From: Guinevere Larsen <guinevere@redhat.com>
To: gdb-patches@sourceware.org, guinevere@redhat.com
Subject: Re: [PING][PATCH v8] gdb: Print linker namespace when showing a frame
Date: Fri, 18 Sep 2026 09:54:16 -0300 [thread overview]
Message-ID: <2dd2c450-3594-49e2-bada-37e0ed59cc05@redhat.com> (raw)
In-Reply-To: <95bf1ee9-55cb-44ae-a0e8-c3e73934f1bc@redhat.com>
ping :)
On 9/2/26 4:30 PM, Guinevere Larsen wrote:
> Ping :)
>
> On 8/6/26 3:42 PM, Guinevere Larsen wrote:
>> Ping :)
>>
>> On 6/22/26 5:34 PM, Guinevere Larsen wrote:
>>> Ping :)
>>>
>>> On 5/7/26 10:48 AM, Guinevere Larsen wrote:
>>>> When a user is stopped in a private linker namespace, the only way for
>>>> them to realize that is using the _linker_namespace convenience
>>>> variable. While serviceable, this is a sub-optimal solution, as most
>>>> users are unaware of convenience variables.
>>>>
>>>> This commit introduces a new way for users to be informed of the
>>>> linker
>>>> namespace of a function, by printing it along with the function name.
>>>> This is done by using the proposed syntax for symbols and locations,
>>>> like so:
>>>>
>>>> #0 [[0]]::main ()
>>>>
>>>> This is done by introducing a new function on solib.h,
>>>> linker_namespace_prefix, that returns either an empty string (if only
>>>> one namespace is active, or we can't determine the PC for the
>>>> frame), or
>>>> the correct identifier. This should be printed manually every time GDB
>>>> will print a frame, so that we don't confuse other systems that
>>>> look at
>>>> function names.
>>>>
>>>> A previous revision would break GDB's ability to strip the
>>>> parameters of
>>>> C++ symbols, so this commit also adds a C++ SO to the test. However,
>>>> libstdc++ never gets unloaded, so many changes were required to
>>>> make the
>>>> rest of the test still valid.
>>>>
>>>> Reviewed-By: Eli Zaretskii <eliz@gnu.org>
>>>> ---
>>>> gdb/NEWS | 4 +
>>>> gdb/doc/gdb.texinfo | 9 +++
>>>> gdb/solib.c | 43 ++++++++---
>>>> gdb/solib.h | 13 ++++
>>>> gdb/stack.c | 2 +
>>>> .../gdb.base/dlmopen-ns-ids-framefilter.py | 31 ++++++++
>>>> .../gdb.base/dlmopen-ns-ids-libc++.cc | 36 +++++++++
>>>> gdb/testsuite/gdb.base/dlmopen-ns-ids-main.c | 19 ++++-
>>>> gdb/testsuite/gdb.base/dlmopen-ns-ids.exp | 74
>>>> ++++++++++++++++---
>>>> gdb/testsuite/gdb.mi/mi-dlmopen.exp | 2 +-
>>>> 10 files changed, 206 insertions(+), 27 deletions(-)
>>>> create mode 100644
>>>> gdb/testsuite/gdb.base/dlmopen-ns-ids-framefilter.py
>>>> create mode 100644 gdb/testsuite/gdb.base/dlmopen-ns-ids-libc++.cc
>>>>
>>>> diff --git a/gdb/NEWS b/gdb/NEWS
>>>> index 480e1854002..941178c06a4 100644
>>>> --- a/gdb/NEWS
>>>> +++ b/gdb/NEWS
>>>> @@ -78,6 +78,10 @@
>>>> * The Windows native target now supports non-stop mode. This feature
>>>> requires Windows 10 or later.
>>>> +* When debugging an inferior with multiple linker namespaces,
>>>> functions
>>>> + will be printed like so [[N]]::foo, where N is the namespace that
>>>> + contains foo. For example [[0]]::main.
