lkml.org 
[lkml]   [2002]   [Feb]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] 2.5.5-pre1 IDE cleanup 9
Gérard Roudier wrote:
>
> On Fri, 22 Feb 2002, Jeff Garzik wrote:
>
> > Gérard Roudier wrote:
> > > Basically at the moment, if the driver allows upper 'seeming cleaner and
> > > smarter' PCI probing things to deal with the HBA attachment order, at
> > > least all my machines running Linux will not even reboot.
> > >
> > > Being smart is doing what user expects, here.
> >
> > Oh come on, how hard is the following?
> >
> > > static int __init foo_init(void)
> > > {
> > > int rc = pci_module_init(&sym2_pci_driver);
> > > if (rc) return rc;
> > > do_deferred_work();
> > > }
> > > module_init(foo_init);
> >
> > You have tons of flexibility you are ignoring here... For the
> > non-hotplug hosts (ie. present at boot), just use pci_driver::probe to
> > register hosts on a list, and little other work. do_deferred_work()
> > handles the list in a manner that ensures proper boot and/or host
> > ordering.
> >
> > So for non-hotplug hosts you do a init_module time:
> > register N hosts with PCI API
> > register N hosts with SCSI API
> >
> > And hotplugged hosts would do the same, with N==1.
> >
> > What you describe -is- supported with the PCI API.
>
> At the time I investigated the API it just mixed the probing and the
> registering by performing some auto-registration based on return value.
> May-be the API did evolve since that time or I missed something important.
>
> For now I will be in vacation for 1 week. I will re-investigate this when
> I will be back.

Thanks!

One thing that is slowly becoming apparently to me during this thread is
the importance of separating ordering [of hosts, of disks] from the
registration of the resource itself.

Thinking about the problem a bit more (NVRAM boot disk ordering, etc.) I
believe that what I describe above might be considered a transition
step... In Step Two, do_deferred_work() [above] would likely be moved
to userspace, running on initramfs.

Jeff



--
Jeff Garzik | "UNIX enhancements aren't."
Building 1024 | -- says /usr/games/fortune
MandrakeSoft |
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:24    [W:0.521 / U:0.828 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site