Re[2]: your mail on mmap() to the kernel list

Peter Zaitsev (pz@spylog.ru)
Sat, 1 Dec 2001 11:31:30 +0300


Hello theowl,

Saturday, December 01, 2001, 2:20:20 AM, you wrote:

Well. Thank you. I've allready found this - in recent kernels it's
even regulated via proc fs.

The only question is why map anonymous is rather fast (i get
1000req/sec even then mapped 300.000 of blocks), therefore with
mapping a fd the perfomance drops to 20req/second at this number.

Any ideas why does this happen or any patches which increases the
speed exists ?

tfch> in include/linux/sched.h:

tfch> /* Maximum number of active map areas.. This is a random (large) number */
tfch> #define MAX_MAP_COUNT (65536)

tfch> this should probably be (TASK_SIZE / PAGE_SIZE) on 32 bit architectures
tfch> and something 'reasonably big' on 64 bit (too big of a value would
tfch> allow for a nice DoS against the kernel).

-- 
Best regards,
 Peter                            mailto:pz@spylog.ru

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/