>>>> +
>>>> * New targets
>>>> GNU/Linux/MicroBlaze (gdbserver) microblazeel-*linux*
>>>> diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
>>>> index ab0216ff477..98a43f3a57e 100644
>>>> --- a/gdb/doc/gdb.texinfo
>>>> +++ b/gdb/doc/gdb.texinfo
>>>> @@ -22431,6 +22431,15 @@ possibility that those objects may
>>>> cross-talk. Each set of isolated
>>>> shared objects is said to belong to a ``namespace'', and linker
>>>> related
>>>> actions such as relocations do not cross namespace boundaries.
>>>> +When multiple linker namespaces are active, @value{GDBN} will
>>>> print the
>>>> +identifier of the namespace that contains the function, when a
>>>> function
>>>> +is printed. It will look like this:
>>>> +
>>>> +@smallexample
>>>> +#0 [[1]]::inc (n=0) at dlmopen-ns-ids-lib.c
>>>> +#1 0x0000000000400591 in [[0]]::main () at dlmopen-ns-ids-main.c
>>>> +@end smallexample
>>>> +
>>>> @kindex info dll
>>>> @item info dll @var{regex}
>>>> This is an alias of @code{info sharedlibrary}.
>>>> diff --git a/gdb/solib.c b/gdb/solib.c
>>>> index 782b844aa2f..add842b5771 100644
>>>> --- a/gdb/solib.c
>>>> +++ b/gdb/solib.c
>>>> @@ -1796,6 +1796,36 @@ solib_linker_namespace_count (program_space
>>>> *pspace)
>>>> return 0;
>>>> }
>>>> +/* See solib.h. */
>>>> +
>>>> +int
>>>> +linker_namespace_for_addr (CORE_ADDR addr, program_space *pspace)
>>>> +{
>>>> + for (const solib &so : pspace->solibs ())
>>>> + if (solib_contains_address_p (so, addr))
>>>> + {
>>>> + if (so.ops ().supports_namespaces ())
>>>> + return so.ops ().find_solib_ns (so);
>>>> + break;
>>>> + }
>>>> +
>>>> + return 0;
>>>> +}
>>>> +
>>>> +/* See solib.h. */
>>>> +
>>>> +std::string
>>>> +linker_namespace_prefix (const frame_info_ptr &frame)
>>>> +{
>>>> + std::optional<CORE_ADDR> pc = get_frame_pc_if_available (frame);
>>>> + program_space *pspace = get_frame_program_space (frame);
>>>> + if (!pc.has_value () || solib_linker_namespace_count (pspace) <= 1)
>>>> + return std::string ("");
>>>> +
>>>> + return string_printf ("[[%d]]::", linker_namespace_for_addr
>>>> (pc.value (),
>>>> + pspace));
>>>> +}
>>>> +
>>>> /* Implementation of the linker_namespace convenience variable.
>>>> This returns the GDB internal identifier of the linker
>>>> namespace,
>>>> @@ -1804,19 +1834,10 @@ solib_linker_namespace_count (program_space
>>>> *pspace)
>>>> static value *
>>>> linker_namespace_make_value (gdbarch *gdbarch, internalvar *var,
>>>> - void *ignore)
>>>> + void *ignore)
>>>> {
>>>> - int nsid = 0;
>>>> CORE_ADDR curr_pc = get_frame_pc (get_selected_frame ());
>>>> -
>>>> - for (const solib &so : current_program_space->solibs ())
>>>> - if (solib_contains_address_p (so, curr_pc))
>>>> - {
>>>> - if (so.ops ().supports_namespaces ())
>>>> - nsid = so.ops ().find_solib_ns (so);
>>>> -
>>>> - break;
>>>> - }
>>>> + int nsid = linker_namespace_for_addr (curr_pc,
>>>> current_program_space);
>>>> /* If the PC is not in an SO, or the solib_ops doesn't support
>>>> linker namespaces, the inferior is in the default
>>>> namespace. */
>>>> diff --git a/gdb/solib.h b/gdb/solib.h
>>>> index 564655c4a57..39931bb1b93 100644
>>>> --- a/gdb/solib.h
>>>> +++ b/gdb/solib.h
>>>> @@ -334,6 +334,19 @@ extern const char *solib_name_from_address
>>>> (struct program_space *, CORE_ADDR);
>>>> extern bool solib_contains_address_p (const solib &, CORE_ADDR);
>>>> +/* Given the address ADDR, return which linker namespace contains
