mpic discovery on JS20

mpic discovery on JS20

[Amos Waterland]

Current Linus and Paulus trees do this on JS20 blades with SLOF:

 Failed to locate the MPIC interrupt controller
 PID hash table entries: 4096 (order: 12, 32768 bytes)
 Maple: Found RTC at IO 0x1070
 cpu 0x0: Vector: 700 (Program Check) at [c00000007ef83ab0]
     pc: c00000000002e0c8: .mpic_request_ipis+0x34/0xc8
     lr: c00000000036b484: .smp_mpic_probe+0x3c/0x58
     sp: c00000007ef83d30
    msr: 9000000000029032
   current = 0xc00000000194d610
   paca    = 0xc00000000038f180
     pid   = 1, comm = swapper
 kernel BUG in mpic_request_ipis at arch/powerpc/sysdev/mpic.c:1132!

Doing a git-bisect produces this:

 0ebfff1491ef85d41ddf9c633834838be144f69f is first bad commit
 Author: Benjamin Herrenschmidt <benh@kernel.crashing.org>
 Date:   Mon Jul 3 21:36:01 2006 +1000

    [POWERPC] Add new interrupt mapping core and change platforms to use it
    
    This adds the new irq remapper core and removes the old one.  Because
    there are some fundamental conflicts with the old code, like the value
    of NO_IRQ which I'm now setting to 0 (as per discussions with Linus),
    etc..., this commit also changes the relevant platform and driver code
    over to use the new remapper (so as not to cause difficulties later
    in bisecting).
    
    This patch removes the old pre-parsing of the open firmware interrupt
    tree along with all the bogus assumptions it made to try to renumber
    interrupts according to the platform. This is all to be handled by the
    new code now.
    
    For the pSeries XICS interrupt controller, a single remapper host is
    created for the whole machine regardless of how many interrupt
    presentation and source controllers are found, and it's set to match
    any device node that isn't a 8259.  That works fine on pSeries and
    avoids having to deal with some of the complexities of split source
    controllers vs. presentation controllers in the pSeries device trees.
    
    The powerpc i8259 PIC driver now always requests the legacy interrupt
    range. It also has the feature of being able to match any device node
    (including NULL) if passed no device node as an input. That will help
    porting over platforms with broken device-trees like Pegasos who don't
    have a proper interrupt tree.
    
    Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
    Signed-off-by: Paul Mackerras <paulus@samba.org>

Re: mpic discovery on JS20

[Benjamin Herrenschmidt]

On Thu, 2006-07-20 at 19:16 -0400, Amos Waterland wrote:
> Current Linus and Paulus trees do this on JS20 blades with SLOF:

I need a tarball of /proc/device-tree on these. 

>  Failed to locate the MPIC interrupt controller
>  PID hash table entries: 4096 (order: 12, 32768 bytes)
>  Maple: Found RTC at IO 0x1070
>  cpu 0x0: Vector: 700 (Program Check) at [c00000007ef83ab0]
>      pc: c00000000002e0c8: .mpic_request_ipis+0x34/0xc8
>      lr: c00000000036b484: .smp_mpic_probe+0x3c/0x58
>      sp: c00000007ef83d30
>     msr: 9000000000029032
>    current = 0xc00000000194d610
>    paca    = 0xc00000000038f180
>      pid   = 1, comm = swapper
>  kernel BUG in mpic_request_ipis at arch/powerpc/sysdev/mpic.c:1132!
> 
> Doing a git-bisect produces this:
> 
>  0ebfff1491ef85d41ddf9c633834838be144f69f is first bad commit
>  Author: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>  Date:   Mon Jul 3 21:36:01 2006 +1000
> 
>     [POWERPC] Add new interrupt mapping core and change platforms to use it
>     
>     This adds the new irq remapper core and removes the old one.  Because
>     there are some fundamental conflicts with the old code, like the value
>     of NO_IRQ which I'm now setting to 0 (as per discussions with Linus),
>     etc..., this commit also changes the relevant platform and driver code
>     over to use the new remapper (so as not to cause difficulties later
>     in bisecting).
>     
>     This patch removes the old pre-parsing of the open firmware interrupt
>     tree along with all the bogus assumptions it made to try to renumber
>     interrupts according to the platform. This is all to be handled by the
>     new code now.
>     
>     For the pSeries XICS interrupt controller, a single remapper host is
>     created for the whole machine regardless of how many interrupt
>     presentation and source controllers are found, and it's set to match
>     any device node that isn't a 8259.  That works fine on pSeries and
>     avoids having to deal with some of the complexities of split source
>     controllers vs. presentation controllers in the pSeries device trees.
>     
>     The powerpc i8259 PIC driver now always requests the legacy interrupt
>     range. It also has the feature of being able to match any device node
>     (including NULL) if passed no device node as an input. That will help
>     porting over platforms with broken device-trees like Pegasos who don't
>     have a proper interrupt tree.
>     
>     Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>     Signed-off-by: Paul Mackerras <paulus@samba.org>
> 
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-dev

