Date: Tue, 4 Mar 2003 23:34:11 +0200 From: "Petri Helenius" <pete@he.iki.fi> To: <freebsd-current@FreeBSD.ORG>, <freebsd-net@FreeBSD.ORG> Subject: mbuf cache Message-ID: <0ded01c2e295$cbef0940$932a40c1@PHE>
index | next in thread | raw e-mail
I did some profiling on -CURRENT from a few days back, and I think I havenīt
figured the new tunables out or the code is not doing what itīs supposed to
or Iīm asking more than it is supposed to do but it seems that mb_free
is being quite wasteful...
Any pointers to how the new high/low watermark tunables should be used?
Is it normal that after almost all traffic has been stopped there is still 8k+
mbufs in "cache"?
Pete
granularity: each sample hit covers 16 byte(s) for 0.00% of 708.90 seconds
% cumulative self self total
time seconds seconds calls ms/call ms/call name
18.9 134.04 134.04 778488459 0.00 0.00 mb_free [5]
10.0 204.99 70.95 290131104 0.00 0.00 ether_input [8]
9.0 268.46 63.47 __mcount [9]
6.3 313.42 44.96 198223061 0.00 0.00 m_move_pkthdr [15]
5.1 349.68 36.27 18238430 0.00 0.02 em_intr [2]
5.0 385.09 35.41 778488459 0.00 0.00 mb_alloc [17]
4.8 418.87 33.77 198510151 0.00 0.00 generic_bcopy [18]
4.5 450.64 31.77 2341 13.57 63.33 m_freem [4]
4.1 479.81 29.17 967684 0.03 0.03 call_fast_unpend [20]
3.5 504.53 24.72 17641942 0.00 0.01 em_process_receive_interru
pts [3]
1.8 517.26 12.73 m_pullup [6]
1.6 528.51 11.25 290131104 0.00 0.00 em_get_buf [10]
mbuf usage:
GEN cache: 56/256 (in use/in pool)
CPU #0 cache: 8138/12064 (in use/in pool)
Total: 8194/12320 (in use/in pool)
Mbuf cache high watermark: 4096
Mbuf cache low watermark: 128
Maximum possible: 51200
Allocated mbuf types:
8194 mbufs allocated to data
24% of mbuf map consumed
mbuf cluster usage:
GEN cache: 4/16 (in use/in pool)
CPU #0 cache: 8188/12280 (in use/in pool)
Total: 8192/12296 (in use/in pool)
Cluster cache high watermark: 4096
Cluster cache low watermark: 16
Maximum possible: 25600
48% of cluster map consumed
27672 KBytes of wired memory reserved (66% in use)
0 requests for memory denied
0 requests for memory delayed
0 calls to protocol drain routines
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-net" in the body of the message
help
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?0ded01c2e295$cbef0940$932a40c1>