>>>> + this address in PSPACE. If ADDR is present in multiple
>>>> namespaces,
>>>> + an arbitrary one is returned. If ADDR is not present in any
>>>> + namespace, or if the target doesn't support linker namespaces,
>>>> + returns 0. */
>>>> +
>>>> +extern int linker_namespace_for_addr (CORE_ADDR addr,
>>>> program_space *pspace);
>>>> +
>>>> +/* Calculate the linker namespace identifier, and turn it into a
>>>> prefix
>>>> + to be used to print a frame. */
>>>> +
>>>> +extern std::string linker_namespace_prefix (const frame_info_ptr
>>>> &frame);
>>>> +
>>>> /* Return whether the data starting at VADDR, size SIZE, must be
>>>> kept
>>>> in a core file for shared libraries loaded before "gcore" is used
>>>> to be handled correctly when the core file is loaded. This only
>>>> diff --git a/gdb/stack.c b/gdb/stack.c
>>>> index 7329430adab..ac270718a91 100644
>>>> --- a/gdb/stack.c
>>>> +++ b/gdb/stack.c
>>>> @@ -1306,6 +1306,7 @@ print_frame (struct ui_out *uiout,
>>>> annotate_frame_function_name ();
>>>> string_file stb;
>>>> + gdb_puts (linker_namespace_prefix (frame).c_str (), &stb);
>>>> gdb_puts (funname ? funname.get () : "??", &stb);
>>>> uiout->field_stream ("func", stb, function_name_style.style ());
>>>> uiout->wrap_hint (3);
>>>> @@ -1466,6 +1467,7 @@ info_frame_command_core (const frame_info_ptr
>>>> &fi, bool selected_frame_p)
>>>> fputs_styled ("<unavailable>", metadata_style.style (),
>>>> gdb_stdout);
>>>> gdb_stdout->wrap_here (3);
>>>> + gdb_puts (linker_namespace_prefix (fi).c_str ());
>>>> if (funname.get () != nullptr)
>>>> {
>>>> gdb_puts (" in ");
>>>> diff --git a/gdb/testsuite/gdb.base/dlmopen-ns-ids-framefilter.py
>>>> b/gdb/testsuite/gdb.base/dlmopen-ns-ids-framefilter.py
>>>> new file mode 100644
>>>> index 00000000000..11d96fcfbc4
>>>> --- /dev/null
>>>> +++ b/gdb/testsuite/gdb.base/dlmopen-ns-ids-framefilter.py
>>>> @@ -0,0 +1,31 @@
>>>> +# Copyright (C) 2026 Free Software Foundation, Inc.
>>>> +
>>>> +# This program is free software; you can redistribute it and/or
>>>> modify
>>>> +# it under the terms of the GNU General Public License as
>>>> published by
>>>> +# the Free Software Foundation; either version 3 of the License, or
>>>> +# (at your option) any later version.
>>>> +#
>>>> +# This program is distributed in the hope that it will be useful,
>>>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>>>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>>>> +# GNU General Public License for more details.
>>>> +#
>>>> +# You should have received a copy of the GNU General Public License
>>>> +# along with this program. If not, see
>>>> <http://www.gnu.org/licenses/>.
>>>> +
>>>> +
>>>> +# This file is part of the GDB testsuite. It tests if printing the
>>>> +# linker namespace id is affected by a null frame filter
>>>> +
>>>> +import gdb
>>>> +
>>>> +
>>>> +class NullFilter:
>>>> + def __init__(self):
>>>> + self.name = "null filter"
>>>> + self.priority = 100
>>>> + self.enabled = True
>>>> + gdb.frame_filters[self.name] = self
>>>> +
>>>> + def filter(self, iterator):
>>>> + return iterator
>>>> diff --git a/gdb/testsuite/gdb.base/dlmopen-ns-ids-libc++.cc
>>>> b/gdb/testsuite/gdb.base/dlmopen-ns-ids-libc++.cc
>>>> new file mode 100644
>>>> index 00000000000..9abf3f25096
>>>> --- /dev/null
>>>> +++ b/gdb/testsuite/gdb.base/dlmopen-ns-ids-libc++.cc
>>>> @@ -0,0 +1,36 @@
>>>> +/* This testcase is part of GDB, the GNU debugger.
