
is not updated, which means it will be out of sync with the actual
current style of m_stream. If we then try to make a call to
m_stream->emit_style_escape, if the style we are changing too happens
to match the out of date style in m_stream->m_applied_style, then the
style change will be ignored.
And this is indeed what we see in pager_file::prompt_for_continue with
the call:
m_stream->emit_style_escape (ui_file_style ());
As m_stream->m_applied_style is not being updated, it will always be
the default style, however m_stream itself might not actually be in
the default style. This call then will not emit an escape sequence as
the desired style matches the out of date m_applied_style.
The fix in this case is to call m_stream->puts directly, passing in
the escape sequence for the desired style. This will result in an
immediate change of style for m_stream, which fixes some of the
problems described above.
In fact, given that m_stream's m_applied_style is always going to be
out of sync, I think we should change all of the
m_stream->emit_style_escape calls to instead call m_stream->puts.
However, just changing to use puts doesn't fix all the problems.
I found that, if I run 'apropos time', then quit at the first
pagination prompt. If for the next command I run 'maintenance time' I
see the expected output:
"maintenance time" takes a numeric argument.
However, everything after the first double quote is given the command
name style rather than only styling the text between the double
quotes.
Here is GDB's stack while printing the above output:
#2 0x0000000001050d56 in ui_out::vmessage (this=0x7fff1238a150, in_style=..., format=0x1c05af0 "", args=0x7fff1238a288) at ../../src/gdb/ui-out.c:754
#3 0x000000000104db88 in ui_file::vprintf (this=0x3f9edb0, format=0x1c05ad0 "\"%ps\" takes a numeric argument.\n", args=0x7fff1238a288) at ../../src/gdb/ui-file.c:73
#4 0x00000000010bc754 in gdb_vprintf (stream=0x3f9edb0, format=0x1c05ad0 "\"%ps\" takes a numeric argument.\n", args=0x7fff1238a288) at ../../src/gdb/utils.c:1905
#5 0x00000000010bca20 in gdb_printf (format=0x1c05ad0 "\"%ps\" takes a numeric argument.\n") at ../../src/gdb/utils.c:1945
#6 0x0000000000b6b29e in maintenance_time_display (args=0x0, from_tty=1) at ../../src/gdb/maint.c:128
The interesting frames here are #3, in here `this` is the pager_file
for GDB's stdout, and this passes its m_applied_style to frame #2 as
the `in_style` argument.
If the m_applied_style is wrong, then frame #2 will believe that the
wrong style is currently in use as the default style, and so, after
printing 'maintenance time' GDB will switch back to the wrong style.
So the question is, why is pager_file::m_applied_style wrong?
In pager_file::prompt_for_continue, there is an attempt to switch back
to the default style using:
m_stream->emit_style_escape (ui_file_style ());
If this is changed to a puts call (see above) then this still leaves
pager_file::m_applied_style out of date.
The right fix in this case is, I think, to instead do this:
this->emit_style_escape (ui_file_style ());
this will update pager_file::m_applied_style, and also send the
default style to m_stream using a puts call.
While writing the tests I noticed that I was getting unnecessary style
reset sequences emitted.
The problem is that, around pagination, we don't really know what
style is currently applied to m_stream. The
pager_file::m_applied_style tracks the style at the end of
m_wrap_buffer, but this can run ahead of the current m_stream style.
For example, if the screen is currently full, such that the next
character of output will trigger the pagination prompt, if the next
call is actually to pager_file::emit_style_escape, then
pager_file::m_applied_style will be updated, but the style of m_stream
will remain unchanged. When the next character is written to
pager_file::puts then the pagination prompt will be presented, and GDB
will try to switch m_stream back to the default style. Whether an
escape is emitted or not will depend on the m_applied_style value,
which we know is different than the actual style of m_stream.
It is, after all, only when m_wrap_buffer is flushed to m_stream that
the style of m_stream actually change.
And so, this commit also adds pager_file::m_stream_style. This new
variable tracks the current style of m_stream. This really is a
replacement for m_stream's ui_file::m_applied_style, which is not
accessible from pager_file.
