Hi Martin, I cut the threads number from 16 to 8, and now the virtual memory has reduced from 131M to 67M. But the error still exists. If error code "93" is EPROTONOSUPPORT, I think maybe it is still the problem of linux kernel (version under 2.6.29) about the IPv6 kernel modules. But I actually patched the kernel using the patch you provided for me and I also patch the strongswan using your provided patch. The module I installed into kernel is listed as followed: >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> xfrm4_tunnel 2752 0 - Live 0xbf07e000 tunnel4 3752 1 xfrm4_tunnel, Live 0xbf078000 ipcomp 3232 0 - Live 0xbf072000 xfrm_ipcomp 5384 1 ipcomp, Live 0xbf06e000 xfrm_user 20544 0 - Live 0xbf063000 esp4 6528 0 - Live 0xbf05c000 ah4 5248 0 - Live 0xbf055000 af_key 32464 0 - Live 0xbf048000 <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< Best Regards, David
-----邮件原件----- 发件人: Martin Willi [mailto:[email protected]] 发送时间: 2009年9月18日 18:16 收件人: weiping deng 抄送: [email protected] 主题: Re: 答复: How to peel off strongswan code for running in an space-stressed ARM Hi, > 1) If the used virtual memory exceed, the following error will be occurred, > is it right? > Resource temporarily unavailable-93: received netlink error I have never seen such a "Resource temporarily unavailable" error from netlink, and 93 is actually EPROTONOSUPPORT. You are probably missing a kernel module (tunnel/transport/esp/...), make sure to have a kernel with all options for your setup. I don't think this has anything to do with memory consumption. Virtual memory does not really exceed, unless you reach the end of address space. > it can be set in strongswan.conf file as followed: > Charon > { threads = 8; ....} Yes, but strongswan.conf does not use semicolons. Regards Martin _______________________________________________ Users mailing list [email protected] https://lists.strongswan.org/mailman/listinfo/users
