Messages in this thread | | | Date | Tue, 07 Sep 1999 18:20:53 -0400 | From | "Jonathan A. George" <> | Subject | Re: > 15,000 Simultaneous Connections -- This is Great! ... Who's going to implement |
| |
> Mike Jagdis wrote: > > On Tue, 7 Sep 1999, Rogier Wolff wrote: > > > > > I would then prefer to go with the API that the digital guys suggested > > > in: http://www.cs.rice.edu/~gaurav/papers/usenix99.ps > > > > > > #define EVENT_READ 0x01 > > > #define EVENT_WRITE 0x02 > > > #define EVENT_EXCEPT 0x03 > > Note the typo: That should be 4. > > > > typedef struct { > > > int fd; > > > unsigned mask; > > > } event_descr_t; > > > > > > int declare_interest (int fd, int interestmask, int *statemask); > > > int get_next_event (int array_max, event_descr_t *ev_array, struct timeval *timeout); > > > > This is basically what I was thinking. However I think we want > > to use the existing POLL* flags rather than these EVENT* values. > > Fine with me. > > I recommend going with a compatibility define to set the EVENT_ > variables to get a source-level compatibility with those other guys > API. > > Even if there is just one other implementation, compatiblity wiht just > one other is to be preferred to something unique to Linux. > > Roger. >
1. I hope that someone follows through on the BSD like poll() optimization.
2. This declare/event API is one of the coolest things I've seen in years! I've always found select() is quite kludgy and poll() seemed inefficient. However, I'm not really sure I like the API label (declare_fd() & get_next_fd() would be my suggestion).
This is so cool I wouldn't mind implementing it myself. Does anyone else think they can get a 2.3.17 based patch ready by the end of next week? ;)
We _should_ be able to match the performance of Solaris 7 and AIX 4.3.3 even without this new API. With this API Linux will have the _potential_ for unbeatable network performance!
-- Jonathan A. George --- [Adjunct MIS Professor, Unix Developer, Linux Supporter]
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |