| Message ID | 20200122151410.30012-1-simon.marchi@efficios.com |
|---|---|
| State | New |
| Headers | show |
On 2020-01-22 10:14 a.m., Simon Marchi wrote: > displaced_step_inferior_state::reset and displaced_step_clear appear to > have the same goal, but they don't do the same thing. > displaced_step_inferior_state::reset clears more things than > displaced_step_clear, but it misses free'ing the closure, which > displaced_step_clear does. > > This patch replaces displaced_step_clear's implementation with just a call to > displaced_step_inferior_state::reset. It then changes > displaced_step_inferior_state::step_closure to be a unique_ptr, to indicate the > fact that displaced_step_inferior_state owns the closure (and so that it is > automatically freed when the field is reset). > > It should be possible to get rid of displaced_step_clear entirely, but I'm not > sure what the best way, give that it's used in scope exit macros. > > The test gdb.base/step-over-syscall.exp caught a problem when doing this, which > I consider to be a latent bug which my cleanup exposes. In > handle_inferior_event, in the TARGET_WAITKIND_FORKED case, if we displaced-step > over a fork syscall, we make sure to restore the memory that we used as a > displaced-stepping buffer in the child. We do so using the > displaced_step_inferior_state of the parent. However, we do it after calling > displaced_step_fixup for the parent, which clears the information in the > parent's displaced_step_inferior_state. It worked fine before, because > displaced_step_clear didn't completely clear the displaced_step_inferior_state > structure, so the required information (in this case the gdbarch) was > still available after clearing. > > I fixed it by making GDB restore the child's memory before calling the > displaced_step_fixup on the parent. This way, the data in the > displaced_step_inferior_state structure is still valid when we use it for the > child. This is the error you would get in > gdb.base/step-over-syscall.exp without this fix: > > /home/smarchi/src/binutils-gdb/gdb/gdbarch.c:3911: internal-error: ULONGEST gdbarch_max_insn_length(gdbarch*): Assertion `gdbarch != NULL' failed. If there's no objection, I would push these two patches next week. Simon
On 1/22/20 3:14 PM, Simon Marchi wrote: > displaced_step_inferior_state::reset and displaced_step_clear appear to > have the same goal, but they don't do the same thing. > displaced_step_inferior_state::reset clears more things than > displaced_step_clear, but it misses free'ing the closure, which > displaced_step_clear does. > > This patch replaces displaced_step_clear's implementation with just a call to > displaced_step_inferior_state::reset. It then changes > displaced_step_inferior_state::step_closure to be a unique_ptr, to indicate the > fact that displaced_step_inferior_state owns the closure (and so that it is > automatically freed when the field is reset). > > It should be possible to get rid of displaced_step_clear entirely, but I'm not > sure what the best way, give that it's used in scope exit macros. The reason it needs to be wrapped in a cleanup instead of say SCOPE_EXIT, is that it needs to be discardable with the "cleanup.release ();" call. I'm not sure there's a better way. Renaming displaced_step_clear to displaced_step_reset and adding a comment may result in clearer code, though. > diff --git a/gdb/infrun.h b/gdb/infrun.h > index 8040b28f0172..c6329c844d9b 100644 > --- a/gdb/infrun.h > +++ b/gdb/infrun.h > @@ -290,7 +290,7 @@ struct displaced_step_inferior_state > failed_before = 0; > step_thread = nullptr; > step_gdbarch = nullptr; > - step_closure = nullptr; > + step_closure.reset (); I see people sometimes doing this change and I'm curious. Is it for clarity? Anyway, this LGTM. Thanks, Pedro Alves
diff --git a/gdb/infrun.c b/gdb/infrun.c index 9c4a07daba97..1fee65730dbc 100644 --- a/gdb/infrun.c +++ b/gdb/infrun.c @@ -1548,7 +1548,7 @@ get_displaced_step_closure_by_addr (CORE_ADDR addr) /* If checking the mode of displaced instruction in copy area. */ if (displaced->step_thread != nullptr && displaced->step_copy == addr) - return displaced->step_closure; + return displaced->step_closure.get (); return NULL; } @@ -1606,13 +1606,9 @@ use_displaced_stepping (struct thread_info *tp) /* Clean out any stray displaced stepping state. */ static void -displaced_step_clear (struct displaced_step_inferior_state *displaced) +displaced_step_clear (displaced_step_inferior_state *displaced) { - /* Indicate that there is no cleanup pending. */ - displaced->step_thread = nullptr; - - delete displaced->step_closure; - displaced->step_closure = NULL; + displaced->reset (); } /* A cleanup that wraps displaced_step_clear. */ @@ -1762,7 +1758,7 @@ displaced_step_prepare_throw (thread_info *tp) succeeds. */ displaced->step_thread = tp; displaced->step_gdbarch = gdbarch; - displaced->step_closure = closure; + displaced->step_closure.reset (closure); displaced->step_original = original; displaced->step_copy = copy; @@ -1887,7 +1883,7 @@ displaced_step_fixup (thread_info *event_thread, enum gdb_signal signal) { /* Fix up the resulting state. */ gdbarch_displaced_step_fixup (displaced->step_gdbarch, - displaced->step_closure, + displaced->step_closure.get (), displaced->step_original, displaced->step_copy, get_thread_regcache (displaced->step_thread)); @@ -2480,8 +2476,8 @@ resume_1 (enum gdb_signal sig) pc = regcache_read_pc (get_thread_regcache (tp)); displaced = get_displaced_stepping_state (tp->inf); - step = gdbarch_displaced_step_hw_singlestep (gdbarch, - displaced->step_closure); + step = gdbarch_displaced_step_hw_singlestep + (gdbarch, displaced->step_closure.get ()); } } @@ -5313,6 +5309,15 @@ Cannot fill $_exitsignal with the correct signal number.