Mac Mini G4 hang on boot with git master

Hi all,

I've just done a git pull and rebuilt master on my Mac Mini G4 in order to test
Michael's merge of my KVM PR fix, and unfortunately my kernel now hangs on boot :(

My last working git checkout was somewhere around the 5.0-rc stage, so I suspect it's
something that's been merged for 5.1.

The hang occurs just after the boot console is disabled which makes me wonder if
something is going wrong during PCI bus enumeration. Does anyone have an idea as to
what may be causing this? I can obviously bisect it down, but on slow hardware it can
take some time...


ATB,

Mark.

Mark,

On Fri, Mar 15, 2019 at 3:08 PM Mark Cave-Ayland
[off-list ref] wrote:

Hi all,

I've just done a git pull and rebuilt master on my Mac Mini G4 in order to test
Michael's merge of my KVM PR fix, and unfortunately my kernel now hangs on boot :(

Ouch :(

My last working git checkout was somewhere around the 5.0-rc stage, so I suspect it's
something that's been merged for 5.1.

OK. My last kernel is also somewhere around here on same hardware.

The hang occurs just after the boot console is disabled which makes me wonder if
something is going wrong during PCI bus enumeration. Does anyone have an idea as to
what may be causing this? I can obviously bisect it down, but on slow hardware it can
take some time...

When doing git bisect I compile from my amd64 machine using:

make O=g4 ARCH=powerpc CROSS_COMPILE=powerpc-linux-gnu- my_defconfig
make -j8 O=g4 ARCH=powerpc CROSS_COMPILE=powerpc-linux-gnu- bindeb-pkg
scp *image*.deb macminig4:

On Debian simply install:

# apt-get install crossbuild-essential-powerpc

ATB,

Mark.

On 15/03/2019 13:37, Mark Cave-Ayland wrote:

Hi all,

I've just done a git pull and rebuilt master on my Mac Mini G4 in order to test
Michael's merge of my KVM PR fix, and unfortunately my kernel now hangs on boot :(

My last working git checkout was somewhere around the 5.0-rc stage, so I suspect it's
something that's been merged for 5.1.

The hang occurs just after the boot console is disabled which makes me wonder if
something is going wrong during PCI bus enumeration. Does anyone have an idea as to
what may be causing this? I can obviously bisect it down, but on slow hardware it can
take some time...

This was a weird one: bisecting directly from git master gave a nonsense result,
however by manually rebasing Michael's PR onto my last known good commit from master
I was able to finally pin it down to this commit:


7a0d6955f3f7a4250da63d528bfff7a9c91b5725 is the first bad commit
commit 7a0d6955f3f7a4250da63d528bfff7a9c91b5725
Author: Christophe Leroy [off-list ref]
Date:   Thu Feb 21 10:37:55 2019 +0000

    powerpc/6xx: Store PGDIR physical address in a SPRG

    Use SPRN_SPRG2 to store the current thread PGDIR and
    avoid reading thread_struct.pgdir at every TLB miss.

    Signed-off-by: Christophe Leroy [off-list ref]
    Signed-off-by: Michael Ellerman [off-list ref]


ATB,

Mark.

Le 17/03/2019 à 15:15, Mark Cave-Ayland a écrit :

On 15/03/2019 13:37, Mark Cave-Ayland wrote:

quoted

Hi all,

I've just done a git pull and rebuilt master on my Mac Mini G4 in order to test
Michael's merge of my KVM PR fix, and unfortunately my kernel now hangs on boot :(

My last working git checkout was somewhere around the 5.0-rc stage, so I suspect it's
something that's been merged for 5.1.

The hang occurs just after the boot console is disabled which makes me wonder if
something is going wrong during PCI bus enumeration. Does anyone have an idea as to
what may be causing this? I can obviously bisect it down, but on slow hardware it can
take some time...

This was a weird one: bisecting directly from git master gave a nonsense result,
however by manually rebasing Michael's PR onto my last known good commit from master
I was able to finally pin it down to this commit:


7a0d6955f3f7a4250da63d528bfff7a9c91b5725 is the first bad commit
commit 7a0d6955f3f7a4250da63d528bfff7a9c91b5725
Author: Christophe Leroy [off-list ref]
Date:   Thu Feb 21 10:37:55 2019 +0000

     powerpc/6xx: Store PGDIR physical address in a SPRG

     Use SPRN_SPRG2 to store the current thread PGDIR and
     avoid reading thread_struct.pgdir at every TLB miss.

     Signed-off-by: Christophe Leroy [off-list ref]
     Signed-off-by: Michael Ellerman [off-list ref]

Hi,

The fix is there:

https://patchwork.ozlabs.org/patch/1053385/

Christophe

---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus

On 17/03/2019 16:25, christophe leroy wrote:

quoted

This was a weird one: bisecting directly from git master gave a nonsense result,
however by manually rebasing Michael's PR onto my last known good commit from master
I was able to finally pin it down to this commit:


7a0d6955f3f7a4250da63d528bfff7a9c91b5725 is the first bad commit
commit 7a0d6955f3f7a4250da63d528bfff7a9c91b5725
Author: Christophe Leroy [off-list ref]
Date:   Thu Feb 21 10:37:55 2019 +0000

     powerpc/6xx: Store PGDIR physical address in a SPRG

     Use SPRN_SPRG2 to store the current thread PGDIR and
     avoid reading thread_struct.pgdir at every TLB miss.

     Signed-off-by: Christophe Leroy [off-list ref]
     Signed-off-by: Michael Ellerman [off-list ref]

Hi,

The fix is there:

https://patchwork.ozlabs.org/patch/1053385/

Christophe

Hi Christophe,

Thank you! I've tried the patch here and have confirmed that it fixes the problem
with the hang on boot.


ATB,

Mark.