Re: mpic discovery on JS20

[Michael Ellerman]

[-- Attachment #1.1: Type: text/plain, Size: 716 bytes --]

On Thu, 2006-07-20 at 19:16 -0400, Amos Waterland wrote:
> Current Linus and Paulus trees do this on JS20 blades with SLOF:
> 
>  Failed to locate the MPIC interrupt controller
>  PID hash table entries: 4096 (order: 12, 32768 bytes)
>  Maple: Found RTC at IO 0x1070
>  cpu 0x0: Vector: 700 (Program Check) at [c00000007ef83ab0]
>      pc: c00000000002e0c8: .mpic_request_ipis+0x34/0xc8
>      lr: c00000000036b484: .smp_mpic_probe+0x3c/0x58
>      sp: c00000007ef83d30
>     msr: 9000000000029032
>    current = 0xc00000000194d610
>    paca    = 0xc00000000038f180
>      pid   = 1, comm = swapper
>  kernel BUG in mpic_request_ipis at arch/powerpc/sysdev/mpic.c:1132!

Does this help?

cheers


[-- Attachment #1.2: mpic-fixup.patch --]
[-- Type: text/x-patch, Size: 1132 bytes --]

Index: to-merge/arch/powerpc/platforms/maple/setup.c
===================================================================
--- to-merge.orig/arch/powerpc/platforms/maple/setup.c
+++ to-merge/arch/powerpc/platforms/maple/setup.c
@@ -216,6 +216,7 @@ static void __init maple_init_IRQ(void)
 	int naddr, n, i, opplen, has_isus = 0;
 	struct mpic *mpic;
 	unsigned int flags = MPIC_PRIMARY;
+	char *typep;
 
 	/* Locate MPIC in the device-tree. Note that there is a bug
 	 * in Maple device-tree where the type of the controller is
@@ -226,9 +227,19 @@ static void __init maple_init_IRQ(void)
 		break;
 	}
 	if (mpic_node == NULL) {
-		printk(KERN_ERR
-		       "Failed to locate the MPIC interrupt controller\n");
-		return;
+		for_each_node_by_type(np, "interrupt-controller") {
+			typep = (char *)get_property(np, "compatible", NULL);
+			if (strstr(typep, "open-pic")) {
+				mpic_node = np;
+				break;
+			}
+		}
+
+		if (mpic_node == NULL) {
+			printk(KERN_ERR "Failed to locate the MPIC interrupt "
+				"controller\n");
+			return;
+		}
 	}
 
 	/* Find address list in /platform-open-pic */

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 191 bytes --]

Re: mpic discovery on JS20

[Segher Boessenkool]

> Does this help?

Yes it does.  I thought that went upstream already though?


Segher

Re: mpic discovery on JS20

[Amos Waterland]

On Mon, Jul 24, 2006 at 05:41:36PM +1000, Michael Ellerman wrote:
> On Thu, 2006-07-20 at 19:16 -0400, Amos Waterland wrote:
> > Current Linus and Paulus trees do this on JS20 blades with SLOF:
> > 
> >  Failed to locate the MPIC interrupt controller
> >  PID hash table entries: 4096 (order: 12, 32768 bytes)
> >  Maple: Found RTC at IO 0x1070
> >  cpu 0x0: Vector: 700 (Program Check) at [c00000007ef83ab0]
> >      pc: c00000000002e0c8: .mpic_request_ipis+0x34/0xc8
> >      lr: c00000000036b484: .smp_mpic_probe+0x3c/0x58
> >      sp: c00000007ef83d30
> >     msr: 9000000000029032
> >    current = 0xc00000000194d610
> >    paca    = 0xc00000000038f180
> >      pid   = 1, comm = swapper
> >  kernel BUG in mpic_request_ipis at arch/powerpc/sysdev/mpic.c:1132!
> 
> Does this help?