\n")); struct regcache *child_regcache; CORE_ADDR parent_pc; + if (ecs->ws.kind == TARGET_WAITKIND_FORKED) + { + struct displaced_step_inferior_state *displaced + = get_displaced_stepping_state (parent_inf); + + /* Restore scratch pad for child process. */ + displaced_step_restore (displaced, ecs->ws.value.related_pid); + } + /* GDB has got TARGET_WAITKIND_FORKED or TARGET_WAITKIND_VFORKED, indicating that the displaced stepping of syscall instruction has been done. Perform cleanup for parent process here. Note @@ -5323,15 +5328,6 @@ Cannot fill $_exitsignal with the correct signal number.\n")); that needs it. */ start_step_over (); - if (ecs->ws.kind == TARGET_WAITKIND_FORKED) - { - struct displaced_step_inferior_state *displaced - = get_displaced_stepping_state (parent_inf); - - /* Restore scratch pad for child process. */ - displaced_step_restore (displaced, ecs->ws.value.related_pid); - } - /* Since the vfork/fork syscall instruction was executed in the scratchpad, the child's PC is also within the scratchpad. Set the child's PC to the parent's PC value, which has already been fixed up. diff --git a/gdb/infrun.h b/gdb/infrun.h index 8040b28f0172..c6329c844d9b 100644 --- a/gdb/infrun.h +++ b/gdb/infrun.h @@ -290,7 +290,7 @@ struct displaced_step_inferior_state failed_before = 0; step_thread = nullptr; step_gdbarch = nullptr; - step_closure = nullptr; + step_closure.reset (); step_original = 0; step_copy = 0; step_saved_copy.clear (); @@ -310,7 +310,7 @@ struct displaced_step_inferior_state /* The closure provided gdbarch_displaced_step_copy_insn, to be used for post-step cleanup. */ - displaced_step_closure *step_closure; + std::unique_ptr<displaced_step_closure> step_closure; /* The address of the original instruction, and the copy we made. */
displaced_step_inferior_state::reset and displaced_step_clear appear to have the same goal, but they don't do the same thing. displaced_step_inferior_state::reset clears more things than displaced_step_clear, but it misses free'ing the closure, which displaced_step_clear does. This patch replaces displaced_step_clear's implementation with just a call to displaced_step_inferior_state::reset. It then changes displaced_step_inferior_state::step_closure to be a unique_ptr, to indicate the fact that displaced_step_inferior_state owns the closure (and so that it is automatically freed when the field is reset). It should be possible to get rid of displaced_step_clear entirely, but I'm not sure what the best way, give that it's used in scope exit macros. The test gdb.base/step-over-syscall.exp caught a problem when doing this, which I consider to be a latent bug which my cleanup exposes. In handle_inferior_event, in the TARGET_WAITKIND_FORKED case, if we displaced-step over a fork syscall, we make sure to restore the memory that we used as a displaced-stepping buffer in the child. We do so using the displaced_step_inferior_state of the parent. However, we do it after calling displaced_step_fixup for the parent, which clears the information in the parent's displaced_step_inferior_state. It worked fine before, because displaced_step_clear didn't completely clear the displaced_step_inferior_state structure, so the required information (in this case the gdbarch) was still available after clearing. I fixed it by making GDB restore the child's memory before calling the displaced_step_fixup on the parent. This way, the data in the displaced_step_inferior_state structure is still valid when we use it for the child. This is the error you would get in gdb.base/step-over-syscall.exp without this fix: /home/smarchi/src/binutils-gdb/gdb/gdbarch.c:3911: internal-error: ULONGEST gdbarch_max_insn_length(gdbarch*): Assertion `gdbarch != NULL' failed. gdb/ChangeLog: * infrun.c (get_displaced_step_closure_by_addr): Adjust to std::unique_ptr. (displaced_step_clear): Just call displaced->reset (). (displaced_step_prepare_throw): Adjust to std::unique_ptr. (displaced_step_fixup): Likewise. (resume_1): Likewise. (handle_inferior_event): Restore child's memory before calling displaced_step_fixup on the parent. * infrun.h (displaced_step_inferior_state) <reset>: Adjust to std::unique_ptr. <step_closure>: Change type to std::unique_ptr. --- gdb/infrun.c | 36 ++++++++++++++++-------------------- gdb/infrun.h | 4 ++-- 2 files changed, 18 insertions(+), 22 deletions(-)