On Mon, Sep 08, 2025 at 03:09:24PM +0200, David Hildenbrand wrote:
On 08.09.25 14:47, Chunyu Hu wrote:
The test will fail as below on x86_64 with cpu la57 support (will skip if no la57 support). Note, the test requries nr_hugepages to be set first.
# running bash ./va_high_addr_switch.sh # ------------------------------------- # mmap(addr_switch_hint - pagesize, pagesize): 0x7f55b60fa000 - OK # mmap(addr_switch_hint - pagesize, (2 * pagesize)): 0x7f55b60f9000 - OK # mmap(addr_switch_hint, pagesize): 0x800000000000 - OK # mmap(addr_switch_hint, 2 * pagesize, MAP_FIXED): 0x800000000000 - OK # mmap(NULL): 0x7f55b60f9000 - OK # mmap(low_addr): 0x40000000 - OK # mmap(high_addr): 0x1000000000000 - OK # mmap(high_addr) again: 0xffff55b6136000 - OK # mmap(high_addr, MAP_FIXED): 0x1000000000000 - OK # mmap(-1): 0xffff55b6134000 - OK # mmap(-1) again: 0xffff55b6132000 - OK # mmap(addr_switch_hint - pagesize, pagesize): 0x7f55b60fa000 - OK # mmap(addr_switch_hint - pagesize, 2 * pagesize): 0x7f55b60f9000 - OK # mmap(addr_switch_hint - pagesize/2 , 2 * pagesize): 0x7f55b60f7000 - OK # mmap(addr_switch_hint, pagesize): 0x800000000000 - OK # mmap(addr_switch_hint, 2 * pagesize, MAP_FIXED): 0x800000000000 - OK # mmap(NULL, MAP_HUGETLB): 0x7f55b5c00000 - OK # mmap(low_addr, MAP_HUGETLB): 0x40000000 - OK # mmap(high_addr, MAP_HUGETLB): 0x1000000000000 - OK # mmap(high_addr, MAP_HUGETLB) again: 0xffff55b5e00000 - OK # mmap(high_addr, MAP_FIXED | MAP_HUGETLB): 0x1000000000000 - OK # mmap(-1, MAP_HUGETLB): 0x7f55b5c00000 - OK # mmap(-1, MAP_HUGETLB) again: 0x7f55b5a00000 - OK # mmap(addr_switch_hint - pagesize, 2*hugepagesize, MAP_HUGETLB): 0x800000000000 - FAILED # mmap(addr_switch_hint , 2*hugepagesize, MAP_FIXED | MAP_HUGETLB): 0x800000000000 - OK # [FAIL]
addr_switch_hint is defined as DFEFAULT_MAP_WINDOW in the failed test (for x86_64, DFEFAULT_MAP_WINDOW is defined as (1UL<<47) - pagesize) in 64 bit.
Before commit cc92882ee218 ("mm: drop hugetlb_get_unmapped_area{_*} functions"), for x86_64 hugetlb_get_unmapped_area() is handled in arch code arch/x86/mm/hugetlbpage.c and addr is checked with map_address_hint_valid() after align with 'addr &= huge_page_mask(h)' which is a round down way, and it will fail the check because the addr is within the DEFAULT_MAP_WINDOW but (addr + len) is above the DFEFAULT_MAP_WINDOW. So it wil go through the hugetlb_get_unmmaped_area_top_down() to find an area within the DFEFAULT_MAP_WINDOW.
After commit cc92882ee218 ("mm: drop hugetlb_get_unmapped_area{_*} functions"). The addr hint for hugetlb_get_unmmaped_area() will be rounded up and aligned to hugepage size with ALIGN() for all arches. And after the align, the addr will be above the default MAP_DEFAULT_WINDOW, and the map_addresshint_valid() check will pass because both aligned addr (addr0) and (addr + len) are above the DEFAULT_MAP_WINDOW, and the aligned hint address (0x800000000000) is returned as an suitable gap is found there, in arch_get_unmapped_area_topdown().
To still cover the case that addr is within the DEFAULT_MAP_WINDOW, and addr + len is above the DFEFAULT_MAP_WINDOW, make the addr hint one hugepage lower, so that after the align it's still within DEFAULT_MAP_WINDOW, and the addr + len (2 hugepages) will be above DEFAULT_MAP_WINDOW.
Fixes: cc92882ee218 ("mm: drop hugetlb_get_unmapped_area{_*} functions") Signed-off-by: Chunyu Hu chuhu@redhat.com
tools/testing/selftests/mm/va_high_addr_switch.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/mm/va_high_addr_switch.c b/tools/testing/selftests/mm/va_high_addr_switch.c index 896b3f73fc53..bd96dc3b5931 100644 --- a/tools/testing/selftests/mm/va_high_addr_switch.c +++ b/tools/testing/selftests/mm/va_high_addr_switch.c @@ -230,10 +230,10 @@ void testcases_init(void) .msg = "mmap(-1, MAP_HUGETLB) again", }, {
.addr = (void *)(addr_switch_hint - pagesize),
.addr = (void *)(addr_switch_hint - pagesize - hugepagesize),Wouldn't it be more deterministic to do the alignment/rounding ourselves?
(void *)(ALIGN_DOWN(addr_switch_hint - pagesize), hugepagesize)
Unfortunately we don't have an ALIGN_DOWN helper available yet.
We could just move the one in pkey-helpers.h into vm_util.h
Thanks for the review!
This is good idea and it would be more deterministic if we provide an aligned address directly, then the kernel change won't affect the test.
But now I realize that, likely,
.addr = (void *)(addr_switch_hint - hugepagesize),
would just work and be aligned?
Yes, it's aligned to the hugepagesize, align down and works. I prefer this way as it's easier and all other tests in the file do like this. Thanks for the suggestion.
I thought we would lose some test coverage on testing if it will work when an un-hugepagesize aligned addr is provided. Do you think it's necessary? If not, I'll change to:
.addr = (void *)(addr_switch_hint - hugepagesize),
or we can add them both if necesasry.
-- Cheers
David / dhildenb