[PATCH] Support minimum kernel version for glibc build other than CT_KERNEL_VERSION

Yann E. MORIN yann.morin.1998@anciens.enib.fr
Tue Aug 26 18:42:00 GMT 2008


Nate,
All,

On Tuesday 26 August 2008 19:35:41 Nate Case wrote:
> Makes sense.  I just implemented this and it seems to work as you
> described.  However, I wasn't sure what you wanted the default behavior
> to be.  My personal preference would be to default to 2.6.9 (or
> whatever) rather than the kernel headers version simply so people don't
> do what I did and have to rebuild since their toolchain was unusable for
> them.   But since it is just a default setting after all, I don't have a
> strong opinion on it.

My feeling here is that people ought to build against the kernel they are
going to run on. So IMHO the default should be to use the same version as
the kernel headers. Especially, running older kernels seems a little bit
weird...

On the projects I usually work, the kernel does not change so often. Once
the kernel version has been choosen, it gets only security fixes.
Only once in a blue moon does the version changes, if at all (once the
devices are out in the wild, you can't aford upgrading them, or you don't
want the maintenance burden to support another version for the new units).

But we could argue for ever on the subject, depending on the kind of target
we are used to... So let's take the current behavior to be the default, and
keep using the kernel headers version, if you will.

> Let me know and I'll send the updated patch.

Please note that I have updated the glibc and eglibc config options to a
common file, so that it's easier to add common options (obviously). That's
in rev 960 if you care to update your patch. That's no problem if you send
it against r959, however.

> > Is there a good reason why 2.6.9 is a _sane_ default over other
> > versions?
> Admittedly, it was somewhat arbitrary :)  But the reason I chose 2.6.9
> was:
> 1) It's old enough that most people are probably targeting kernels newer
> than it
> 2) It's new enough that it wouldn't burden glibc with too much legacy
> cruft
> 3) So why not 2.6.8 or 2.6.12 ?  Well, most importantly, Gentoo builds
> glibc with --enable-kernel=2.6.9 in its ebuilds :)
> Therefore, I declared 2.6.9 to be sane :)  The old crosstool used 2.6.4
> for NPTL builds and 2.4.3 for non-NPTL, for what it's worth.

That's fine with me. Let's make 2.6.9 a sane default, then! :-)

Awaiting your patch, then! :-)

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |   ^                |
| --==< ^_^ >==-- `------------.-------:  X  AGAINST      |  /e\  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |  """  conspiracy.  |
`------------------------------^-------^------------------^--------------------'


--
For unsubscribe information see http://sourceware.org/lists.html#faq



More information about the crossgcc mailing list