When content is flushed from m_wrap_buffer to m_stream then the
current value of pager_file::m_applied_style becomes the current style
of m_stream. But, when m_wrap_buffer is filling up, but before it is
flushed, then pager_file::m_applied_style can change, but
m_stream_style will remain unchanged.
Now in pager_file::emit_style_escape we are able to skip some of the
direct calls to m_stream->puts() used to emit style escapes.
After all this there are still a few calls to
m_stream->emit_style_escape(). These are all in the wrap_here support
code. I think that these calls are technically broken, but don't
actually cause any issues due to the way styling works in GDB. I
certainly haven't been able to trigger any bugs from these calls yet.
I plan to "fix" these in the next commit just for completeness.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31033
Approved-By: Tom Tromey <tom@tromey.com>
|
|
On openSUSE Tumbleweed (with python 3.13), I get:
...
(gdb) PASS: gdb.python/py-warning.exp: python gdb.warning("")
python gdb.warning()^M
Python Exception <class 'TypeError'>: \
function missing required argument 'text' (pos 1)^M
Error occurred in Python: function missing required argument 'text' (pos 1)^M
(gdb) PASS: gdb.python/py-warning.exp: python gdb.warning()
...
But on openSUSE Leap 15.6 (with python 3.6), I get instead:
...
(gdb) PASS: gdb.python/py-warning.exp: python gdb.warning("")
python gdb.warning()^M
Python Exception <class 'TypeError'>: \
Required argument 'text' (pos 1) not found^M
Error occurred in Python: Required argument 'text' (pos 1) not found^M
(gdb) FAIL: gdb.python/py-warning.exp: python gdb.warning()
...
Fix this by updating the regexp.
Tested on x86_64-linux.
PR testsuite/33104
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=33104
|
|
Disclaimer: this issue cannot occur with Object Attributes v1 (OAv1) while
using the GNU binutils because a value of '\0' (empty string) for a tag
with a string value is considered as the default value for the attribute,
and consequently is eliminated by gas from the output object file during
the serialization.
An empty string is a valid value for a NTBS tag in both OAv1 and OAv2 [1]
cases. However, contrarily to OAv1, a OAv2 subsection can be required and
so, tags in this subsection might have to be present even if the value is
the default. To comply with this requirement, the OAv2 serializer won't
drop the default values.
In the case where a NTBS tag has the value '\0' and is last in the object
attributes section, the current code in readelf used for dumping the object
attributes incorrectly detects an overflow, and prints out an error message
for a corrupted string tag.
This patch fixes the detection of the overflow so that it now accept an
empty string in the last tag of the object attributes section.
It also fixes the previous tests for the empty NTBS case and the non-null
terminated string one. The fix was also tested in the context of OAv2's
patch series [1] where the issue was originally detected. No regression
was found.
[1]: https://inbox.sourceware.org/binutils/20250509151319.88725-1-matthieu
.longo@arm.com/
|
|
The current testsuite for gas/readelf lacked two tests for EABI build
attributes:
- one when the final attribute is an empty string.
- one when the final attribute is a string missing the NULL terminator.
Those two issues cannot occur with Object Attributes v1 (OAv1) sections
created by the GNU binutils. Indeed a value of '\0' (empty string) for a
tag with a string value is considered as the default value for the
attribute, and consequently is eliminated by Gas from the output object
file during the serialization.
However, readelf should be able to process correctly files of an unknown
origin that could contain those two use cases.
This patch adds the two tests mentioned above. The first one is marked
as XFAIL because the empty string is not processed correctly by readelf
when it is in the last position. The second one passes, but simply print
out "[...]" without mentioning that the NTBS is corrupted.
A following patch will fix the bug in readelf, and will amend the newly
introduced tests.
|
|
On x86_64-freebsd with test-case gdb.base/infcall-failure.exp I get:
...
(gdb) continue
Continuing.
Program received signal SIGSEGV, Segmentation fault.
Address not mapped to object.
0x0000000000400522 in func_segfault () at infcall-failure.c:24
24 return *p; /* Segfault here. */
Error in testing condition for breakpoint 2:
The program being debugged was signaled while in a function called from GDB.
GDB remains in the frame where the signal was received.
To change this behavior use "set unwind-on-signal on".
Evaluation of the expression containing the function
(func_segfault) will be abandoned.
