[uml-devel] NPTL status?

[uml-devel] NPTL status?

[Christopher Faylor]

As I understand it NPTL support in UML is still a work in progress.  Is
this correct?  If so, is there an experimental patch set which works
better with NPTL and might allow things like RHEL4 to boot
out-of-the-box?

I am currently using a 2.6.12 guest kernel with RH9 rootfs and have
moved things around in /lib so as to avoid problems with NPTL but
I was hoping that there might be some improvements which would allow
me to use later versions of glibc without modification.

Is this possible?

cgf


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] NPTL status?

[Blaisorblade]

On Friday 09 September 2005 18:24, Christopher Faylor wrote:
> As I understand it NPTL support in UML is still a work in progress.  Is
> this correct?  If so, is there an experimental patch set which works
> better with NPTL
Yes, but I'd really underline _experimental_.
> and might allow things like RHEL4 to boot 
> out-of-the-box?
Basically, no. At least nothing you'd rely onto.
There is something like that, but it doesn't work very well yet.

If you browse my website, in the patches directory:

http://www.user-mode-linux.org/~blaisorblade/patches/devel-guest/

there are some test trees which include TLS support code.

It depends - at this stage, because of the difference between distros (glibc 
version mainly), with the same code I can see one behaviour (say "nothing 
works") and somebody else have no problem.

My Debian Sarge appears to be very picky, while Jeff, with more or less the 
same patches, has FC4 working "well enough" (there are still some spurious 
segfaults, however).

And Sébastien Dugué tried LinuxFromScratch 6.1 with glibc 2.3.5 and it worked 
well - but I think LFS is not a heavy stress tool for NPTL.

> I am currently using a 2.6.12 guest kernel with RH9 rootfs and have
> moved things around in /lib so as to avoid problems with NPTL but
> I was hoping that there might be some improvements which would allow
> me to use later versions of glibc without modification.
RH9 is entirely different - that wouldn't even work with a plain 2.6 kernel on 
i386, I think.

In fact, I've seen some early NPTL code on Ingo Molnar's site and it isn't at 
all compatible with current API. So RH9's NPTL will never be supported.
> Is this possible?

> cgf

-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

	

	
		
___________________________________ 
Yahoo! Mail: gratis 1GB per i messaggi e allegati da 10MB 
http://mail.yahoo.it



-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] NPTL status?

[Christopher Faylor]

On Sat, Sep 10, 2005 at 04:19:38PM +0200, Blaisorblade wrote:
>On Friday 09 September 2005 18:24, Christopher Faylor wrote:
>>As I understand it NPTL support in UML is still a work in progress.  Is
>>this correct?  If so, is there an experimental patch set which works
>>better with NPTL
>
>Yes, but I'd really underline _experimental_.

Understood.

>>and might allow things like RHEL4 to boot out-of-the-box?
>Basically, no.  At least nothing you'd rely onto.  There is something
>like that, but it doesn't work very well yet.
>
>If you browse my website, in the patches directory:
>
>http://www.user-mode-linux.org/~blaisorblade/patches/devel-guest/
>
>there are some test trees which include TLS support code.

Yes, apologies for not mentioning that I'd already looked in the
archives and tried those patches.  I don't know why I didn't think to
mention that.  IIRC, I wasn't able to completely boot RHEL4 with those,
unfortunately.  That's really my goal.  I can provide more details, if
it would help.

I saw that you were working on this and hoped that there might be newer
versions of these patches.  The newest one in the above directory is
31-Jul-2005.

cgf


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel