Eliminating gettimeofday from construct_reply
Andrew Theurer
habanero at us.ibm.com
Thu Aug 1 10:08:06 GMT 2002
From man readprofile:
The readprofile command uses the /proc/profile information to print
ascii data on standard output. The output is organized in three
columns: the first is the number of clock ticks, the second is the name
of the C function in the kernel where those many ticks occurred, and
the third is the normalized `load' of the procedure, calculated as a
ratio between the number of ticks and the length of the procedure. The
output is filled with blanks to ease readability.
-Andrew
On Thursday 01 August 2002 10:43 am, Green, Paul wrote:
> Andrew Theurer [mailto:habanero at us.ibm.com] wrote:
> > Here is the latest profile:
> > 8867 total 0.0044
> > 700 default_idle 13.4615
> > 689 __generic_copy_to_user 11.4833
> > 419 generic_file_write 0.2272
> > 291 do_softirq 1.4265
> > 260 fget 3.8235
> > 250 _text_lock_inode 1.1521
> > * 241 system_call 4.3036 *
> > 237 qdisc_restart 0.6237
> > 205 schedule 0.2330
> > 153 d_lookup 0.5709
> > 145 mod_timer 0.6144
> > 134 skb_release_data 1.1552
> > 134 ip_queue_xmit 0.1047
> > 132 _text_lock_locks 0.6600
> > 120 tcp_v4_rcv 0.0769
> > 118 link_path_walk 0.0530
> > 113 kfree 0.7847
> > 107 tcp_recvmsg 0.0459
> > 105 get_hash_table 0.6562
> > 105 __kfree_skb 0.3409
> > * 99 do_gettimeofday 0.8250 *
> >
> > *these two are what I am focusing on. I expect both to be reduced
> > significantly if we can reduce the number of calls to
> > gettimeofday. I agree, we would need to see at least a few percent
> > difference to make this worth while. I am obviously going to test
> > a before/after in any case.
>
> Cool! Only I don't understand the columns; is the first column a count and
> the second column milliseconds per invocation? Counts aren't all that
> interesting, but calculating the percent
> of total run time spent by each function would be highly interesting...
>
> PG
More information about the samba-technical
mailing list