>>>> +
>>>> + Copyright 2025-2026 Free Software Foundation, Inc.
>>>> +
>>>> + This program is free software; you can redistribute it and/or
>>>> modify
>>>> + it under the terms of the GNU General Public License as
>>>> published by
>>>> + the Free Software Foundation; either version 3 of the License, or
>>>> + (at your option) any later version.
>>>> +
>>>> + This program is distributed in the hope that it will be useful,
>>>> + but WITHOUT ANY WARRANTY; without even the implied warranty of
>>>> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>>>> + GNU General Public License for more details.
>>>> +
>>>> + You should have received a copy of the GNU General Public License
>>>> + along with this program. If not, see
>>>> <http://www.gnu.org/licenses/>.
>>>> +
>>>> +*/
>>>> +
>>>> +int gdb_dlmopen_glob = 0;
>>>> +
>>>> +int
>>>> +cpp_inc (int n)
>>>> +{
>>>> + int amount = gdb_dlmopen_glob;
>>>> + return n + amount; /* bp.inc. */
>>>> +}
>>>> +
>>>> +extern "C" {
>>>> + __attribute__((visibility ("default")))
>>>> + int
>>>> + inc (int n)
>>>> + {
>>>> + return cpp_inc (n);
>>>> + }
>>>> +}
>>>> diff --git a/gdb/testsuite/gdb.base/dlmopen-ns-ids-main.c
>>>> b/gdb/testsuite/gdb.base/dlmopen-ns-ids-main.c
>>>> index 4f08a5135b2..f13eb9168de 100644
>>>> --- a/gdb/testsuite/gdb.base/dlmopen-ns-ids-main.c
>>>> +++ b/gdb/testsuite/gdb.base/dlmopen-ns-ids-main.c
>>>> @@ -27,7 +27,7 @@
>>>> int
>>>> main (void)
>>>> {
>>>> - void *handle[4];
>>>> + void *handle[5];
>>>> int (*fun) (int);
>>>> Lmid_t lmid;
>>>> int dl;
>>>> @@ -41,20 +41,31 @@ main (void)
>>>> handle[2] = dlmopen (LM_ID_NEWLM, DSO_NAME, RTLD_LAZY |
>>>> RTLD_LOCAL);
>>>> assert (handle[2] != NULL);
>>>> + /* We open the C++ SO on the main namespace because libstdc++
>>>> marks a
>>>> + lot of objects as GNU_UNIQUE meaning that they effectively
>>>> can't be
>>>> + unloaded, so the final namespace would never be fully
>>>> unloaded. */
>>>> + handle[3] = dlmopen (LM_ID_BASE, CPP_DSO_NAME, RTLD_LAZY |
>>>> RTLD_LOCAL);
>>>> + assert (handle[3] != NULL);
>>>> +
>>>> for (dl = 2; dl >= 0; dl--)
>>>> {
>>>> fun = dlsym (handle[dl], "inc");
>>>> fun (dl);
>>>> }
>>>> + /* C++ call, to make sure the linker namespace doesn't break
>>>> anything
>>>> + of the C++ handling of symbols. */
>>>> + fun = dlsym (handle[3], "inc");
>>>> + fun (-1);
>>>> dlclose (handle[0]); /* TAG: first dlclose */
>>>> dlclose (handle[1]); /* TAG: second dlclose */
>>>> dlclose (handle[2]); /* TAG: third dlclose */
>>>> + dlclose (handle[3]); /* TAG: fourth dlclose */
>>>> - handle[3] = dlmopen (LM_ID_NEWLM, DSO_NAME, RTLD_LAZY |
>>>> RTLD_LOCAL);
>>>> - dlinfo (handle[3], RTLD_DI_LMID, &lmid);
>>>> + handle[4] = dlmopen (LM_ID_NEWLM, DSO_NAME, RTLD_LAZY |
>>>> RTLD_LOCAL);
>>>> + dlinfo (handle[4], RTLD_DI_LMID, &lmid);
>>>> - dlclose (handle[3]); /* TAG: fourth dlclose */
>>>> + dlclose (handle[4]); /* TAG: final dlclose */