When the function is done executing, GDB will silently stop.
(gdb) FAIL: $exp: target_async=on: target_non_stop=on: \
run_cond_hits_segfault_test: continue
...
The problem is that the regexp in the test-case doesn't expect the
"Address not mapped to object." bit.
Fix this by updating the regexp.
Approved-by: Kevin Buettner <kevinb@redhat.com>
Tested on x86_64-freebsd and x86_64-linux.
|
|
|
|
This patch adds support for the RISC-V Profiles RVA23S64 and RVB23S64.
Version log:
Fix wrong test for rvb23s.
bfd/ChangeLog:
* elfxx-riscv.c: New Profiles.
gas/ChangeLog:
* testsuite/gas/riscv/attribute-rva23s.d: New test.
* testsuite/gas/riscv/attribute-rvb23s.d: New test.
|
|
PR testsuite/31831 reports the following failure in the
gdb.dap/log-message.exp test-case (formatted for readability):
...
{ "type": "event",
"event": "output",
"body": {
"category": "stdout",
"output": "Breakpoint 1 at 0x681: file log-message.c, line 23.\n"
},
"seq": 13
}
FAIL: $exp: logging output (checking body category)
...
for a gdb 14.2 based package.
The output event listed above is a result from the setBreakpoints request.
The test-case issues the setBreakpoints request and waits for the
corresponding response, but doesn't wait for the output event, and
consequently the output event is read by:
...
dap_wait_for_event_and_check "logging output" output \
{body category} console \
{body output} "got 23 - 23 = 0"
...
which triggers the failure.
I'm not able to reproduce this, but it looks worth fixing regardless.
We're fixing this on trunk though, and the output event looks different, and
there's one more output event:
...
{ "type": "event",
"event": "output",
"body": {
"category": "stdout",
"output": "No source file named log-message.c.\n"
},
"seq": 4
}
{ "type": "event",
"event": "output",
"body": {
"category": "stdout",
"output": "Breakpoint 1 (-source log-message.c -line 23) pending.\n"
},
"seq": 5
}
...
Fix this by waiting for these two output events, making the test-case a bit
more robust.
It is possible that one or both of these output events will be read by
dap_check_request_and_response "set breakpoint", and in that case restashing
them (for which there's currently no infrastructure) would be an easy way of
handling this. But I haven't been able to trigger that, so I'm leaving that
for if and when it does.
Tested on x86_64-linux.
Approved-By: Tom Tromey <tom@tromey.com>
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=31831
|
|
A user pointed out that DAP allows the "threads" request to work when
the inferior is running. This is documented in the overview, not the
specification.
While looking into this, I found a few other issues:
* The _thread_name function was not marked @in_gdb_thread.
This isn't very important but is still an oversight.
* DAP requires all threads to have a name -- the field is not optional
in the "Thread" type.
* There was no test examining events resulting from the inferior
printing to stdout.
This patch fixes all these problems.
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=33080
|
|
I changed my system linker to 'mold', but then I saw some gdb test
failures. This patch fixes a subset of the failures.
dw2-strp.exp was failing, and investigating showed that there were two
.debug_str sections. I tracked this down to the .S file not using the
correct section flags.
This patch fixes this problem, plus the other instances I could find.
(Strangely, these did not all cause problems, however.) I also
changed the DWARF assembler to always use these flags for .debug_str.
|
|
Split -v from -version/--version. They aren't the same; -v long form is
--verbose, which so far wasn't mentioned at all.
|
|
Update the Profiles string in RV23 to include the extensions 'b' and 'supm'.
bfd/ChangeLog:
* elfxx-riscv.c: Update Profiles string in RV23.
gas/ChangeLog:
* testsuite/gas/riscv/attribute-19.d: Update test string.
* testsuite/gas/riscv/attribute-20.d: Ditto.
|
|
|
|
|
|
|
|
This patch adds support for generating unlabeled landing pad PLT entries
for the RISC-V architecture. Unlabeled landing pad will place a LPAD
instruction at the PLT entry and PLT header, also PLT header will have
few changes due to the offset is different from the original one.
