> -----Original Message-----
> From: Rob van der Heij [mailto:[EMAIL PROTECTED]
> Sent: Monday, March 03, 2003 12:48 PM
> To: [EMAIL PROTECTED]
> Subject: Re: Kernel patch to add VM IPL PARM support
>
>
> Lucius, Leland wrote:
> > So, which should it be?  Append or prepend the PARMs to the
> command line?  I
> > haven't looked too deeply, but it appears that processing
> kernel parameters
> > isn't too consistent and some rtns take the first parameter
> encountered.  As
> > it is, appending PARMs wouldn't allow you to override in
> those situations.
>
> A similar implementation was published in the ISP/ASP
> redbook. So far I have
> failed in trying to get our friends in Boeblingen to take the
> code I offered.
That's a bummer.  A shame they aren't more open to "user contributed"
patches.

> I can justify spending some type in protoyping code, but if
> it does not make
> it into the official sources it is a burden to support it
>
My take on this is a tad different.  Since we will be using the patch, I
don't have a problem with maintaining it.  Not much different than the
usermods we maintain on our OS/390 LPARs.  But, I will only maintain a
version that fits our needs.  Since the patch is/will be made available to
anyone, then it would be a simple matter for "anyone" to modify it to suit
their needs.

> We've argued also about the order of the parameters. In our
> case the rest of
> the options is in the NSS, and in real life it turns out we
> mainly needed it
> to specify the root= (which takes the last one) and to
> specify a few more disk
> devices (takes them all in the order specified). Also the
> option to suppress
> the DEVFS mount came in handy (when you forget to disable it
> in the kernel).
> But the main thing remains the simplicity to save the NSS.
>
My guess is that appending would be quite satisfactory, but it won't be
difficult to allow both prefixing and suffixing.

Leland

Reply via email to