[Bug gdb/27889] New: jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[Bug gdb/27889] New: jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

            Bug ID: 27889
           Summary: jit.c:1178: internal-error: void
                    jit_event_handler(gdbarch*, objfile*): Assertion
                    `jiter->jiter_data != nullptr' failed.
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: gdb
          Assignee: unassigned at sourceware dot org
          Reporter: vries at gcc dot gnu.org
  Target Milestone: ---

Take example from https://llvm.org/docs/DebuggingJITedCode.html, save as
test.c. [ Probably works with any example, just listing what I used. ]

Verify compilation and error code:
...
$ gcc test.c; ./a.out; echo $?; rm a.out
255
...

Compile into llvm byte code:
...
$ clang -g -S -emit-llvm --target=x86_64-unknown-unknown-elf test.c
...

Verify that byte-code works:
...
$ lli test.ll; echo $?
255
...

Now run under gdb:
...
$ gdb -q -batch -ex run --args /usr/bin/lli test.ll
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
[Inferior 1 (process 23272) exited with code 0377]
...

Ok, that went fine.  Using -ex "set debug infrun 1" we can also verify that
handle_jit_event was called:
...
  [infrun] handle_jit_event: handling bp_jit_event
...

Now install debuginfo for libllvm:
...
$ sudo zypper in libLLVM10-debuginfo
...

and try again:
...
$ gdb -q -batch -ex run --args /usr/bin/lli test.ll
Dwarf Error: Cannot not find DIE at 0x18a936e7 [from module
/usr/lib/debug/usr/lib64/libLLVM.so.10-10.0.1-lp152.30.4.x86_64.debug]

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
/home/vries/gdb_versions/devel/src/gdb/jit.c:1178: internal-error: void
jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr'
failed.
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #1 from Tom de Vries <vries at gcc dot gnu.org> ---
At the assert site, we have:
...
(gdb) 
#6  0x00000000007c454a in jit_event_handler (gdbarch=0x39cd880,
jiter=0x22fe700)
    at /home/vries/gdb_versions/devel/src/gdb/jit.c:1178
1178      gdb_assert (jiter->jiter_data != nullptr);
(gdb) p jiter
$1 = (objfile *) 0x22fe700
(gdb) p *jiter
$2 = {original_name = 0x2497fe0 "/usr/bin/../lib64/libLLVM.so.10", addr_low =
140737277312592, 
...

and at the point where we set jiter_data, we have:
...
181         objf->jiter_data.reset (new jiter_objfile_data ());
(gdb) p *objf
$5 = {
  original_name = 0x38e5c80
"/usr/lib/debug/usr/lib64/libLLVM.so.10-10.0.1-lp152.30.4.x86_64.debug"
...

So it's another on of those: exec vs exec.debug.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #2 from Tom de Vries <vries at gcc dot gnu.org> ---
Workaround: use eu-unstrip:
...
$ pwd
/home/vries/try
$ eu-unstrip /usr/lib64/libLLVM.so.10
/usr/lib/debug/usr/lib64/libLLVM.so.10-10.0.1-lp152.30.4.x86_64.debug -o
libLLVM.so.10
$ gdb -q -batch -ex "set env LD_LIBRARY_PATH = /home/vries/try" -ex run --args
/usr/bin/lli ~/test.ll
Dwarf Error: Cannot not find DIE at 0x18a936e7 [from module
/home/vries/try/libLLVM.so.10]

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
[Inferior 1 (process 24428) exited with code 0377]
$
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[pafee at tycoint dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

Paul Fee <pafee at tycoint dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pafee at tycoint dot com

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #3 from Tom de Vries <vries at gcc dot gnu.org> ---
Tentative patch:
...
$ git diff
diff --git a/gdb/jit.c b/gdb/jit.c
index 296b436c796..00097881d83 100644
--- a/gdb/jit.c
+++ b/gdb/jit.c
@@ -833,7 +833,7 @@ jit_breakpoint_re_set_internal (struct gdbarch *gdbarch,
program_s
pace *pspace)
        }

       jiter_objfile_data *objf_data
-       = get_jiter_objfile_data (reg_symbol.objfile);
+       = get_jiter_objfile_data (the_objfile);
       objf_data->register_code = reg_symbol.minsym;
       objf_data->descriptor = desc_symbol.minsym;

...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|gdb                         |breakpoints

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #4 from Tom de Vries <vries at gcc dot gnu.org> ---
Patch now testing:
...
diff --git a/gdb/jit.c b/gdb/jit.c
index 296b436c796..be10f197fd6 100644
--- a/gdb/jit.c
+++ b/gdb/jit.c
@@ -807,6 +807,10 @@ jit_breakpoint_re_set_internal (struct gdbarch *gdbarch,
program_
space *pspace)
 {
   for (objfile *the_objfile : pspace->objfiles ())
     {
+      /* Skip separate debug objects.  */
+      if (the_objfile->separate_debug_objfile_backlink != nullptr)
+       continue;
+
       if (the_objfile->skip_jit_symbol_lookup)
        continue;

@@ -833,7 +837,7 @@ jit_breakpoint_re_set_internal (struct gdbarch *gdbarch,
program_s
pace *pspace)
        }

       jiter_objfile_data *objf_data
-       = get_jiter_objfile_data (reg_symbol.objfile);
+       = get_jiter_objfile_data (the_objfile);
       objf_data->register_code = reg_symbol.minsym;
       objf_data->descriptor = desc_symbol.minsym;

...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #5 from Tom de Vries <vries at gcc dot gnu.org> ---
[ FTR, originally reported at https://bugzilla.suse.com/show_bug.cgi?id=1186040
. ]

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #6 from Tom de Vries <vries at gcc dot gnu.org> ---
https://sourceware.org/pipermail/gdb-patches/2021-May/179072.html

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[cvs-commit at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #7 from cvs-commit at gcc dot gnu.org <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Tom de Vries <vries@sourceware.org>:

https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=6d1a09b77a9ab52d9fc627b8f5eb952892c81f90

commit 6d1a09b77a9ab52d9fc627b8f5eb952892c81f90
Author: Tom de Vries <tdevries@suse.de>
Date:   Fri May 21 15:09:14 2021 +0200

    [gdb/breakpoint] Fix assert in jit_event_handler

    Consider a minimal test-case test.c:
    ...
    int main (void) { return 0; }
    ...
    which we can compile into llvm byte code using clang:
    ...
    $ clang -g -S -emit-llvm --target=x86_64-unknown-unknown-elf test.c
    ...
    and then run using lli, which uses the llvm jit:
    ...
    $ lli test.ll
    ...

    If we run this under gdb, we run into an assert:
    ...
    $ gdb -q -batch -ex run --args /usr/bin/lli test.ll
    Dwarf Error: Cannot not find DIE at 0x18a936e7 \
      [from module libLLVM.so.10-10.0.1-lp152.30.4.x86_64.debug]

    [Thread debugging using libthread_db enabled]
    Using host libthread_db library "/lib64/libthread_db.so.1".
    src/gdb/jit.c:1178: internal-error: \
      void jit_event_handler(gdbarch*, objfile*): \
      Assertion `jiter->jiter_data != nullptr' failed.
    ...

    This is caused by the following.

    When running jit_breakpoint_re_set_internal, we first handle
    libLLVM.so.10.debug, and set a jit breakpoint.