Yes, it does to some degree.  Current paulus tree with your patch applied:

 Built 1 zonelists.  Total pages: 524288
 Kernel command line: 
 mpic: Setting up MPIC " MPIC     " version <unknown> at f8040000, max 28 CPUs
 mpic: ISU size: 516, shift: 10, mask: 3ff
 Badness in mpic_init at arch/powerpc/sysdev/mpic.c:875
 Call Trace:
 [C0000000004D3880] [C00000000000EEB8] .show_stack+0x68/0x1b0 (unreliable)
 [C0000000004D3920] [C00000000001D558] .program_check_exception+0x1cc/0x598
 [C0000000004D39F0] [C0000000000044EC] program_check_common+0xec/0x100
 --- Exception: 700 at .mpic_init+0x5c/0x784
     LR = .maple_init_IRQ+0x204/0x290
 [C0000000004D3CE0] [C0000000004D3DB0] init_thread_union+0x3db0/0x4000 
 [C0000000004D3DB0] [C00000000036C130] .maple_init_IRQ+0x204/0x290
 [C0000000004D3E80] [C000000000361A88] .init_IRQ+0x34/0x48
 [C0000000004D3EF0] [C00000000035B6E0] .start_kernel+0x154/0x2b8
 [C0000000004D3F90] [C0000000000084FC] .start_here_common+0x50/0x54
 mpic: Initializing for 251 sources
 mpic: Setting up HT PICs workarounds for U3/U4
 mpic:   - HT:01.0 [0xb8] vendor 1022 device 7450 has 4 irqs
 mpic:   - HT:02.0 [0xb8] vendor 1022 device 7450 has 4 irqs
 mpic:   - HT:03.0 [0xf0] vendor 1022 device 7460 has 24 irqs
 PID hash table entries: 4096 (order: 12, 32768 bytes)
 Maple: Found RTC at IO 0x1070
 cpu 0x0: Vector: 0  at [c00000007ef83940]
     pc: c0000000000216bc: .smp_call_function+0x18c/0x1e4
     lr: c0000000000216bc: .smp_call_function+0x18c/0x1e4
     sp: c00000007ef83ab0
    msr: 9000000000009032
   current = 0xc00000000194d610
   paca    = 0xc00000000038c180
     pid   = 1, comm = swapper

The warning is this:

         WARN_ON(mpic->num_sources > MPIC_VEC_IPI_0);

Re: mpic discovery on JS20

[Amos Waterland]

[-- Attachment #1: Type: text/plain, Size: 548 bytes --]

***********************
Warning: Your file, js20-devtree.tar.gz, contains more than 32 files after decompression and cannot be scanned.
***********************


On Fri, Jul 21, 2006 at 01:37:56PM -0400, Benjamin Herrenschmidt wrote:
> On Thu, 2006-07-20 at 19:16 -0400, Amos Waterland wrote:
> > Current Linus and Paulus trees do this on JS20 blades with SLOF:
> 
> I need a tarball of /proc/device-tree on these. 

I booted 2.6.17-rc2 and tarred up /proc/device-tree.  Since the
compressed tarball is only 4KB I have attached it to this email.



[-- Attachment #2: js20-devtree.tar.gz --]
[-- Type: application/octet-stream, Size: 4747 bytes --]

Re: mpic discovery on JS20

[Michael Ellerman]

[-- Attachment #1: Type: text/plain, Size: 499 bytes --]

On Mon, 2006-07-24 at 20:20 +0200, Segher Boessenkool wrote:
> > Does this help?
> 
> Yes it does.  I thought that went upstream already though?

Doh, pays to read the mailing list huh :)

It's probably sitting in Paul's inbox, waiting for him to merge it.

cheers

-- 
Michael Ellerman
IBM OzLabs

wwweb: http://michael.ellerman.id.au
phone: +61 2 6212 1183 (tie line 70 21183)

We do not inherit the earth from our ancestors,
we borrow it from our children. - S.M.A.R.T Person

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 191 bytes --]

Re: mpic discovery on JS20

[Segher Boessenkool]

>>> Does this help?
>>
>> Yes it does.  I thought that went upstream already though?
>
> Doh, pays to read the mailing list huh :)
>
> It's probably sitting in Paul's inbox, waiting for him to merge it.

I could have sworn it went in, I don't see it anymore though.
More patches are needed anyway, I'll take care of it ASAP.


Segher