aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorPedro Alves <palves@redhat.com>2016-08-22 20:05:09 +0100
committerPedro Alves <palves@redhat.com>2016-08-22 20:05:09 +0100
commit6bb90213cb7b8e2f3be20f2e46f11f57f0c9ce55 (patch)
treec4d9f2bf1199d01f59be4dc9b0c181432b008a49
parent5fbe0d878a691b9be42bb2bdebd027ac3dfd38c2 (diff)
downloadgdb-6bb90213cb7b8e2f3be20f2e46f11f57f0c9ce55.zip
gdb-6bb90213cb7b8e2f3be20f2e46f11f57f0c9ce55.tar.gz
gdb-6bb90213cb7b8e2f3be20f2e46f11f57f0c9ce55.tar.bz2
Fix PR gdb/20505 - Make vDSO detection work with core files
Loading a core dump that was either generated on a system running pristine glibc master, or on a Fedora/RHEL system with LD_DEBUG=unused set in the environment, solib-svr4.c:svr4_current_sos fails to filter out the vDSO, resulting in: (gdb) core-file corefile.core^M [New LWP 2362]^M warning: Could not load shared library symbols for linux-vdso.so.1.^M Do you need "set solib-search-path" or "set sysroot"?^M Core was generated by `build-gdb/gdb/testsuite/outputs/gdb.base/corefile/'.^M ... The problem is that gdbarch_vsyscall_range does not support core inferiors at all. When live debugging, we're finding the vDSO's start address with auxv/AT_SYSINFO_EHDR, and then we find the vDSO's size by look for the corresponding mapping, by parsing /proc/PID/maps. When debugging a core dump, we can also determine the starting address from auxv/AT_SYSINFO_EHDR. However, we obviously can't read the core mappings out of the host's /proc. But we can instead look for a corresponding load segment in the core's bfd. gdb/ChangeLog: 2016-08-22 Pedro Alves <palves@redhat.com> PR gdb/20505 * linux-tdep.c (linux_vsyscall_range_raw): For core inferiors, find the vDSO's start address with AT_SYSINFO_EHDR too, and determine the vDSO's size by finding the PT_LOAD segment that matches AT_SYSINFO_EHDR. gdb/testsuite/ChangeLog: 2016-08-22 Pedro Alves <palves@redhat.com> PR gdb/20505 * gdb.base/vdso-warning.exp: Test core dumps too. Use with_test_prefix. Factor out bits to ... (test_no_vdso): ... this new procedure.
-rw-r--r--gdb/ChangeLog8
-rw-r--r--gdb/linux-tdep.c35
-rw-r--r--gdb/testsuite/ChangeLog7
-rw-r--r--gdb/testsuite/gdb.base/vdso-warning.exp76
4 files changed, 97 insertions, 29 deletions
diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index db3527b..8de59d0 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,11 @@
+2016-08-22 Pedro Alves <palves@redhat.com>
+
+ PR gdb/20505
+ * linux-tdep.c (linux_vsyscall_range_raw): For core inferiors,
+ find the vDSO's start address with AT_SYSINFO_EHDR too, and
+ determine the vDSO's size by finding the PT_LOAD segment that
+ matches AT_SYSINFO_EHDR.
+
2016-08-19 Yao Qi <yao.qi@linaro.org>
* aarch64-tdep.c (aarch64_analyze_prologue): Handle register
diff --git a/gdb/linux-tdep.c b/gdb/linux-tdep.c
index ab110b0..718dc1a 100644
--- a/gdb/linux-tdep.c
+++ b/gdb/linux-tdep.c
@@ -2287,17 +2287,42 @@ linux_vsyscall_range_raw (struct gdbarch *gdbarch, struct mem_range *range)
long pid;
char *data;
- /* Can't access /proc if debugging a core file. */
- if (!target_has_execution)
+ if (target_auxv_search (&current_target, AT_SYSINFO_EHDR, &range->start) <= 0)
return 0;
+ /* It doesn't make sense to access the host's /proc when debugging a
+ core file. Instead, look for the PT_LOAD segment that matches
+ the vDSO. */
+ if (!target_has_execution)
+ {
+ Elf_Internal_Phdr *phdrs;
+ long phdrs_size;
+ int num_phdrs, i;
+
+ phdrs_size = bfd_get_elf_phdr_upper_bound (core_bfd);
+ if (phdrs_size == -1)
+ return 0;
+
+ phdrs = (Elf_Internal_Phdr *) alloca (phdrs_size);
+ num_phdrs = bfd_get_elf_phdrs (core_bfd, phdrs);
+ if (num_phdrs == -1)
+ return 0;
+
+ for (i = 0; i < num_phdrs; i++)
+ if (phdrs[i].p_type == PT_LOAD
+ && phdrs[i].p_vaddr == range->start)
+ {
+ range->length = phdrs[i].p_memsz;
+ return 1;
+ }
+
+ return 0;
+ }
+
/* We need to know the real target PID to access /proc. */
if (current_inferior ()->fake_pid_p)
return 0;
- if (target_auxv_search (&current_target, AT_SYSINFO_EHDR, &range->start) <= 0)
- return 0;
-
pid = current_inferior ()->pid;
/* Note that reading /proc/PID/task/PID/maps (1) is much faster than
diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 9f45c39..7e5fce6 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,10 @@
+2016-08-22 Pedro Alves <palves@redhat.com>
+
+ PR gdb/20505
+ * gdb.base/vdso-warning.exp: Test core dumps too. Use
+ with_test_prefix. Factor out bits to ...
+ (test_no_vdso): ... this new procedure.
+
2016-08-19 Carl Love <cel@us.ibm.com>
* gdb.arch/altivec-regs.exp: Use standard_testfile instead of
diff --git a/gdb/testsuite/gdb.base/vdso-warning.exp b/gdb/testsuite/gdb.base/vdso-warning.exp
index af2b2b0..aeb85a2 100644
--- a/gdb/testsuite/gdb.base/vdso-warning.exp
+++ b/gdb/testsuite/gdb.base/vdso-warning.exp
@@ -13,42 +13,70 @@
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
+# Test that on Linux, we don't warn about not finding the vDSO. E.g.:
+#
+# warning: Could not load shared library symbols for linux-vdso.so.1.
+
standard_testfile
if { [prepare_for_testing "failed to prepare" ${testfile} $srcfile] } {
return -1
}
-gdb_breakpoint "main"
+with_test_prefix "setup" {
+ gdb_breakpoint "main"
-# At least some versions of Fedora/RHEL glibc have local patches that
-# hide the vDSO. This lines re-exposes it. See PR libc/13097,
-# comment 2. There's no support for passing environment variables in
-# the remote protocol, but that's OK -- if we're testing against a
-# glibc that doesn't list the vDSO without this, the test should still
-# pass.
-gdb_test_no_output "set environment LD_DEBUG=unused"
+ # At least some versions of Fedora/RHEL glibc have local patches that
+ # hide the vDSO. This lines re-exposes it. See PR libc/13097,
+ # comment 2. There's no support for passing environment variables in
+ # the remote protocol, but that's OK -- if we're testing against a
+ # glibc that doesn't list the vDSO without this, the test should still
+ # pass.
+ gdb_test_no_output "set environment LD_DEBUG=unused"
+}
-gdb_run_cmd
+proc test_no_vdso {command} {
+ global srcfile
+ global gdb_prompt
-set test "stop without warning"
-gdb_test_multiple "" $test {
- -re "Could not load shared library symbols .*\r\n$gdb_prompt $" {
- fail $test
+ set message "startup"
+ gdb_test_multiple "$command" $message {
+ -re "Could not load shared library symbols .*\r\n$gdb_prompt $" {
+ fail $message
+ }
+ -re "main \\(\\) at .*$srcfile.*\r\n$gdb_prompt $" {
+ pass $message
+ }
}
- -re "\r\nBreakpoint \[0-9\]+, main .*\r\n$gdb_prompt $" {
- pass $test
+
+ # Extra testing in case the warning changes and we miss updating
+ # the above.
+ set test "no vdso without symbols is listed"
+ gdb_test_multiple "info shared" $test {
+ -re "No\[^\r\n\]+linux-(vdso|gate).*$gdb_prompt $" {
+ fail $test
+ }
+ -re "$gdb_prompt $" {
+ pass $test
+ }
}
}
-# Extra testing in case the warning changes and we miss updating the
-# above.
-set test "no vdso without symbols is listed"
-gdb_test_multiple "info shared" $test {
- -re "No\[^\r\n\]+linux-(vdso|gate).*$gdb_prompt $" {
- fail $test
- }
- -re "$gdb_prompt $" {
- pass $test
+# First, try a live process.
+with_test_prefix "run" {
+ gdb_run_cmd
+ test_no_vdso ""
+}
+
+# Now, dump a core, and reload it.
+with_test_prefix "core" {
+ set corefile [standard_output_file $testfile.core]
+ set core_supported [gdb_gcore_cmd "$corefile" "save a corefile"]
+ if {!$core_supported} {
+ return -1
}
+
+ clean_restart ${testfile}
+
+ test_no_vdso "core-file $corefile"
}