    Next we handle libLLVM.so.10:
    ...
    (gdb) p the_objfile.original_name
    $42 = 0x2494170 "libLLVM.so.10"
    ...
    but the minimal symbols we find are from libLLVM.so.10.debug:
    ...
    (gdb) p reg_symbol.objfile.original_name
    $43 = 0x38e7c50 "libLLVM.so.10-10.0.1-lp152.30.4.x86_64.debug"
    (gdb) p desc_symbol.objfile.original_name
    $44 = 0x38e7c50 "libLLVM.so.10-10.0.1-lp152.30.4.x86_64.debug"
    ...
    and consequently, the objf_data is the one from libLLVM.so.10.debug:
    ...
          jiter_objfile_data *objf_data
            = get_jiter_objfile_data (reg_symbol.objfile);
    ...
    and so we hit this:
    ...
          if (objf_data->cached_code_address == addr)
            continue;
    ...
    and no second jit breakpoint is inserted.

    Subsequently, the jit breakpoint is triggered and handled, but when finding
    the symbol for the breakpoint address we get:
    ...
    (gdb) p jit_bp_sym.objfile.original_name
    $52 = 0x2494170 "libLLVM.so.10"
    ...

    The assert 'jiter->jiter_data != nullptr' triggers because it checks
    libLLVM.so.10 while the one with jiter_data setup is libLLVM.so.10.debug.

    This fixes the assert:
    ...
           jiter_objfile_data *objf_data
    -        = get_jiter_objfile_data (reg_symbol.objfile);
    -        = get_jiter_objfile_data (the_objfile);
    ...
    but consequently we'll have two jit breakpoints, so we also make sure we
don't
    set a jit breakpoint on separate debug objects like libLLVM.so.10.debug.

    Tested on x86_64-linux.

    gdb/ChangeLog:

    2021-05-21  Tom de Vries  <tdevries@suse.de>

            PR breakpoint/27889
            * jit.c (jit_breakpoint_re_set_internal): Skip separate debug
            objects.  Call get_jiter_objfile_data with the_objfile.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|NEW                         |RESOLVED
   Target Milestone|---                         |11.1

--- Comment #8 from Tom de Vries <vries at gcc dot gnu.org> ---
Patch committed, marked resolved-fixed.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[hector.oron at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

HectorOron <hector.oron at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hector.oron at gmail dot com

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[hector.oron at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #9 from HectorOron <hector.oron at gmail dot com> ---
I am hitting this issue with 11.2 and 12.1, I tried this patch and fixes the
issue for me.

```
--- gdb-12.1.orig/gdb/breakpoint.c
+++ gdb-12.1/gdb/breakpoint.c
@@ -5549,7 +5549,10 @@ handle_jit_event (CORE_ADDR address)
      function needs to be updated too.  */
   bound_minimal_symbol jit_bp_sym = lookup_minimal_symbol_by_pc (address);
   gdb_assert (jit_bp_sym.objfile != nullptr);
-  jit_event_handler (gdbarch, jit_bp_sym.objfile);
+  if (jit_bp_sym.objfile->separate_debug_objfile_backlink)
+    jit_event_handler (gdbarch,
jit_bp_sym.objfile->separate_debug_objfile_backlink);
+  else
+    jit_event_handler (gdbarch, jit_bp_sym.objfile);

   target_terminal::inferior ();
 }

```

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[guillaume at morinfr dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

Guillaume Morin <guillaume at morinfr dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |guillaume at morinfr dot org

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #10 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to HectorOron from comment #9)
> I am hitting this issue with 11.2 and 12.1

Do you mean you hit the same assert?

If so, please file a PR with a reproducer.

As for the issue described in this PR, I tried to reproduce it on trunk, and
found that it's still fixed.  I also verified that I could still trigger the
assert by reverting the fixing patch.

So, this PR stays resolved-fixed, with target milestone 11.1.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug breakpoints/27889] jit.c:1178: internal-error: void jit_event_handler(gdbarch*, objfile*): Assertion `jiter->jiter_data != nullptr' failed.

[guillaume at morinfr dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27889

--- Comment #11 from Guillaume Morin <guillaume at morinfr dot org> ---
Hi Tom, as you guessed, yes we've hit the same assert but this is not the same
bug. We've opened a new PR:
https://sourceware.org/bugzilla/show_bug.cgi?id=29277

-- 
You are receiving this mail because:
You are on the CC list for the bug.