Re: [PATCH v2 10/44] mmap: Change do_mas_munmap and do_mas_aligned_munmap() to use vma iterator

[Sven Schnelle <svens@linux.ibm.com>]

Liam Howlett <liam.howlett@oracle.com> writes:

> * Sven Schnelle <svens@linux.ibm.com> [230110 09:54]:
>> Liam Howlett <liam.howlett@oracle.com> writes:
>> 
>> > From: "Liam R. Howlett" <Liam.Howlett@Oracle.com>
>> >
>> > Start passing the vma iterator through the mm code.  This will allow for
>> > reuse of the state and cleaner invalidation if necessary.
>> >
>> > Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
>> > ---
>> >  include/linux/mm.h |  2 +-
>> >  mm/mmap.c          | 77 +++++++++++++++++++++-------------------------
>> >  mm/mremap.c        |  6 ++--
>> >  3 files changed, 39 insertions(+), 46 deletions(-)
>> >
>> 
>> Starting with this patch i see the following oops on s390:
>> [..]
>> This happens on every boot, always killing the init process. The oops
>> doesn't happen with next-20230110. With next-20230110 i see shmat
>> testcase failures in ltp (shmat returning with -EINVAL because
>> find_vma_intersection() tells shmat that there's already a mapping
>> present).
>> 
>> Trying to bisect that i stumbled above the oops above. Any ideas before
>> i start trying to understand the patch?
>
> Yes, try the patch for fixing the invalidated state I sent out yesterday
> [1].  This should come before ("mm: expand vma iterator interface").
>
> 1. https://lore.kernel.org/linux-mm/20230109165455.647400-1-Liam.Howlett@oracle.com/

Thanks, missed that. I can report that the crash i've seen seems to be
fixed. Also the shmat01 testcase in ltp is working now.