Ref: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/417
|
|
GNU_PROPERTY_RISCV_FEATURE_1_AND will perform a bitwise AND operation
on the properties of the input files.
|
|
GNU_PROPERTY_RISCV_FEATURE_1_CFI_LP_UNLABELED
This patch adds two new GNU properties for RISC-V:
GNU_PROPERTY_RISCV_FEATURE_1_CFI_SS and GNU_PROPERTY_RISCV_FEATURE_1_CFI_LP_UNLABELED.
We only add readelf and define the properties in this patch.
Ref: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/417
|
|
The goal of this refactor is to improve the possiblity of having
different PLT generation code for different RISC-V ABIs. The changes
include:
- Extract PLT generation logic into individual functions.
- Keep the PLT generation data in riscv_elf_link_hash_table.
In the following patches, we will use this framework to implement
different PLT.
|
|
|
|
Make sure we bail out early from amd64_analyze_prologue if CURRENT_PC
is reached to avoid unnecessary call to amd64_analyze_frame_setup.
Approved-By: Andrew Burgess <aburgess@redhat.com>
|
|
Compilers can put a sequence aligning the stack at the entry of a
function. However with -fcf-protection enabled, "endbr64" is
generated before. Current implementation of amd64 prologue analyzer
first checks for stack alignment and then for "endbr64", which is not
correct. This behavior was introduced with patch "gdb: handle endbr64
instruction in amd64_analyze_prologue". In case both are generated,
prologue will not be skipped. This patch swaps the order so that
"endbr64" is checked first and adds a regression test. i386-tdep
implementation also already had those checked in the correct order,
that is stack alignment is after endbr64.
Given such source compiled with gcc 11.4.0 via:
gcc -O0 main.c -o main
```
#include <alloca.h>
void
foo (int id)
{
volatile __attribute__ ((__aligned__ (64))) int a;
volatile char *p = (char *) alloca (id * 12);
p[2] = 'b';
}
int
main (int argc, char **argv)
{
foo (argc + 1);
return 1;
}
```
we get such function entry for foo (generated with objdump -d):
```
0000000000001149 <foo>:
1149: f3 0f 1e fa endbr64
114d: 4c 8d 54 24 08 lea 0x8(%rsp),%r10
1152: 48 83 e4 c0 and $0xffffffffffffffc0,%rsp
1156: 41 ff 72 f8 push -0x8(%r10)
115a: 55 push %rbp
115b: 48 89 e5 mov %rsp,%rbp
115e: 41 52 push %r10
1160: 48 81 ec a8 00 00 00 sub $0xa8,%rsp
1167: 89 7d 8c mov %edi,-0x74(%rbp)
...
```
The 3 instructions following endbr64 align the stack. If we were to set
a breakpoint on foo, gdb would set it at function's entry:
```
(gdb) b foo
Breakpoint 1 at 0x1149
(gdb) r
...
Breakpoint 1, 0x0000555555555149 in foo ()
(gdb) disassemble
Dump of assembler code for function foo:
=> 0x0000555555555149 <+0>: endbr64
0x000055555555514d <+4>: lea 0x8(%rsp),%r10
0x0000555555555152 <+9>: and $0xffffffffffffffc0,%rsp
0x0000555555555156 <+13>: push -0x8(%r10)
0x000055555555515a <+17>: push %rbp
0x000055555555515b <+18>: mov %rsp,%rbp
0x000055555555515e <+21>: push %r10
0x0000555555555160 <+23>: sub $0xa8,%rsp
0x0000555555555167 <+30>: mov %edi,-0x74(%rbp)
...
```
With this patch fixing the order of checked instructions, gdb can
properly analyze the prologue:
```
(gdb) b foo
Breakpoint 1 at 0x115e
(gdb) r
...
Breakpoint 1, 0x000055555555515e in foo ()
(gdb) disassemble
Dump of assembler code for function foo:
0x0000555555555149 <+0>: endbr64
0x000055555555514d <+4>: lea 0x8(%rsp),%r10
0x0000555555555152 <+9>: and $0xffffffffffffffc0,%rsp
0x0000555555555156 <+13>: push -0x8(%r10)
0x000055555555515a <+17>: push %rbp
0x000055555555515b <+18>: mov %rsp,%rbp
=> 0x000055555555515e <+21>: push %r10
0x0000555555555160 <+23>: sub $0xa8,%rsp
0x0000555555555167 <+30>: mov %edi,-0x74(%rbp)
...
