[OpenAFS-devel] Issue building 1.5.77 rpms for i386
Phillip Moore
w.phillip.moore@gmail.com
Fri, 8 Oct 2010 11:37:44 -0400
--20cf300fb205b23a8304921cca61
Content-Type: text/plain; charset=ISO-8859-1
No, the problem is that the openafs.spec file behaves a little wierd for the
i386 and i686 targets. I confess to not really understand the practical
difference here.
See my last response here -- it sums up the issue.
I'm assuming the logic was to create user-mode binaries that are generic as
possible (i386), but to build the kernel module for the newer architecture
(i686).
It's annoying, because it means you have to make TWO passes with rpmbuild to
get a complete set of rpms for the 32bit platform.
On Fri, Oct 8, 2010 at 9:21 AM, Derrick Brashear <shadow@gmail.com> wrote:
> On Thu, Oct 7, 2010 at 9:14 PM, Phillip Moore <w.phillip.moore@gmail.com>
> wrote:
> > I managed to get a complete set of 1.5.77 rpms for x86_64, but when I
> > repeated the entire process for i386, I got all the same rpms, except
> one:
> > kmod-openafs-*. IOW, it didn't build the most important one :-(
>
> you installed the i386 kernel-devel rpm also, right? (i'm betting no)
>
> > This is on CentOS 5.4. The list of rpms I get on x86_64 is:
> > -rw-r--r-- 1 efsops efsops 1154050 Oct 7 17:11
> > dkms-openafs-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 299341 Oct 7 17:11
> > kmod-openafs-1.5.77-1.1.2.6.18_164.el5.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 4244312 Oct 7 17:11
> > openafs-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 245869 Oct 7 17:11
> > openafs-authlibs-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 560495 Oct 7 17:11
> > openafs-authlibs-devel-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 764714 Oct 7 17:11
> > openafs-client-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 13287 Oct 7 17:11
> > openafs-compat-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 1190887 Oct 7 17:11
> > openafs-devel-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 1816226 Oct 7 17:11
> > openafs-docs-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 1161461 Oct 7 17:11
> > openafs-kernel-source-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 143921 Oct 7 17:11
> > openafs-kpasswd-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 300444 Oct 7 17:11
> > openafs-krb5-1.5.77-1.1.1.x86_64.rpm
> > -rw-r--r-- 1 efsops efsops 3979585 Oct 7 17:11
> > openafs-server-1.5.77-1.1.1.x86_64.rpm
> > The list on i386:
> > -rw-r--r-- 1 efsops efsops 1154119 Oct 7 17:50
> > dkms-openafs-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 3927398 Oct 7 17:50
> > openafs-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 238659 Oct 7 17:50
> > openafs-authlibs-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 525398 Oct 7 17:50
> > openafs-authlibs-devel-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 702579 Oct 7 17:50
> > openafs-client-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 13318 Oct 7 17:50
> > openafs-compat-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 1081577 Oct 7 17:50
> > openafs-devel-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 1816259 Oct 7 17:50
> > openafs-docs-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 1161445 Oct 7 17:50
> > openafs-kernel-source-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 129702 Oct 7 17:50
> > openafs-kpasswd-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 270950 Oct 7 17:50
> > openafs-krb5-1.5.77-1.1.1.i386.rpm
> > -rw-r--r-- 1 efsops efsops 3835766 Oct 7 17:50
> > openafs-server-1.5.77-1.1.1.i386.rpm
> > Is there anything special that needs to be done to coerce the build of
> the
> > kmod-openafs rpm on the 32 bit platform?
> > I can't find the code/file/whatever that determines whether or not it
> gets
> > built....
