From: Michael M. <mik...@am...> - 2003-11-25 23:12:46
|
Hello, I'm somewhat of a newbie so bear with me, please. I'm trying to run ltp-full-20031106 on a RH7.2 system with a stock 2.4.17 kernel ( I didn't get to choose the OS ). LTP was compiled with gcc 2.95.3 ( not RH's compiler ). The platform is based on an NSC Geode processor ( roughly compatible to an Intel PII ). In trying to compare errors from a new version of LTP on an older system, I realize I'm probably comparing apples and oranges, but I'm also operating under the assumption that newer software generally works better than older. Most of the errors I'm getting I can match to older kernels, glibc, etc. The error I'm having problems with is from the sendmsg01 test. From the log file I get : "sendmsg01 11 FAIL : invalid iovec count ; returned -1 (expected -1), errno 22 (expected 90)" I've seen similar errors in the results lists, but on a PPC using later kernels and glibc. No details were posted so I don't know if this is the same error. I have several questions: 1) Is sendmsg01 a fairly new test? 2) My gut is telling me that this is probably a software error, possibly related to libraries, but I don't know for sure. Is there someone who can shed a little light on any possible differences in libraries/platforms which might account for the error? Thanks in advance, MikeM |
From: Robert W. <ro...@us...> - 2003-11-26 16:01:36
|
> 1) Is sendmsg01 a fairly new test? Not really, but it has been changed a little over the past year...for better support across different archs. > 2) My gut is telling me that this is probably a software error, possibly > related to libraries, but I don't know for sure. Is there someone who > can shed a little light on any possible differences in > libraries/platforms which might account for the error? This is a glibc library bug, where the wrong error code is being returned. The test fails for me on a Linux box with glibc 2.2.5-184. However, on another Linux box with glibc 2.3.2-27.9.7, it passes. What level are the glibc packages on your box? (# rpm -qa | grep glibc) FYI, We try to adhere to the Open Group's Single UNIX Specification version 3 when it comes to tests. Granted this is Linux, not UNIX, but we need some type of standard. The SUSv3 states that sendmsg() shall fail and return EMSGSIZE if: The message is too large to be sent all at once (as the socket requires), or the msg_iovlen member of the msghdr structure pointed to by message is less than or equal to 0 or is greater than {IOV_MAX}. In Linux, EMSGSIZE maps to errno 90, and in the test the msg_iovlen member, which represents the number of elements in msg_iov, is set to -1. Taking all this into consideration, the errno should be 90 b/c the msg_iovlen member is less than 0. Open Group Website: http://www.unix-systems.org/single_unix_specification/ -Robbie Robert V. Williamson <ro...@us...> Linux Test Project IBM Linux Technology Center Web: http://ltp.sourceforge.net IRC: #ltp on freenode.irc.net ==================== "Only two things are infinite, the universe and human stupidity, and I'm not sure about the former." -Albert Einstein "Michael Meador" <mik...@am...> To: ltp...@li... Sent by: cc: ltp...@li...ur Subject: [LTP] sendmsg01 error ceforge.net 11/25/2003 05:12 PM Hello, I'm somewhat of a newbie so bear with me, please. I'm trying to run ltp-full-20031106 on a RH7.2 system with a stock 2.4.17 kernel ( I didn't get to choose the OS ). LTP was compiled with gcc 2.95.3 ( not RH's compiler ). The platform is based on an NSC Geode processor ( roughly compatible to an Intel PII ). In trying to compare errors from a new version of LTP on an older system, I realize I'm probably comparing apples and oranges, but I'm also operating under the assumption that newer software generally works better than older. Most of the errors I'm getting I can match to older kernels, glibc, etc. The error I'm having problems with is from the sendmsg01 test. From the log file I get : "sendmsg01 11 FAIL : invalid iovec count ; returned -1 (expected -1), errno 22 (expected 90)" I've seen similar errors in the results lists, but on a PPC using later kernels and glibc. No details were posted so I don't know if this is the same error. I have several questions: 1) Is sendmsg01 a fairly new test? 2) My gut is telling me that this is probably a software error, possibly related to libraries, but I don't know for sure. Is there someone who can shed a little light on any possible differences in libraries/platforms which might account for the error? Thanks in advance, MikeM ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ Ltp-list mailing list Ltp...@li... https://lists.sourceforge.net/lists/listinfo/ltp-list |