>>>> return 0;
>>>> }
>>>> diff --git a/gdb/testsuite/gdb.base/dlmopen-ns-ids.exp
>>>> b/gdb/testsuite/gdb.base/dlmopen-ns-ids.exp
>>>> index e40bc213454..cc1696f4229 100644
>>>> --- a/gdb/testsuite/gdb.base/dlmopen-ns-ids.exp
>>>> +++ b/gdb/testsuite/gdb.base/dlmopen-ns-ids.exp
>>>> @@ -21,20 +21,29 @@
>>>> require allow_dlmopen_tests
>>>> -standard_testfile -main.c -lib.c
>>>> +standard_testfile -main.c -lib.c -framefilter.py -libc++.cc
>>>> set srcfile_lib $srcfile2
>>>> +set srcfile_libcpp $srcfile4
>>>> set so_name dlmopen-lib.so
>>>> +set cpp_so_name dlmopen-libc++.so
>>>> set binfile_lib [standard_output_file $so_name]
>>>> +set binfile_libcpp [standard_output_file $cpp_so_name]
>>>> if { [build_executable "build shlib" $binfile_lib $srcfile_lib \
>>>> [list debug shlib]] == -1 } {
>>>> return
>>>> }
>>>> +if { [build_executable "build shlib" $binfile_libcpp
>>>> $srcfile_libcpp \
>>>> + [list debug shlib c++]] == -1 } {
>>>> + return
>>>> +}
>>>> +
>>>> if { [build_executable "failed to build" $testfile $srcfile \
>>>> [list additional_flags=-DDSO_NAME=\"$binfile_lib\" \
>>>> - shlib_load debug]] } {
>>>> + additional_flags=-DCPP_DSO_NAME=\"$binfile_libcpp\" \
>>>> + shlib_load debug]] } {
>>>> return
>>>> }
>>>> @@ -107,11 +116,13 @@ proc check_no_duplicates {} {
>>>> proc get_first_so_ns {} {
>>>> set ns -1
>>>> set lib_regexp [string_to_regexp ${::binfile_lib}]
>>>> + set libcpp_regexp [string_to_regexp ${::binfile_libcpp}]
>>>> + set both_regexp "($lib_regexp|$libcpp_regexp)"
>>>> gdb_test_multiple "info sharedlibrary $::so_name" "get SO
>>>> namespace" -lbl {
>>>> -re "\r\nFrom\\s+To\\s+\(Linker
>>>> NS\\s+\)?Syms\\s+Read\\s+Shared Object Library(?=\r\n)" {
>>>> exp_continue
>>>> }
>>>> - -re
>>>> "\r\n$::hex\\s+$::hex\\s+($::decimal)\\s+\[^\r\n]+${lib_regexp}(?=\r\n)"
>>>> {
>>>> + -re
>>>> "\r\n$::hex\\s+$::hex\\s+($::decimal)\\s+\[^\r\n]+${both_regexp}(?=\r\n)"
>>>> {
>>>> if {$ns == -1} {
>>>> set ns $expect_out(1,string)
>>>> }
>>>> @@ -155,8 +166,8 @@ proc test_info_shared {} {
>>>> gdb_test "next" ".*third dlclose.*" "close second library"
>>>> gdb_assert {[get_first_so_ns] == 3} "before closing two
>>>> libraries"
>>>> - gdb_breakpoint [gdb_get_line_number "TAG: fourth dlclose"]
>>>> - gdb_continue_to_breakpoint "TAG: fourth dlclose"
>>>> + gdb_breakpoint [gdb_get_line_number "TAG: final dlclose"]
>>>> + gdb_continue_to_breakpoint "TAG: final dlclose"
>>>> # As of writing this test, glibc's LMID is just an index on
>>>> an array of
>>>> # namespaces. After closing a namespace, requesting a new
>>>> one will
>>>> # return the index of the lowest-closed namespace, so this
>>>> will likely
>>>> @@ -192,6 +203,11 @@ proc_with_prefix test_conv_vars {} {
>>>> gdb_test "print \$_linker_namespace" ".* = 0" \
>>>> "Still in the default namespace"
>>>> + # There should be no namespace ID visible, since there's
>>>> + # only one namespace loaded.
