| Message ID | 20210519141436.11961-1-rppt@kernel.org (mailing list archive) |
|---|---|
| Headers | show |
| Series | memblock, arm: fixes for freeing of the memory map | expand |
On 2021/5/19 22:14, Mike Rapoport wrote: > From: Mike Rapoport <rppt@linux.ibm.com> > > Hi, > > The coordination between freeing of unused memory map, pfn_valid() and core > mm assumptions about validity of the memory map in various ranges was not > designed for complex layouts of the physical memory with a lot of holes all > over the place. > > Kefen Wang reported crashes in move_freepages() on a system with the > following memory layout: > > node 0: [mem 0x0000000080a00000-0x00000000855fffff] > node 0: [mem 0x0000000086a00000-0x0000000087dfffff] > node 0: [mem 0x000000008bd00000-0x000000008c4fffff] > node 0: [mem 0x000000008e300000-0x000000008ecfffff] > node 0: [mem 0x0000000090d00000-0x00000000bfffffff] > node 0: [mem 0x00000000cc000000-0x00000000dc9fffff] > node 0: [mem 0x00000000de700000-0x00000000de9fffff] > node 0: [mem 0x00000000e0800000-0x00000000e0bfffff] > node 0: [mem 0x00000000f4b00000-0x00000000f6ffffff] > node 0: [mem 0x00000000fda00000-0x00000000ffffefff] > > The crashes can be mitigated by enabling CONFIG_HOLES_IN_ZONE and > essentially turning pfn_valid_within() to pfn_valid() instead of having it > hardwired to 1. > > Alternatively, we can update ARM's implementation of pfn_valid() to take > into accounting rounding of the freed memory map to pageblock boundaries > and make sure it returns true for PFNs that have memory map entries even if > there is no physical memory. > > I can take the entire series via memblock tree. > > @Kefen, I didn't add your Tested-by yet because the patch is slightly > different from the version you've tested. Backport this version(also with link[2]) and oom test could pass too. > > v2: > * Use single memblock_overlaps_region() instead of several > memblock_is_map_memory() lookups. This makes this series depend on update > of MEMBLOCK_NOMAP handling in the memory map [2] > > v1: Link: https://lore.kernel.org/lkml/20210518090613.21519-1-rppt@kernel.org > > [1] https://lore.kernel.org/lkml/2a1592ad-bc9d-4664-fd19-f7448a37edc0@huawei.com > [2] https://lore.kernel.org/lkml/20210511100550.28178-1-rppt@kernel.org > > Mike Rapoport (3): > memblock: free_unused_memmap: use pageblock units instead of MAX_ORDER > memblock: align freed memory map on pageblock boundaries with SPARSEMEM > arm: extend pfn_valid to take into accound freed memory map alignment > > arch/arm/mm/init.c | 13 ++++++++++++- > mm/memblock.c | 23 ++++++++++++----------- > 2 files changed, 24 insertions(+), 12 deletions(-) > > > base-commit: d07f6ca923ea0927a1024dfccafc5b53b61cfecc >
From: Mike Rapoport <rppt@linux.ibm.com> Hi, The coordination between freeing of unused memory map, pfn_valid() and core mm assumptions about validity of the memory map in various ranges was not designed for complex layouts of the physical memory with a lot of holes all over the place. Kefen Wang reported crashes in move_freepages() on a system with the following memory layout: node 0: [mem 0x0000000080a00000-0x00000000855fffff] node 0: [mem 0x0000000086a00000-0x0000000087dfffff] node 0: [mem 0x000000008bd00000-0x000000008c4fffff] node 0: [mem 0x000000008e300000-0x000000008ecfffff] node 0: [mem 0x0000000090d00000-0x00000000bfffffff] node 0: [mem 0x00000000cc000000-0x00000000dc9fffff] node 0: [mem 0x00000000de700000-0x00000000de9fffff] node 0: [mem 0x00000000e0800000-0x00000000e0bfffff] node 0: [mem 0x00000000f4b00000-0x00000000f6ffffff] node 0: [mem 0x00000000fda00000-0x00000000ffffefff] The crashes can be mitigated by enabling CONFIG_HOLES_IN_ZONE and essentially turning pfn_valid_within() to pfn_valid() instead of having it hardwired to 1. Alternatively, we can update ARM's implementation of pfn_valid() to take into accounting rounding of the freed memory map to pageblock boundaries and make sure it returns true for PFNs that have memory map entries even if there is no physical memory. I can take the entire series via memblock tree. @Kefen, I didn't add your Tested-by yet because the patch is slightly different from the version you've tested. v2: * Use single memblock_overlaps_region() instead of several memblock_is_map_memory() lookups. This makes this series depend on update of MEMBLOCK_NOMAP handling in the memory map [2] v1: Link: https://lore.kernel.org/lkml/20210518090613.21519-1-rppt@kernel.org [1] https://lore.kernel.org/lkml/2a1592ad-bc9d-4664-fd19-f7448a37edc0@huawei.com [2] https://lore.kernel.org/lkml/20210511100550.28178-1-rppt@kernel.org Mike Rapoport (3): memblock: free_unused_memmap: use pageblock units instead of MAX_ORDER memblock: align freed memory map on pageblock boundaries with SPARSEMEM arm: extend pfn_valid to take into accound freed memory map alignment arch/arm/mm/init.c | 13 ++++++++++++- mm/memblock.c | 23 ++++++++++++----------- 2 files changed, 24 insertions(+), 12 deletions(-) base-commit: d07f6ca923ea0927a1024dfccafc5b53b61cfecc