From mboxrd@z Thu Jan 1 00:00:00 1970 From: xuyang Date: Thu, 16 May 2019 14:44:55 +0800 Subject: [LTP] [PATCH] ltp-cap.m4: add libcap version detection && fix compiler error In-Reply-To: <20190515143746.GA7507@dell5510> References: <1557732392-7049-1-git-send-email-xuyang2018.jy@cn.fujitsu.com> <20190513131451.GA22761@dell5510> <5CDA5124.3070708@cn.fujitsu.com> <20190514120119.GA26452@dell5510> <5CDBD564.8010300@cn.fujitsu.com> <20190515143746.GA7507@dell5510> Message-ID: <5CDD06E7.7050100@cn.fujitsu.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ltp@lists.linux.it Hi Petr > Hi Yang, > >>>> At the beginning, I >>>> think it makes no sense because libcap-1 is too old. But I notices that ltp-quota.m4 >>>> has quota_v1 and quota_v2. I changes my mind. >>>> Since we have drop libnuma v1, I agree with you that we also drop libcap v1(too old). >>>> Also, I think we should drop quota_v1 because it is too old (since Linux 2.4.22, Q_GETINFO has >>>> been supported in kernel). >>> Good point. But I'm not sure if it's the same, as QFMT_VFS_OLD is part of kernel >>> source tree (part of kernel space code) so it makes sense to test it, whereas >>> libcap v1 and libnuma v1 are just old versions of user space library. >>> Or am I wrong? >> Hi Petr >> I have seen kernel source about quota and try it. I think you are right. But current quota_v1 >> and quota_v2 detection has problem. In kernel source, I disable CONFIG_QFMT_V2 and enable CONFIG_QFMT_V1 >> ,then I found ltp still detects quotav2 instead of quotav1, and run ./quotactl01, the following error occurs: > LTP detection in ltp-quota.m4 defines _LINUX_QUOTA_VERSION for detection. > I guess you'd have to define HAVE_QUOTAV1 and undefine HAVE_QUOTAV2 in in > include/config.h, but not tested. > > BTW _LINUX_QUOTA_VERSION was removed from glibc in 2.25 [1] [2], which based setup > on. And musl defines only _LINUX_QUOTA_VERSION 2. > > Regards man info that v1 is from kernel v2.4.22 we should probably also remove > it's support. Thanks for your reply. > We got a bit away from libcap :). :-) let's go back to libcap. I will send a v2 patch to only fix possible compiler error. > Kind regards, > Petr > > [1] https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=4d728087ef8cc826b05bd21d0c74d4eca9b1a27d;hp=de6b6d14e9b4d8e9a7333addca55163d9ebb7ff6 > [2] https://patchwork.ozlabs.org/patch/664156/ > > > . >