>>>> + gdb_test "backtrace" "\#0\\s+main .*" \
>>>> + "No namespace ID in backtrace"
>>>> +
>>>> gdb_breakpoint "inc" allow-pending
>>>> gdb_breakpoint [gdb_get_line_number "TAG: first dlclose"]
>>>> @@ -200,14 +216,41 @@ proc_with_prefix test_conv_vars {} {
>>>> gdb_test "print \$_linker_namespace" ".* = $dl" \
>>>> "Verify we're in namespace $dl"
>>>> +
>>>> + gdb_test "frame" "\#0\\s+\\\[\\\[$dl\\\]\\\]::inc.*" \
>>>> + "Namespace ID in the frame"
>>>> }
>>>> # Check that we display the namespace of the selected
>>>> # frame, not the lowermost one.
>>>> - gdb_test "up" "\#1.*in main.*"
>>>> + gdb_test "up" "\#1.*in \\\[\\\[0\\\]\\\]::main.*"
>>>> gdb_test "print \$_linker_namespace" ".* = 0" \
>>>> "print namespace of selected frame"
>>>> + gdb_test "backtrace" [multi_line \
>>>> + "\#0\\s+\\\[\\\[1\\\]\\\]::inc \[^\r\n\]*" \
>>>> + "\#1\\s+$::hex in \\\[\\\[0\\\]\\\]::main \[^\r\n\]*" ] \
>>>> + "raw backtrace"
>>>> +
>>>> + set remote_python_file [gdb_remote_download host \
>>>> + $::srcdir/$::subdir/$::srcfile3]
>>>> + gdb_test_no_output "source $remote_python_file" "load frame
>>>> filter"
>>>> + gdb_test "backtrace" [multi_line \
>>>> + "\#0\\s+\\\[\\\[1\\\]\\\]::inc \[^\r\n\]*" \
>>>> + "\#1\\s+$::hex in \\\[\\\[0\\\]\\\]::main \[^\r\n\]*" ] \
>>>> + "With null frame filter"
>>>> +
>>>> + # Ensure that the code for identifying and removing parameters
>>>> for
>>>> + # C++ symbols still works with linker namespace IDs.
>>>> + gdb_continue_to_breakpoint "inc on c++ SO"
>>>> + gdb_test "step" "\\\[\\\[0\\\]\\\]::cpp_inc \\\(n=-1\\\).*"
>>>> "step into c++ symbol"
>>>> + gdb_test "backtrace" [multi_line \
>>>> + "\#0\\s+\\\[\\\[0\\\]\\\]::cpp_inc \\\(n=-1\\\)\[^\r\n\]*" \
>>>> + "\#1\\s+$::hex in \\\[\\\[0\\\]\\\]::inc
>>>> \\\(n=-1\\\)\[^\r\n\]*" \
>>>> + "\#2\\s+$::hex in \\\[\\\[0\\\]\\\]::main \[^\r\n\]*" ] \
>>>> + "bt through c++ symbols"
>>>> +
>>>> + # Check that everything still works as we close namespaces.
>>>> gdb_continue_to_breakpoint "first dlclose"
>>>> gdb_test "print \$_linker_namespace_count" "4" "all SOs loaded"
>>>> @@ -264,15 +307,23 @@ proc test_info_linker_namespaces {} {
>>>> # Some systems may add libc and libm to every loaded
>>>> namespace,
>>>> # others may load only one or neither, because the SO doesn't
>>>> - # actually use either library. The best we can do is check if
>>>> - # we found the dynamic linker, and up to 2 more libraries.