>
>
>
> --
> Derrick
>
--20cf300fb205b23a8304921cca61
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
No, the problem is that the openafs.spec file behaves a little wierd for th=
e i386 and i686 targets. =A0I confess to not really understand the practica=
l difference here.<div><br></div><div>See my last response here -- it sums =
up the issue.</div>
<div><br></div><div>I'm assuming the logic was to create user-mode bina=
ries that are generic as possible (i386), but to build the kernel module fo=
r the newer architecture (i686).</div><div><br></div><div>It's annoying=
, because it means you have to make TWO passes with rpmbuild to get a compl=
ete set of rpms for the 32bit platform.</div>
<div><br></div><div><br><div class=3D"gmail_quote">On Fri, Oct 8, 2010 at 9=
:21 AM, Derrick Brashear <span dir=3D"ltr"><<a href=3D"mailto:shadow@gma=
il.com">shadow@gmail.com</a>></span> wrote:<br><blockquote class=3D"gmai=
l_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left=
:1ex;">
<div class=3D"im">On Thu, Oct 7, 2010 at 9:14 PM, Phillip Moore <<a href=
=3D"mailto:w.phillip.moore@gmail.com">w.phillip.moore@gmail.com</a>> wro=
te:<br>
> I managed to get a complete set of 1.5.77 rpms for x86_64, but when I<=
br>
> repeated the entire process for i386, I got all the same rpms, except =
one:<br>
> =A0kmod-openafs-*. =A0=A0IOW, it didn't build the most important o=
ne :-(<br>
<br>
</div>you installed the i386 kernel-devel rpm also, right? (i'm betting=
no)<br>
<div><div></div><div class=3D"h5"><br>
> This is on CentOS 5.4. =A0 The list of rpms I get on x86_64 is:<br>
> -rw-r--r-- 1 efsops efsops 1154050 Oct =A07 17:11<br>
> dkms-openafs-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0299341 Oct =A07 17:11<br>
> kmod-openafs-1.5.77-1.1.2.6.18_164.el5.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops 4244312 Oct =A07 17:11<br>
> openafs-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0245869 Oct =A07 17:11<br>
> openafs-authlibs-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0560495 Oct =A07 17:11<br>
> openafs-authlibs-devel-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0764714 Oct =A07 17:11<br>
> openafs-client-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0 13287 Oct =A07 17:11<br>
> openafs-compat-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops 1190887 Oct =A07 17:11<br>
> openafs-devel-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops 1816226 Oct =A07 17:11<br>
> openafs-docs-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops 1161461 Oct =A07 17:11<br>
> openafs-kernel-source-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0143921 Oct =A07 17:11<br>
> openafs-kpasswd-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0300444 Oct =A07 17:11<br>
> openafs-krb5-1.5.77-1.1.1.x86_64.rpm<br>
> -rw-r--r-- 1 efsops efsops 3979585 Oct =A07 17:11<br>
> openafs-server-1.5.77-1.1.1.x86_64.rpm<br>
> The list on i386:<br>
> -rw-r--r-- 1 efsops efsops 1154119 Oct =A07 17:50<br>
> dkms-openafs-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops 3927398 Oct =A07 17:50<br>
> openafs-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0238659 Oct =A07 17:50<br>
> openafs-authlibs-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0525398 Oct =A07 17:50<br>
> openafs-authlibs-devel-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0702579 Oct =A07 17:50<br>
> openafs-client-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0 13318 Oct =A07 17:50<br>
> openafs-compat-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops 1081577 Oct =A07 17:50<br>
> openafs-devel-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops 1816259 Oct =A07 17:50<br>
> openafs-docs-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops 1161445 Oct =A07 17:50<br>
> openafs-kernel-source-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0129702 Oct =A07 17:50<br>
> openafs-kpasswd-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops =A0270950 Oct =A07 17:50<br>
> openafs-krb5-1.5.77-1.1.1.i386.rpm<br>
> -rw-r--r-- 1 efsops efsops 3835766 Oct =A07 17:50<br>
> openafs-server-1.5.77-1.1.1.i386.rpm<br>
> Is there anything special that needs to be done to coerce the build of=
the<br>
> kmod-openafs rpm on the 32 bit platform?<br>
> I can't find the code/file/whatever that determines whether or not=
it gets<br>
> built....<br>
<br>
<br>
<br>
</div></div>--<br>
<font color=3D"#888888">Derrick<br>
</font></blockquote></div><br></div>
--20cf300fb205b23a8304921cca61--