public inbox for cygwin-cvs@sourceware.org
help / color / mirror / Atom feed
* [newlib-cygwin] Cygwin: Remove waitloop argument from try_to_debug()
@ 2020-08-30 15:27 Jon TURNEY
0 siblings, 0 replies; only message in thread
From: Jon TURNEY @ 2020-08-30 15:27 UTC (permalink / raw)
To: cygwin-cvs
https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin.git;h=a30cd7a5b95675adf31b6092235b6c1f42c59520
commit a30cd7a5b95675adf31b6092235b6c1f42c59520
Author: Jon Turney <jon.turney@dronecode.org.uk>
Date: Thu Aug 27 16:12:47 2020 +0100
Cygwin: Remove waitloop argument from try_to_debug()
Currently, when using CYGWIN's error_start facility, the faulting
process isn't stopped while the error_start process is started when the
fault is caused by an exception. (it even seems possible in theory that
the faulting process could have exited before the error_start process
attaches).
This leads to e.g. the core dump written by CYGWIN='error_start=dumper'
in response to an exception being non-deterministic.
Remove the waitloop argument from try_to_debug(), only used in the
exception case, so the faulting process busy-waits until the error_start
process attaches.
Code archaeology to determine why the code is this way didn't really
turn up any answers, but this seems a low-risk change, as this only
changes the behaviour when:
- a debugger isn't already attached
- an error_start is specified in CYGWIN env var
- an exception has occurred which will be translated to a signal
If error_start invokes something which doesn't attach using
DebugActiveProcess(), we will spin indefinitely, but that will also
currently occur for any of the existing other uses of try_to_debug(),
which default to waitloop=TRUE.
Diff:
---
winsup/cygwin/exceptions.cc | 14 ++++++++------
winsup/cygwin/winsup.h | 2 +-
2 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/winsup/cygwin/exceptions.cc b/winsup/cygwin/exceptions.cc
index bb7704f94..c98b92d30 100644
--- a/winsup/cygwin/exceptions.cc
+++ b/winsup/cygwin/exceptions.cc
@@ -461,10 +461,8 @@ cygwin_stackdump ()
exc.dumpstack ();
}
-#define TIME_TO_WAIT_FOR_DEBUGGER 10000
-
extern "C" int
-try_to_debug (bool waitloop)
+try_to_debug ()
{
if (!debugger_command)
return 0;
@@ -532,13 +530,17 @@ try_to_debug (bool waitloop)
&si,
&pi);
+ /* We want to stop here and wait until the error_start process attaches. But
+ we can't wait here for the error_start process to exit, as if it's a
+ debugger, it might want to continue this thread. So we busy wait until a
+ debugger attaches, which stops this process, after which it can decide if
+ we continue or not. */
+
*dbg_end = L'\0';
if (!dbg)
system_printf ("Failed to start debugger, %E");
else
{
- if (!waitloop)
- return dbg;
SetThreadPriority (GetCurrentThread (), THREAD_PRIORITY_IDLE);
while (!being_debugged ())
Sleep (1);
@@ -812,7 +814,7 @@ exception::handle (EXCEPTION_RECORD *e, exception_list *frame, CONTEXT *in,
if (exit_state >= ES_SIGNAL_EXIT
&& (NTSTATUS) e->ExceptionCode != STATUS_CONTROL_C_EXIT)
api_fatal ("Exception during process exit");
- else if (!try_to_debug (0))
+ else if (!try_to_debug ())
rtl_unwind (frame, e);
else
{
diff --git a/winsup/cygwin/winsup.h b/winsup/cygwin/winsup.h
index 79844cb87..0ffd8c5af 100644
--- a/winsup/cygwin/winsup.h
+++ b/winsup/cygwin/winsup.h
@@ -190,7 +190,7 @@ void close_all_files (bool = false);
/* debug_on_trap support. see exceptions.cc:try_to_debug() */
extern "C" void error_start_init (const char*);
-extern "C" int try_to_debug (bool waitloop = 1);
+extern "C" int try_to_debug ();
void ld_preload ();
void fixup_hooks_after_fork ();
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2020-08-30 15:27 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-08-30 15:27 [newlib-cygwin] Cygwin: Remove waitloop argument from try_to_debug() Jon TURNEY
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).