```
Approved-By: Andrew Burgess <aburgess@redhat.com>
|
|
Refactor amd64_analyze_prologue so it clearly reflects what is the order
of operations in the prologue that we expect to encounter, as is the
case for i386's implementation.
Approved-By: Andrew Burgess <aburgess@redhat.com>
|
|
Update the make-check-all.sh script to use TESTS rather than passing
the test names within RUNTESTFLAGS. This addresses the following
issue:
I was running some tests like this:
make -C gdb check-all-boards TESTS="gdb.base/break*.exp"
And I was finding that I would get lots of DUPLICATE test results,
which is not what I expected.
What's happening here is that the 'make check-all-boards' rule runs
the 'make-check-all.sh' script, which then runs 'make check' with
various board files.
However, passing TESTS=... to the initial 'make check-all-boards'
command invocation automatically causes the TESTS value to be added to
the MAKEFLAGS environment variable, this is then picked up by the
later calls to 'make check'.
Now, in GDB's testfile/Makefile, we check for TESTS, and if this is
set, we expand the value and set `expanded_tests_or_none`. Otherwise,
if TESTS is not set, expanded_tests_or_none is left empty.
Finally, when handling 'make check', the value of
`expanded_tests_or_none` is passed through to dejagnu, along with the
RUNTESTFLAGS value.
What this means is that, when make-check-all.sh passes the test names
in the RUNTESTFLAGS, then dejagnu ends up seeing the list of tests
twice, once from RUNTESTFLAGS, and once from expanded_tests_or_none,
and this is why I was seeing duplicate testnames.
The easiest fix for the above is to have make-check-all.sh pass the
test names using TESTS="...", this will override the TESTS="..." value
already present in MAKEFLAGS, and means dejagnu will see the test
names just once.
Additionally, this is a start towards allowing parallel test running
from the make-check-all.sh script. Parallel test running only works
if the test names are passed in TESTS, and not in RUNTESTFLAGS.
Currently, in testsuite/Makefile, if RUNTESTFLAGS is not empty, then
we force single threaded test running. But with this change, at least
for the `local` board, we can now benefit from multi-threaded test
running, as this board has an empty RUNTESTFLAGS now. For the other
boards we'd need to set FORCE_PARALLEL in order to benefit from
parallel test running, but we'll need to double check that all the
board files actually support parallel test running first, so I'm
leaving that for another day.
|
|
Since the output section contents are copied from the input, don't
extend the output section size beyond the input section size.
PR binutils/33049
* objcopy.c (copy_section): Don't extend the output section
size beyond the input section size.
Signed-off-by: H.J. Lu <hjl.tools@gmail.com>
|
|
Report corrupted group section instead of trying to recover.
PR binutils/33050
* elf.c (bfd_elf_set_group_contents): Report corrupted group
section.
Signed-off-by: H.J. Lu <hjl.tools@gmail.com>
|
|
This commit adds filename completion for the shell command part of
the pipe command. This is a follow on from this commit:
commit 036e5c0c9121d0ac691dbf408a3bdf2bf3501d0f
Date: Mon May 19 20:54:54 2025 +0100
gdb: use quoted filename completion for the shell command
which fixed the completion for the 'shell' command itself.
Like with the 'shell' command, we don't offer completions of command
names pulled from $PATH, we just offer filename completion, which is
often useful for arguments being passed to commands. Maybe in the
future we could add completion for command names too (for both 'pipe'
and the 'shell' command), but that is left for a future commit.
There's some additional testing.
|
|
The use of user namespaces is required for normal users to use mount
namespaces. Consider trying this as an unprivileged user:
$ unshare --mount /bin/true
unshare: unshare failed: Operation not permitted
The problem here is that an unprivileged user doesn't have the
required permissions to create a new mount namespace. If, instead, we
do this:
$ unshare --mount --map-root-user /bin/true
then this will succeed. The new option causes unshare to create a
user namespace in which the unprivileged user is mapped to UID/GID 0,
and so gains all privileges (inside the namespace), the user is then
able to create the mount namespace as required.