>>>> - gdb_assert {$n_libraries <= 3} "the correct number of
>>>> libraries was reported"
>>>> + # actually use either library. Similarly, because of the c++
>>>> + # SO, libstdc++ and libgcc_s may be included. The best we can
>>>> + # do is check if we found the dynamic linker, the SO, and up to
>>>> + # 4 more libraries. Ensuring that only the right one shows up is
>>>> + # in the foreach block.
>>>> + gdb_assert {$n_libraries <= 6} "the correct number of
>>>> libraries was reported"
>>>> set binfile_lib_re [string_to_regexp $::binfile_lib]
>>>> - foreach_with_prefix ns {1 2 3} {
>>>> + foreach_with_prefix ns {1 2 3 0} {
>>>> set found_test_so false
>>>> set n_libraries 999
>>>> + set expected_libs 4
>>>> + if {$ns == 0} {
>>>> + set expected_libs 6
>>>> + set binfile_lib_re [string_to_regexp $::binfile_libcpp]
>>>> + }
>>>> gdb_test_multiple "info linker-namespaces $ns" "print
>>>> namespace $ns" {
>>>> -re ".*($::decimal) librar(?:y|ies) loaded in linker
>>>> namespace $ns:\r\n" {
>>>> @@ -290,7 +341,8 @@ proc test_info_linker_namespaces {} {
>>>> # actually use either library. The best we can do is check if
>>>> # we found the dynamic linker, the test SO, and maybe up to 2
>>>> # more libraries.
>>>> - gdb_assert {$n_libraries <= 4} "the correct number of
>>>> libraries was reported"
>>>> + gdb_assert {$n_libraries <= $expected_libs} \
>>>> + "the correct number of libraries was reported"
>>>> gdb_assert {$found_test_so} "this testfile's SO was reported"
>>>> }
>>>> diff --git a/gdb/testsuite/gdb.mi/mi-dlmopen.exp
>>>> b/gdb/testsuite/gdb.mi/mi-dlmopen.exp
>>>> index ff854ac7dd4..832c6ea6fe6 100644
>>>> --- a/gdb/testsuite/gdb.mi/mi-dlmopen.exp
>>>> +++ b/gdb/testsuite/gdb.mi/mi-dlmopen.exp
>>>> @@ -156,7 +156,7 @@ proc check_solib_unload_events {} {
>>>> -disp keep -func main -file ".*$::srcfile" -line $::bp_main
>>>> # Run past all the dlopen and dlmopen calls.
>>>> - mi_execute_to "exec-continue" "breakpoint-hit" main "" ".*"
>>>> $::bp_loaded \
>>>> + mi_execute_to "exec-continue" "breakpoint-hit"
>>>> {\[\[0\]\]::main} "" ".*" $::bp_loaded \
>>>> {"" "disp=\"keep\""} "continue until all libraries are loaded"
>>>> # Check that the dynamic linker has now been loaded
>>>> multiple times.
>>>>
>>>> base-commit: 9e6e0c3cd9e6d5ee4953f72a805cff5f150aa2fd
>>>
>>>
>>
--
Cheers,
Guinevere Larsen
it/its
she/her (deprecated)
prev parent reply other threads:[~2026-09-18 12:54 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-05-07 13:48 [PATCH " Guinevere Larsen
2026-06-22 20:34 ` [PING][PATCH " Guinevere Larsen
2026-06-23 11:57 ` Eli Zaretskii
2026-06-23 12:07 ` Guinevere Larsen
2026-08-06 18:42 ` Guinevere Larsen
2026-09-02 19:30 ` Guinevere Larsen
2026-09-18 12:54 ` Guinevere Larsen [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=2dd2c450-3594-49e2-bada-37e0ed59cc05@redhat.com \
--to=guinevere@redhat.com \
--cc=gdb-patches@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox