Busy loop in IoLib
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
IOLib |
New
|
Undecided
|
Unassigned |
Bug Description
Using the DBus library (the experimental "server" branch) I was able to trigger a busy loop in IOLIB:
[pid 17998] 14:14:57.805234 epoll_wait(9, [{EPOLLIN|EPOLLHUP, {u32=10, u64=10}}], 1024, -1) = 1
[pid 17998] 14:14:57.805282 recvfrom(10, "", 4096, 0, NULL, NULL) = 0
[pid 17998] 14:14:57.805332 epoll_wait(9, [{EPOLLIN|EPOLLHUP, {u32=10, u64=10}}], 1024, -1) = 1
[pid 17998] 14:14:57.805381 recvfrom(10, "", 4096, 0, NULL, NULL) = 0
[pid 17998] 14:14:57.805431 epoll_wait(9, [{EPOLLIN|EPOLLHUP, {u32=10, u64=10}}], 1024, -1) = 1
[pid 17998] 14:14:57.805505 recvfrom(10, "", 4096, 0, NULL, NULL) = 0
[pid 17998] 14:14:57.805556 epoll_wait(9, [{EPOLLIN|EPOLLHUP, {u32=10, u64=10}}], 1024, -1) = 1
Swank shows this stack - nearly the same one on a few different tries.
0: ("bogus stack frame")
1: ((:METHOD IOLIB/MULTIPLEX
2: (IOLIB/
3: ((:METHOD IOLIB/MULTIPLEX
4: ((FLET CALL-NEXT-METHOD :IN "/home/
5: ((:METHOD IOLIB/MULTIPLEX
6: ((:METHOD WAIT-FOR-REPLY (T STANDARD-
Another try had
0: ((LAMBDA (SB-KERNEL:
1: ((:METHOD IOLIB/MULTIPLEX
etc.
This is with iolib-v0.8.1 from QL, BTW.