| Message ID | 20200221210628.24922-1-luis.machado@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | Harden gdb.arch/aarch64-pauth.exp and fix a failure | expand |
On Fri, 21 Feb 2020 18:06:28 -0300 Luis Machado <luis.machado@linaro.org> wrote: > When running this testcase against a QEMU with PAC support, i noticed we > were failing to recognize the additional [PAC] that is emitted in the > backtrace, resulting in this failure: > > FAIL: gdb.arch/aarch64-pauth.exp: backtrace > > I've made the test use multi_line to make the pattern more clear. > > Tested against aarch64-linux-gnu with and without PAC support. > > gdb/testsuite/ChangeLog: > > 2020-02-21 Luis Machado <luis.machado@linaro.org> > > * gdb.arch/aarch64-pauth.exp: Recognize optional PAC output. > > Signed-off-by: Luis Machado <luis.machado@linaro.org> > --- > gdb/testsuite/gdb.arch/aarch64-pauth.exp | 7 ++++++- > 1 file changed, 6 insertions(+), 1 deletion(-) > > diff --git a/gdb/testsuite/gdb.arch/aarch64-pauth.exp b/gdb/testsuite/gdb.arch/aarch64-pauth.exp > index 816e58be44..e6d25c5d97 100644 > --- a/gdb/testsuite/gdb.arch/aarch64-pauth.exp > +++ b/gdb/testsuite/gdb.arch/aarch64-pauth.exp > @@ -40,4 +40,9 @@ gdb_breakpoint [ gdb_get_line_number "break here" ] > gdb_continue_to_breakpoint "break here" ".*break here.*" > > # Ensure we can get a full backtrace, despite the address signing. > -gdb_test "bt" "^bt\r\n#0 +bar *\\(b=9\\) +at.*\r\n#1 +0x\[0-9a-f\]* +in +foo \\(a=5\\).*\r\n#2 +0x\[0-9a-f\]* +in +main \\(\\).*" "backtrace" > +gdb_test "bt" \ > + [multi_line \ > + "#0\[ \t\]*bar \\(b=9\\) at \[^\r\n\]+" \ > + "#1\[ \t\]*$hex (\\\[PAC\\\] )?in foo \\(a=5\\) at \[^\r\n\]+" \ > + "#2\[ \t\]*$hex (\\\[PAC\\\] )?in main \\(\\) at \[^\r\n\]+" ] \ > + "backtrace" LGTM except for the \[ \t\]* parts of it. I think we actually want + instead of * to ensure that there's whitespace separating the frame number from the hex value. Kevin
On 2/22/20 1:20 AM, Kevin Buettner wrote: > On Fri, 21 Feb 2020 18:06:28 -0300 > Luis Machado <luis.machado@linaro.org> wrote: > >> When running this testcase against a QEMU with PAC support, i noticed we >> were failing to recognize the additional [PAC] that is emitted in the >> backtrace, resulting in this failure: >> >> FAIL: gdb.arch/aarch64-pauth.exp: backtrace >> >> I've made the test use multi_line to make the pattern more clear. >> >> Tested against aarch64-linux-gnu with and without PAC support. >> >> gdb/testsuite/ChangeLog: >> >> 2020-02-21 Luis Machado <luis.machado@linaro.org> >> >> * gdb.arch/aarch64-pauth.exp: Recognize optional PAC output. >> >> Signed-off-by: Luis Machado <luis.machado@linaro.org> >> --- >> gdb/testsuite/gdb.arch/aarch64-pauth.exp | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/gdb/testsuite/gdb.arch/aarch64-pauth.exp b/gdb/testsuite/gdb.arch/aarch64-pauth.exp >> index 816e58be44..e6d25c5d97 100644 >> --- a/gdb/testsuite/gdb.arch/aarch64-pauth.exp >> +++ b/gdb/testsuite/gdb.arch/aarch64-pauth.exp >> @@ -40,4 +40,9 @@ gdb_breakpoint [ gdb_get_line_number "break here" ] >> gdb_continue_to_breakpoint "break here" ".*break here.*" >> >> # Ensure we can get a full backtrace, despite the address signing. >> -gdb_test "bt" "^bt\r\n#0 +bar *\\(b=9\\) +at.*\r\n#1 +0x\[0-9a-f\]* +in +foo \\(a=5\\).*\r\n#2 +0x\[0-9a-f\]* +in +main \\(\\).*" "backtrace" >> +gdb_test "bt" \ >> + [multi_line \ >> + "#0\[ \t\]*bar \\(b=9\\) at \[^\r\n\]+" \ >> + "#1\[ \t\]*$hex (\\\[PAC\\\] )?in foo \\(a=5\\) at \[^\r\n\]+" \ >> + "#2\[ \t\]*$hex (\\\[PAC\\\] )?in main \\(\\) at \[^\r\n\]+" ] \ >> + "backtrace" > > LGTM except for the \[ \t\]* parts of it. I think we actually want + > instead of * to ensure that there's whitespace separating the frame > number from the hex value. > > Kevin > Indeed. Fixed now. Thanks for the review.