So, how does this relate to GDB?
When a user attaches to a process running in a separate mount
namespace, GDB makes use of a separate helper process (see
linux_mntns_get_helper in nat/linux-namespaces.c), which will then use
the `setns` function to enter (or try to enter) the mount namespace of
the process GDB is attaching too. The helper process will then handle
file I/O requests received from GDB, and return the results back to
GDB, this allows GDB to access files within the mount namespace.
The problem here is that, switching to a mount namespace requires that
a process hold CAP_SYS_CHROOT and CAP_SYS_ADMIN capabilities within
its user namespace (actually it's a little more complex, see 'man 2
setns'). Assuming GDB is running as an unprivileged user, then GDB
will not have the required permissions.
However, if GDB enters the user namespace that the `unshare` process
created, then the current user will be mapped to UID/GID 0, and will
have the required permissions.
And so, this patch extends linux_mntns_access_fs (in
nat/linux-namespace.c) to first try and switch to the user namespace
of the inferior before trying to switch to the mount namespace. If
the inferior does have a user namespace, and does have elevated
privileges within that namespace, then this first switch by GDB will
mean that the second step, into the mount namespace, will succeed.
If there is no user namespace, or the inferior doesn't have elevated
privileges within the user namespace, then the switch into the mount
namespace will fail, just as it currently does, and the user will need
to give elevated privileges to GDB via some other mechanism (e.g. run
as root).
This work was originally posted here:
https://inbox.sourceware.org/gdb-patches/20230321120126.1418012-1-benjamin@sipsolutions.net
I (Andrew Burgess) have made some cleanups to the code to comply with
GDB's coding standard, and the test is entirely mine. This commit
message is also entirely mine -- the original message was very terse
and required the reader to understand how the various namespaces
work and interact. The above is my attempt to document what I now
understand about the problem being fixed.
I've left the original author in place as the core of the GDB change
itself is largely as originally presented, but any inaccuracies in the
commit message, or problems with the test, are all mine.
Co-Authored-by: Andrew Burgess <aburgess@redhat.com>
|
|
This commit works around a problem introduced by commit:
commit e58beedf2c8a1e0c78e0f57aeab3934de9416bfc
Date: Tue Jan 23 16:00:59 2024 +0000
gdb: attach to a process when the executable has been deleted
The above commit extended GDB for Linux, so that, of the executable
for a process had been deleted, GDB would instead try to use
/proc/PID/exe as the executable.
This worked by updating linux_proc_pid_to_exec_file to introduce the
/proc/PID/exe fallback. However, the result of
linux_proc_pid_to_exec_file is then passed to exec_file_find to
actually find the executable, and exec_file_find, will take into
account the sysroot. In addition, if GDB is attaching to a process in
a different MNT and/or PID namespace then the executable lookup is
done within that namespace.
This all means two things:
1. Just because linux_proc_pid_to_exec_file cannot see the
executable doesn't mean that GDB is actually going to fail to
find the executable, and
2. returning /proc/PID/exe isn't useful if we know GDB is then going
to look for this within a sysroot, or within some other
namespace (where PIDs might be different).
There was an initial attempt to fix this issue here:
https://inbox.sourceware.org/gdb-patches/20250511141517.2455092-4-kilger@sec.in.tum.de/
This proposal addresses the issue in PR gdb/32955, which is all about
the namespace side of the problem. The fix in this original proposal
is to check the MNT namespace inside linux_proc_pid_to_exec_file, and
for the namespace problem this is fine. But we should also consider
the sysroot problem.
And for the sysroot problem, the fix cannot fully live inside
linux_proc_pid_to_exec_file, as linux_proc_pid_to_exec_file is shared
between GDB and gdbserver, and gdbserver has no sysroot.
And so, I propose a slightly bigger change.
Now, linux_proc_pid_to_exec_file takes a flag which indicates if
GDB (or gdbserver) will look for the inferior executable in the
local file system, where local means the same file system as GDB (or
gdbserver) is running in.
This local file system check is true if:
1. The MNT namespace of the inferior is the same as for GDB, and
2. for GDB only, the sysroot must either be empty, or 'target:'.
