| 1 |
On 10/07/13 01:11, Greg Turner wrote: |
| 2 |
> Alan wrote: |
| 3 |
>> - redirect_alloc_fd mj_control_fd "${pipe}" |
| 4 |
>> + redirect_alloc_fd mj_write_fd "${pipe}" |
| 5 |
>> + redirect_alloc_fd mj_read_fd "${pipe}" |
| 6 |
> The one thing I'd be looking to verify is that there's no race where |
| 7 |
> where the reader must block before the writer starts producing data. |
| 8 |
> |
| 9 |
> You could test that hypothesis by sleeping for a while right before |
| 10 |
> the reader starts listening. |
| 11 |
> |
| 12 |
> If that were the case -- I kinda suspect it is -- then unfortunately, |
| 13 |
> I think the writer has to spin, waiting for the reader to block... not |
| 14 |
> a pretty situation, so let's hope I'm confused. |
| 15 |
> |
| 16 |
|
| 17 |
I don't understand. There is no race, the above two commands just get |
| 18 |
two different file descriptors for the read end and the write end. The |
| 19 |
FIFO is opened in non-blocking IO mode anyway, as the FIFO is still |
| 20 |
opened in read/write mode to get this behaviour. So the actual read from |
| 21 |
one fd, and the write from another fd behaves no differently to reading |
| 22 |
and writing from the same fd. |
| 23 |
|
| 24 |
Alan. |
Archives