On 2/22/20 1:20 AM, Kevin Buettner wrote: > On Fri, 21 Feb 2020 18:06:28 -0300 > Luis Machado <luis.machado@linaro.org> wrote: > >> When running this testcase against a QEMU with PAC support, i noticed we >> were failing to recognize the additional [PAC] that is emitted in the >> backtrace, resulting in this failure: >> >> FAIL: gdb.arch/aarch64-pauth.exp: backtrace >> >> I've made the test use multi_line to make the pattern more clear. >> >> Tested against aarch64-linux-gnu with and without PAC support. >> >> gdb/testsuite/ChangeLog: >> >> 2020-02-21 Luis Machado <luis.machado@linaro.org> >> >> * gdb.arch/aarch64-pauth.exp: Recognize optional PAC output. >> >> Signed-off-by: Luis Machado <luis.machado@linaro.org> >> --- >> gdb/testsuite/gdb.arch/aarch64-pauth.exp | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/gdb/testsuite/gdb.arch/aarch64-pauth.exp b/gdb/testsuite/gdb.arch/aarch64-pauth.exp >> index 816e58be44..e6d25c5d97 100644 >> --- a/gdb/testsuite/gdb.arch/aarch64-pauth.exp >> +++ b/gdb/testsuite/gdb.arch/aarch64-pauth.exp >> @@ -40,4 +40,9 @@ gdb_breakpoint [ gdb_get_line_number "break here" ] >> gdb_continue_to_breakpoint "break here" ".*break here.*" >> >> # Ensure we can get a full backtrace, despite the address signing. >> -gdb_test "bt" "^bt\r\n#0 +bar *\\(b=9\\) +at.*\r\n#1 +0x\[0-9a-f\]* +in +foo \\(a=5\\).*\r\n#2 +0x\[0-9a-f\]* +in +main \\(\\).*" "backtrace" >> +gdb_test "bt" \ >> + [multi_line \ >> + "#0\[ \t\]*bar \\(b=9\\) at \[^\r\n\]+" \ >> + "#1\[ \t\]*$hex (\\\[PAC\\\] )?in foo \\(a=5\\) at \[^\r\n\]+" \ >> + "#2\[ \t\]*$hex (\\\[PAC\\\] )?in main \\(\\) at \[^\r\n\]+" ] \ >> + "backtrace" > > LGTM except for the \[ \t\]* parts of it. I think we actually want + > instead of * to ensure that there's whitespace separating the frame > number from the hex value. > > Kevin > Fixed and pushed. Thanks! I'll also push a fix to gdb.base/backtrace.exp, since it also uses * instead of + for the following tests: # Backtrace with the default options. gdb_test "bt" \ [multi_line \ "#0\[ \t\]*baz \\(\\) at \[^\r\n\]+" \ "#1\[ \t\]*$hex in bar \\(\\) at \[^\r\n\]+" \ "#2\[ \t\]*$hex in foo \\(\\) at \[^\r\n\]+" \ "#3\[ \t\]*$hex in main \\(\\) at \[^\r\n\]+" ] # Backtrace with 'set disassemble-next-line on'. This shouldn't make # any difference to the backtrace. gdb_test "with disassemble-next-line on -- bt" \ [multi_line \ "#0\[ \t\]*baz \\(\\) at \[^\r\n\]+" \ "#1\[ \t\]*$hex in bar \\(\\) at \[^\r\n\]+" \ "#2\[ \t\]*$hex in foo \\(\\) at \[^\r\n\]+" \ "#3\[ \t\]*$hex in main \\(\\) at \[^\r\n\]+" ]
diff --git a/gdb/testsuite/gdb.arch/aarch64-pauth.exp b/gdb/testsuite/gdb.arch/aarch64-pauth.exp index 816e58be44..e6d25c5d97 100644 --- a/gdb/testsuite/gdb.arch/aarch64-pauth.exp +++ b/gdb/testsuite/gdb.arch/aarch64-pauth.exp @@ -40,4 +40,9 @@ gdb_breakpoint [ gdb_get_line_number "break here" ] gdb_continue_to_breakpoint "break here" ".*break here.*" # Ensure we can get a full backtrace, despite the address signing. -gdb_test "bt" "^bt\r\n#0 +bar *\\(b=9\\) +at.*\r\n#1 +0x\[0-9a-f\]* +in +foo \\(a=5\\).*\r\n#2 +0x\[0-9a-f\]* +in +main \\(\\).*" "backtrace" +gdb_test "bt" \ + [multi_line \ + "#0\[ \t\]*bar \\(b=9\\) at \[^\r\n\]+" \ + "#1\[ \t\]*$hex (\\\[PAC\\\] )?in foo \\(a=5\\) at \[^\r\n\]+" \ + "#2\[ \t\]*$hex (\\\[PAC\\\] )?in main \\(\\) at \[^\r\n\]+" ] \ + "backtrace"
When running this testcase against a QEMU with PAC support, i noticed we were failing to recognize the additional [PAC] that is emitted in the backtrace, resulting in this failure: FAIL: gdb.arch/aarch64-pauth.exp: backtrace I've made the test use multi_line to make the pattern more clear. Tested against aarch64-linux-gnu with and without PAC support. gdb/testsuite/ChangeLog: 2020-02-21 Luis Machado <luis.machado@linaro.org> * gdb.arch/aarch64-pauth.exp: Recognize optional PAC output. Signed-off-by: Luis Machado <luis.machado@linaro.org> --- gdb/testsuite/gdb.arch/aarch64-pauth.exp | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) -- 2.17.1