On 11/18/2019 11:17 AM, Michael Ellerman wrote:

Christophe Leroy [off-list ref] writes:

quoted

When not using large TLBs, the IMMR region is still
mapped as a whole block in the FIXMAP area.

Do not remove pages mapped in the FIXMAP region when
initialising paging.

Properly report that the IMMR region is block-mapped even
when not using large TLBs.

Signed-off-by: Christophe Leroy <redacted>
---
  arch/powerpc/mm/mem.c        |  8 --------
  arch/powerpc/mm/nohash/8xx.c | 13 +++++++------
  2 files changed, 7 insertions(+), 14 deletions(-)

This blows up pmac32_defconfig + qemu mac99 for me with:

   NET: Registered protocol family 1
   RPC: Registered named UNIX socket transport module.
   RPC: Registered udp transport module.
   RPC: Registered tcp transport module.
   RPC: Registered tcp NFSv4.1 backchannel transport module.
   PCI: CLS 0 bytes, default 32
   Trying to unpack rootfs image as initramfs...
   BUG: Unable to handle kernel data access on write at 0xfffdf000

I tested it with pmac32_defconfig and qemu mac99 and don't get the problem:

NET: Registered protocol family 1
RPC: Registered named UNIX socket transport module.
RPC: Registered udp transport module.
RPC: Registered tcp transport module.
RPC: Registered tcp NFSv4.1 backchannel transport module.
PCI: CLS 0 bytes, default 32
Initialise system trusted keyrings
workingset: timestamp_bits=30 max_order=15 bucket_order=0
NFS: Registering the id_resolver key type
Key type id_resolver registered
...

Looks like I don't get that 'Trying to unpack rootfs image as 
initramfs...', do you change anything to pmac32_defconfig ?

Anyway, when rebasing this patch on next branch, only the 
arch/powerpc/mm/nohash/8xx.c change remains. The other part is already 
applied through another patch.

So I believe the remaining part is safe to apply

Christophe