Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 4 Feb 2017 00:54:15 -0500
From:      Kurt Lidl <lidl@pix.net>
To:        Jason Harmening <jason.harmening@gmail.com>, Alexander Kabaev <kabaev@gmail.com>
Cc:        "Jason A. Harmening" <jah@freebsd.org>, src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org, Ed Maste <emaste@freebsd.org>
Subject:   Re: svn commit: r313037 - in head/sys: amd64/include kern mips/include net powerpc/include sparc64/include
Message-ID:  <8523aaa5-6c30-9f9f-40f0-fdf82cdf1669@pix.net>
In-Reply-To: <CAM=8qa=h3=sfro02hCQyzqkDnLO7TnQJ8ugCoa5=MfNE_OCgZg@mail.gmail.com>
References:  <201702010332.v113WnYf041362@repo.freebsd.org> <20170203231238.0675c289@kan> <CAM=8qa=h3=sfro02hCQyzqkDnLO7TnQJ8ugCoa5=MfNE_OCgZg@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Having just spent a couple of hours bisecting what broke the kernel on
my mips64 machine, I can definitively state it was this commit.

With this commit in place, the kernel hangs early in the
autoconfiguration:

gcc version 4.2.1 20070831 patched [FreeBSD]
Preloaded elf kernel "kernel" at 0xffffffff80aa96a0.
real memory  = 523239424 (510976K bytes)
Physical memory chunk(s):
0x00bf3000 - 0x080d5fff, 122564608 bytes (29923 pages)
0x08101000 - 0x0ff00fff, 132120576 bytes (32256 pages)
0x410000000 - 0x41f196fff, 253325312 bytes (61847 pages)
avail memory = 504360960 (480MB)
Create COP2 context zone
AP #1 started!
FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
---- hangs here ----

-Kurt

On 2/4/17 12:29 AM, Jason Harmening wrote:
> Hi,
>
> I'm a bit confused as to how this change breaks MIPS.  The new function,
> get_pcpu() is intended to be used only to access the per-cpu data
> pointer locally.  It returns pcpup, which is the per-cpu pointer wired
> into the local TLB to translate to the local CPU's physical data region,
> correct?
>
> This is the same value used by the per-CPU accessors such as PCPU_ADD
> and PCPU_GET.  The MI portions of this change only use get_pcpu() to
> access  the local CPU's data, e.g. under a critical section in the
> rmlock.  It is not intended to be used for iterating all CPUs.
>
> If I've missed something and MIPS is truly broken by this, then I'll
> gladly revert, but (maybe because it's late) I'm not seeing where this
> goes wrong on MIPS.
>
> Thanks,
> Jason
>
> On Fri, Feb 3, 2017 at 8:12 PM, Alexander Kabaev <kabaev@gmail.com
> <mailto:kabaev@gmail.com>> wrote:
>
>     On Wed, 1 Feb 2017 03:32:49 +0000 (UTC)
>     "Jason A. Harmening" <jah@FreeBSD.org> wrote:
>
>     > Author: jah
>     > Date: Wed Feb  1 03:32:49 2017
>     > New Revision: 313037
>     > URL: https://svnweb.freebsd.org/changeset/base/313037
>     <https://svnweb.freebsd.org/changeset/base/313037>;
>     >
>     > Log:
>     >   Implement get_pcpu() for the remaining architectures and use it to
>     >   replace pcpu_find(curcpu) in MI code.
>     >
>     > Modified:
>     >   head/sys/amd64/include/pcpu.h
>     >   head/sys/kern/kern_rmlock.c
>     >   head/sys/mips/include/pcpu.h
>     >   head/sys/net/netisr.c
>     >   head/sys/powerpc/include/cpufunc.h
>     >   head/sys/powerpc/include/pcpu.h
>     >   head/sys/sparc64/include/pcpu.h
>     >
>
>     Hi,
>
>     this change was not reviewed nor testing was thought for all
>     architectures it touches. The change happens to break MIPS quite
>     thoroughly, since MIPS is using different pointers when accessing PCPU
>     area locally and when doing iterations using cpu_to_cpuid array. I
>     therefore officially am requesting this change to be reverted until
>     reasonable solution is found to unbreak architectures that use wired
>     TLBs to access local per-CPU data.
>
>     --
>     Alexander Kabaev
>
>




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8523aaa5-6c30-9f9f-40f0-fdf82cdf1669>