Patchwork Sparc release requalification

login
register
mail settings
Submitter Josip Rodin
Date Sept. 14, 2009, 12:40 p.m.
Message ID <20090914124022.GA24872@orion.carnet.hr>
Download mbox | patch
Permalink /patch/33578/
State RFC
Delegated to: David Miller
Headers show

Comments

Josip Rodin - Sept. 14, 2009, 12:40 p.m.
On Mon, Sep 14, 2009 at 02:36:48PM +0200, Josip Rodin wrote:
> > > 2) Try a UP kernel build, does it work even with all of the NMI
> > >    bits enabled on this machine?
> > 
> > Will try and report shortly.
> 
> It seems to work just fine without SMP.

Just in case it might matter:
David Miller - Sept. 15, 2009, 9:11 a.m.
From: Josip Rodin <joy@entuzijast.net>
Date: Mon, 14 Sep 2009 14:40:22 +0200

> On Mon, Sep 14, 2009 at 02:36:48PM +0200, Josip Rodin wrote:
>> > > 2) Try a UP kernel build, does it work even with all of the NMI
>> > >    bits enabled on this machine?
>> > 
>> > Will try and report shortly.
>> 
>> It seems to work just fine without SMP.
> 
> Just in case it might matter:

I was able to reproduce the hang with your originally posted config.

It only triggers when CONFIG_PROM_CONSOLE is enabled, I keep telling
people to turn it off.  Nobody should be using it, it either breaks
things or it slows the console down considerably.

That's it, I'm completely removing that driver from the tree this
merge window, so you guys can't mess it up any more even if you want
to :-)
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Josip Rodin - Sept. 15, 2009, 9:58 a.m.
On Tue, Sep 15, 2009 at 02:11:24AM -0700, David Miller wrote:
> I was able to reproduce the hang with your originally posted config.
> 
> It only triggers when CONFIG_PROM_CONSOLE is enabled, I keep telling
> people to turn it off.  Nobody should be using it, it either breaks
> things or it slows the console down considerably.
> 
> That's it, I'm completely removing that driver from the tree this
> merge window, so you guys can't mess it up any more even if you want
> to :-)

Oh! Well, that's good.

I googled for it now and found in some Ubuntu kernel changelog that Fabio
did it this way:

linux-source-2.6.17 (2.6.17-10.26) edgy; urgency=low

  *  make promcon driver init a boot option:
    - fix Niagara cpu soft lockups. (Closes Ubuntu: #61821)

  The generic PROM_CONSOLE driver is just too slow and blocks IRQ during
  normal output. Ideally the PROM_CONSOLE should be used only if we fail
  to find the correct driver for a certain hardware. In reality when built in
  it will be always used as default. With this workaround, we init the driver
  only on specific requests.
  The pro is that all machines that have a working console driver will notice
  a speed up at boot. The cons is that machines for which we have no driver
  will need to wait for one or boot with "forcepromconsole" boot option.

 -- Fabio M. Di Nitto <email address hidden>   Mon, 02 Oct 2006 06:48:54 +0200

If the PROM console driver still has some utility, maybe the boot option is
the way to go... does it? Does anyone still manufacture new machines with
new and strange console types that we don't support? :)
David Miller - Sept. 15, 2009, 10:01 a.m.
From: Josip Rodin <joy@entuzijast.net>
Date: Tue, 15 Sep 2009 11:58:33 +0200

> If the PROM console driver still has some utility, maybe the boot option is
> the way to go... does it? Does anyone still manufacture new machines with
> new and strange console types that we don't support? :)

The PROM console driver has no relevance today at all.

It should simply never be used.
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Josip Rodin - Sept. 15, 2009, 10:03 a.m.
On Tue, Sep 15, 2009 at 03:01:29AM -0700, David Miller wrote:
> From: Josip Rodin <joy@entuzijast.net>
> Date: Tue, 15 Sep 2009 11:58:33 +0200
> 
> > If the PROM console driver still has some utility, maybe the boot option is
> > the way to go... does it? Does anyone still manufacture new machines with
> > new and strange console types that we don't support? :)
> 
> The PROM console driver has no relevance today at all.
> 
> It should simply never be used.

