[rea-fbsd@codelabs.ru: Patch for "device_attach: estX attach returned 6" on half of the cores]

[Eygene Ryabinkin]

Jung-uk, good day.

Tue, May 26, 2009 at 12:34:33PM -0400, Jung-uk Kim wrote:
> On Tuesday 26 May 2009 05:54 am, Eygene Ryabinkin wrote:
> > It's mine ;))  Though I hadn't looked at the newer ACPI-CA yet.
> 
> --- >8 --- SNIP!!! --- >8 ---
> 
> As I posted earlier, it should be fixed in the newer ACPI-CA:
> 
> http://docs.freebsd.org/cgi/mid.cgi?200905041551.19904.jkim
> 
> Sorry, if you missed it.

Yes, missed it, thanks for the link!

Having looked at the patch itself I have an impression that it will not
solve the original problem: while ACPI-CA now correctly dereferences
aliases found in the path (currently I am interested only Processor
objects), but still, full walk will return both real nodes and their
aliases.  So for my case of dual-way machine with aliases to both
processors, all four objects seem to be inserted to the ACPI namespace,
so again, acpi_cpu will be attached twice to the first processor.

But I want to be mistaken and will try your patch that imports the
latest ACPI-CA for the -CURRENT.  Will report on my findings.

Thanks!
-- 
Eygene
 _                ___       _.--.   #
 \`.|\..----...-'`   `-._.-'_.-'`   #  Remember that it is hard
 /  ' `         ,       __.--'      #  to read the on-line manual
 )/' _/     \   `-_,   /            #  while single-stepping the kernel.
 `-'" `"\_  ,_.-;_.-\_ ',  fsc/as   #
     _.-'_./   {_.'   ; /           #    -- FreeBSD Developers handbook
    {_.-``-'         {_/            #