If the local file system check is false then GDB (or gdbserver) is
going to look elsewhere for the inferior executable, and so, falling
back to /proc/PID/exe should not be done, as GDB will end up looking
for this file in the sysroot, or within the alternative MNT
namespace (which in also likely to be a different PID namespace).
Now this is all a bit of a shame really. It would be nice if
linux_proc_pid_to_exec_file could return /proc/PID/exe in such a way
that exec_file_find would know that the file should NOT be looked for
in the sysroot, or in the alternative namespace. But fixing that
problem would be a much bigger change, so for now lets just disable
the /proc/PID/exe fallback for cases where it might not work.
For testing, the sysroot case is now tested.
I don't believe we have any alternative namespace testing. It would
certainly be interesting to add some, but I'm not proposing any with
this patch, so the code for checking the MNT namespace has been tested
manually by me, but isn't covered by a new test I'm adding here.
Author of the original fix is listed as co-author here. Credit for
identifying the original problem, and proposing a solution belongs to
them.
Co-Authored-By: Fabian Kilger <kilger@sec.in.tum.de>
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=32955
|
|
Currently, when attaching to a process, if the user hasn't told GDB
which executable they are going to be debugging, GDB will try to
figure out the executable from the running process.
There are two (for this patch) interesting places where this can fail,
both in exec_file_locate_attach.
First GDB calls target_pid_to_exec_file, this does target specific
"stuff" to find the name of the executable file. If this returns NULL
then GDB will give a warning and return.
After this we need to "find" the executable. This is where we apply
things like the sysroot in order to transform the executable path.
This is done by calling exec_file_find, and this too can return NULL
to indicate that the executable couldn't be found.
Currently, if exec_file_find returns NULL then GDB doesn't give a
warning, instead we push on and call try_open_exec_file passing in the
NULL pointer as the filename string. This has the effect of removing
the current executable from the current program space.
However, exec_file_locate_attach already checks there is no executable
attached to the current program space. If there was, then there would
be no need to try and lookup the executable from the running process.
So calling try_open_exec_file with a NULL string is, I claim,
pointless.
But worse, calling try_open_exec_file with a NULL string means that
GDB prints the message: "No executable file now.", which, while
correct, isn't (I think) very helpful. To me this message indicates
that we've moved from a state of having an executable to a state of
not having one, which isn't correct.
I think we should introduce a new warning in exec_file_locate_attach,
which is printed if the executable cannot be found.
So, before this patch GDB's output looked like this:
(gdb) attach 12345
Attaching to process 12345
No executable file now.
warning: Could not load vsyscall page because no executable was specified
0x00007f0978b94557 in ?? ()
(gdb)
After this patch the output now looks like this:
(gdb) attach 12345
Attaching to process 12345
No executable has been specified, and target executable /tmp/my-exec (deleted) could not be found. Try using the "file" command.
warning: Could not load vsyscall page because no executable was specified
0x00007f0978b94557 in ?? ()
(gdb)
This warning includes the name of the file that GDB was looking for,
and gives a hint that the 'file' command should be used to tell GDB
which executable is being debugged. Much better.
There's no test for this change in this commit. The next commit fixes
another (semi-related) bug, and includes a test that checks for this
warning string.
|
|
Tom de Vries reported a build failure on x86_64-w64-mingw32 after
commit:
commit bd389c9515d240f55b117075b43184efdea41287
Date: Wed Jun 11 22:52:16 2025 +0200
gdb: implement linux namespace support for fileio_lstat and vFile::lstat
The build failure looks like this:
../../src/gdbserver/hostio.cc: In function 'void handle_lstat(char*, int*)':
../../src/gdbserver/hostio.cc:544:63: error: cannot convert '_stat64*' to 'stat*'
544 | ret = the_target->multifs_lstat (hostio_fs_pid, filename, &st);
| ^~~
| |
| _stat64*
In file included from ./../../src/gdbserver/server.h:58,
from <command-line>:
./../../src/gdbserver/target.h:448:74: note: initializing argument 3 of 'virtual int process_stratum_target::multifs_lstat(int, const char*, stat*)'
448 | virtual int multifs_lstat (int pid, const char *filename, struct stat *sb);
| ~~~~~~~~~~~~~^~
The problem is that in sys/stat.h for mingw, 'stat' is #defined to
_stat64, but target.h doesn't include sys/stat.h, and so doesn't see
this #define.