OK, cool, please remove it, and also please propagate that to the stable
branches so we don't miss anyone who's not on the bleeding edge.
seb@frankengul.org - Sept. 15, 2009, 10:05 a.m.
David Miller a écrit :
> From: Josip Rodin <joy@entuzijast.net>
> Date: Mon, 14 Sep 2009 14:40:22 +0200
>
>   
>> On Mon, Sep 14, 2009 at 02:36:48PM +0200, Josip Rodin wrote:
>>     
>>>>> 2) Try a UP kernel build, does it work even with all of the NMI
>>>>>    bits enabled on this machine?
>>>>>           
>>>> Will try and report shortly.
>>>>         
>>> It seems to work just fine without SMP.
>>>       
>> Just in case it might matter:
>>     
>
> I was able to reproduce the hang with your originally posted config.
>
> It only triggers when CONFIG_PROM_CONSOLE is enabled, I keep telling
> people to turn it off.  Nobody should be using it, it either breaks
> things or it slows the console down considerably.
>
> That's it, I'm completely removing that driver from the tree this
> merge window, so you guys can't mess it up any more even if you want
> to :-)
>
>
>   
Very good news.
I had a lot of trouble because of this.
I had to disable it for initcall_debug to work, but then, any kernel 
patched or not was working.
I thought I messed up between patches and built kernels.
I realize now that I had the explanation under my nose.
Duh....

    Seb

--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jurij Smakov - Sept. 15, 2009, 7:38 p.m.
On Tue, Sep 15, 2009 at 12:03:06PM +0200, Josip Rodin wrote:
> On Tue, Sep 15, 2009 at 03:01:29AM -0700, David Miller wrote:
> > From: Josip Rodin <joy@entuzijast.net>
> > Date: Tue, 15 Sep 2009 11:58:33 +0200
> > 
> > > If the PROM console driver still has some utility, maybe the boot option is
> > > the way to go... does it? Does anyone still manufacture new machines with
> > > new and strange console types that we don't support? :)
> > 
> > The PROM console driver has no relevance today at all.
> > 
> > It should simply never be used.
> 
> OK, cool, please remove it, and also please propagate that to the stable
> branches so we don't miss anyone who's not on the bleeding edge.

Please feel free to follow up on http://bugs.debian.org/525958 which I've
filed in April to have the CONFIG_PROM_CONSOLE removed.

> 
> -- 
>      2. That which causes joy or happiness.
> 
> 
> -- 
> To UNSUBSCRIBE, email to debian-sparc-REQUEST@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Josip Rodin - Sept. 16, 2009, 8:03 a.m.
On Tue, Sep 15, 2009 at 08:38:22PM +0100, Jurij Smakov wrote:
> > > > If the PROM console driver still has some utility, maybe the boot option is
> > > > the way to go... does it? Does anyone still manufacture new machines with
> > > > new and strange console types that we don't support? :)
> > > 
> > > The PROM console driver has no relevance today at all.
> > > 
> > > It should simply never be used.
> > 
> > OK, cool, please remove it, and also please propagate that to the stable
> > branches so we don't miss anyone who's not on the bleeding edge.
> 
> Please feel free to follow up on http://bugs.debian.org/525958 which I've
> filed in April to have the CONFIG_PROM_CONSOLE removed.

Yeah, definitely, Cc:ed. Dave has in the meantime killed it completely in
his stable tree:

http://git.kernel.org/?p=linux/kernel/git/davem/sparc-2.6.git;a=commit;h=09d3f3f0e02c8a900d076c302c5c02227f33572d

There's another commit that takes it out of the defconfigs completely - it
was already unset there for a while now. So it's pretty much official now
(well, Linus still has to take it in, but that should be a formality).

Patch

--- /boot/config-2.6.31.old     2009-09-11 23:10:03.000000000 +0000
+++ /boot/config-2.6.31 2009-09-14 12:08:04.000000000 +0000
@@ -4 +4 @@ 
-# Fri Sep 11 18:23:35 2009
+# Mon Sep 14 11:47:09 2009
@@ -35 +35 @@ 
-CONFIG_LOCK_KERNEL=y
+CONFIG_BROKEN_ON_SMP=y
@@ -114 +113,0 @@ 
-CONFIG_USE_GENERIC_SMP_HELPERS=y
@@ -131 +129,0 @@ 
-CONFIG_STOP_MACHINE=y
@@ -154,2 +152 @@ 
-CONFIG_SMP=y
-CONFIG_NR_CPUS=32
+# CONFIG_SMP is not set
@@ -167 +163,0 @@ 
-CONFIG_SPARC64_SMP=y
@@ -171 +166,0 @@ 
-# CONFIG_HOTPLUG_CPU is not set
@@ -178 +172,0 @@ 
-# CONFIG_NUMA is not set
@@ -200,2 +193,0 @@ 
-CONFIG_SCHED_SMT=y
-CONFIG_SCHED_MC=y