Hi,
[ Thanks to David Hildenbrand for identifying the root cause of this
issue and proving guidance on how to fix it. The new API idea, bugs
and misconceptions are all mine though ]
Currently, trying to reserve 1G pages with page_owner=on and sparsemem
causes a crash. The reproducer is very simple:
1. Build the kernel with CONFIG_SPARSEMEM=y and the table extensions
2. Pass 'default_hugepagesz=1 page_owner=on' in the kernel command-line
3. Reserve one 1G page at run-time, this should crash (see patch 1 for
backtrace)
[ A crash with page_table_check is also possible, but harder to
trigger ]
Apparently, starting with commit e98337d11bbd ("mm/contig_alloc: support
__GFP_COMP") we now pass the full allocation order to page extension
clients and the page extension implementation assumes that all PFNs of
an allocation range will be stored in the same memory section (which is
not true for 1G pages).
To fix this, this series introduces a new iteration API for page extension
objects. The API ensures that we always lookup the memory sections for
the next page extension in the iteration.
While this series seems to fix the issue, it's RFC because:
1. page_ext_iter_next() uses brute-force. David suggested to maintain
the current API but only do a memory section lookup if the next PFN
aligns with PAGE_PER_SECTION (ie. it's the first PFN of a section).
But I couldn't make it work: I was getting random crashes and RCU
warnings
2. It's very lightly tested
Luiz Capitulino (4):
mm: page_ext: add an iteration API for page extensions
mm: page_owner: use new iteration API
mm: page_table_check: use new iteration API
mm: page_ext: drop page_ext_next()
include/linux/page_ext.h | 15 +++++----
mm/page_ext.c | 55 ++++++++++++++++++++++++++++++++
mm/page_owner.c | 68 ++++++++++++++++++++++++++--------------
mm/page_table_check.c | 21 +++++++------
4 files changed, 120 insertions(+), 39 deletions(-)
--
2.47.1