However, target.h does, by luck, manages to see the actual definition
of 'struct stat', which isn't in sys/stat.h itself, but is in some
other header that just happens to be pulled in by chance.
As a result of all this, the declaration of
process_stratum_target::multifs_lstat in target.h uses 'struct stat'
for its argument type, while the call in hostio.cc, uses 'struct
_stat64' as its argument type, which causes the build error seen
above.
The fix is to include sys/stat.h in target.h so that the declaration's
argument type will change to 'struct _stat64' (via the #define).
|
|
|
|
|
|
There are some OpenRISC CPUs that have their binaries stored in
little-endian format. Using objdump to disassemble these is
problematic, as some instructions fail to disassemble, for example:
objdump -D -b binary -EB -m or1k test_be.bin
0: 18 60 07 27 l.movhi r3,0x727
4: a8 63 0e 00 l.ori r3,r3,0xe00
8: 9c 63 ff ff l.addi r3,r3,-1
c: bc 43 00 00 l.sfgtui r3,0
10: 13 ff ff fe l.bf 0x8
14: 44 00 48 00 l.jr r9
objdump -D -b binary -EL -m or1k test_le.bin
0: 27 07 60 18 *unknown*
4: 00 0e 63 a8 l.ori r3,r3,0xe00
8: ff ff 63 9c *unknown*
c: 00 00 43 bc l.sfgtui r3,0
10: fe ff ff 13 *unknown*
14: 00 48 00 44 l.jr r9
It was found that the hash function was using the still little-endian
buffer to extract the opcode used for the hash lookup. This didn't work
as it was pulling the wrong hashcode causing instruction lookup to fail.
Fix the hash function by using the normalized/byte-swapped value instead
of the buffer.
Signed-off-by: Stafford Horne <shorne@gmail.com>
|
|
|
|
Theoretically, in functions core_addr_to_string_nz() and
core_addr_to_string(), strcat() can overflow, so use a safe
approach using xsnprintf().
Change-Id: Ib9437450b3634dc35077234f462a03a8640242d4
|
|
This patch simplifies the code at two points by removing redundant
null checks. There is no functional impact.
Reviewed-By: Keith Seitz <keiths@redhat.com>
Approved-By: Pedro Alves <pedro@palves.net>
Change-Id: I76e1c7fad00e8fcb24ced7bfd75d19cdd6266c32
|
|
This patch adds support for following system registers and the spec
can be found here[1].
1. PMBSR_EL12, PMBSR_EL2, PMBSR_EL3, PMBMAR_EL1 depends on FEAT_SPE
and Armv9.5-A architecture and these are enabled by passing
-march=armv9.5-a+profile.
2. TRBSR_EL12, TRBSR_EL2, and TRBSR_EL3 depends Armv9.5-A architecture
and these are enabled by passing -march=armv9.5-a.
3. HFGITR2_EL2 depends on Armv8.8-A architecture and enabled by passing
-march=armv8.8-a.
[1]: https://developer.arm.com/documentation/ddi0601/2025-03/AArch64-Registers?lang=en
|
|
Currently gdbserver uses the require_int() function to parse the
requested offset (in vFile::pread packet and the like). This function
allows integers up to 0x7fffffff (to fit in 32-bit int), however the
offset (for the pread system call) has an off_t type which can be
larger than 32-bit.
This patch allows require_int() function to parse offset up to the
maximum value implied by the off_t type.
Approved-By: Pedro Alves <pedro@palves.net>
Change-Id: I3691bcc1ab1838c0db7f8b82d297d276a5419c8c
|
|
|
|
`pre-commit run --all-files` found this.
Change-Id: I8db09b12cf184d32351ff2c579bdaa8cf6f80ac3
|
|
Change the messages to reflect that these numbers includes type units,
not only compile units.
Change-Id: Id2f511d4666e5cf92112be917d72ff76791b7e1d
Approved-by: Kevin Buettner <kevinb@redhat.com>
|
|
FEAT_LSFE - Large System Float Extension - implements A64 base atomic
floating-point in-memory instructions.
|
