Opened 14 years ago
Closed 13 years ago
#7235 closed bug (fixed)
JamVM thread continue to wait for signal when sent
| Reported by: | korli | Owned by: | nobody |
|---|---|---|---|
| Priority: | normal | Milestone: | R1 |
| Component: | System/POSIX | Version: | R1/Development |
| Keywords: | Cc: | ||
| Blocked By: | #5679 | Blocking: | |
| Platform: | All |
Description
When compiling Classpath, JamVM running Eclipse java compiler uses multiple threads, which are suspended and resumed by signals.
I tried to recreate the problem with src/tests/system/libroot/posix/pthread_signal_test.cpp, I hope I got it right. The test name is pthread_signal_test.
Attachments (1)
Change History (8)
comment:1 by , 14 years ago
| Blocked By: | 5679 added |
|---|---|
| Resolution: | → duplicate |
| Status: | new → closed |
comment:2 by , 14 years ago
| Resolution: | duplicate |
|---|---|
| Status: | closed → reopened |
#5679 is indeed a problem but not the only one met with JamVM. pthread_kill send signals to a thread which doesn't receive it, it's not killed either. send_signal() is actually called with a thread id 0.
To init the main thread id when calling pthread_self(), we have to init sMainThread correctly.
by , 14 years ago
| Attachment: | pthread.diff added |
|---|
init thread id so that it can be used by pthread_kill()
comment:3 by , 14 years ago
| patch: | 0 → 1 |
|---|
comment:4 by , 14 years ago
Looks OK. Apparently freeing the pthread_thread structure on error is missing in spawn_thread().
comment:5 by , 14 years ago
I applied with the structure freeing in hrev40665. I'll see if it's enough to close this report.
comment:7 by , 13 years ago
| Resolution: | → fixed |
|---|---|
| Status: | reopened → closed |
Assuming fixed. Please reopen, if not.
Looks like a duplicate of #5679.