Opened 21 years ago
Closed 21 years ago
#828 closed defect (wontfix)
Coreutils uname patch doesn't fix arch bug
Reported by: | Owned by: | ||
---|---|---|---|
Priority: | normal | Milestone: | |
Component: | Book | Version: | 5.0 |
Severity: | minor | Keywords: | |
Cc: |
Description
Hi,
I am having problems getting the uname command to display the architechture of the system being run on.
I have patched the coreutils package with the recommended patches as instructed in the LFS 5.0 book, and uname -i still displays "Unknown"
I have tried using the later version of coreutils from the LFS 5.1 pre CVS release and applying the 5.1 version of the patches, this results in the same problem
as an interesting asside, on another machine I had this same problem and got the coreutils package from the gentoo source tree (already patched) and it worked fine.
The machine I am having problems with at the moment is an IBM Thinkpad, built with gcc 3.3.3 no optimisation. the CPU/Arch is an Intel Pentium3 processor
I built coreutils with the the optimisation flags -Os -march=pentium3
The output of uname -a is as follows
Linux jaguar 2.4.26 #8 Wed Apr 28 19:46:13 BST 2004 i686 GenuineIntel unknown GNU/Linux
uname -i
unknown
The other LFS 5.0 machine I built that was having this problem was an AMD machine, and the output of coreutils was the same until I used the gentoo source which seems to work
uname -a Linux jordan.no-dns.co.uk 2.4.25 #3 Mon Mar 15 19:47:18 EST 2004 i686 AMD Athlon (tm) processor AuthenticAMD GNU/Linux
uname -i
AuthenticAMD
I am tempted to try the gentoo coreutils source on the pentium 3 laptop, but It hought I'd log this first to see if you needed any more information.
Getting proper output for uname is VERY difficult on x86 architecture.
Technically, the proper output for uname -i should ALWAYS be i386, because that's the architecture all x86 processors are....so the proper output for your thinkpad is:
Linux jaguar 2.4.26 #8 Wed Apr 28 19:46:13 BST 2004 i686 pentium3 i386 GNU/Linux
Achieving totally proper output is difficult at best, and really should be handled upstream by the coreutils maintainers. Our patch does about the best we can do.