Scott Long wrote: >> BTW, will you work on CAM, target mode support, or 4G/MP safety of any >> of the two? > CAM: yes > target mode: yes > 4G/SMP: yes Hmm. Is it christmas, or something? :) This is absolutely great news! Do you have any deeper details? BTW, this is what I got on 6-STABLE with two HP BL25p connected to each other with a single FC loop: target: CPU states: 3.0% user, 0.0% nice, 59.9% system, 15.4% interrupt, 21.7% idle initiator: dT: 0.502 flag_I 500000us sizeof 288 i -1 L(q) ops/s r/s kBps ms/r w/s kBps ms/w %busy Name 2 261 147 18869 6.7 114 14534 8.6 99.5| da1 2 265 149 19124 6.7 116 14789 8.6 100.9| da2 2 275 157 20144 6.3 118 15044 8.3 100.0| da3 2 273 155 19889 6.3 118 15044 8.4 100.0| da4 2 267 153 19634 6.4 114 14534 8.6 100.0| da5 2 277 157 20144 6.3 120 15299 8.3 100.7| da6 2 263 151 19379 6.6 112 14279 8.7 99.9| da7 2 267 153 19634 6.5 114 14534 8.7 100.2| da8 The machines have two Opteron 252, 8G of RAM (running with 4G because of isp (I guess)) and a Qlogic 2312 dual port 2G FC adapter (only one port utilized). On the target I did 8 malloc backed RAM disks and shared them with scsi_target. On the initiator, I do this (just blindly testing for now): dd if=/dev/zero of=/dev/daX bs=1024k dd if=/dev/daX of=/dev/zero bs=1024k simultaneously for the 8 LUNs. -- Attila Nagy e-mail: Attila.Nagy_at_fsn.hu Free Software Network (FSN.HU) phone: +3630 306 6758 http://www.fsn.hu/Received on Sun Jan 22 2006 - 18:28:16 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:51 UTC