On 9/15/26 2:56 AM, Muchun Song wrote:
> While working on the HugeTLB Vmemmap Optimization (HVO) generalization,
> Sashiko reported that memmap_init_zone_device() could be asked to
> initialize struct pages with a large-page order for a physical range
> that was not naturally aligned to that order.
>
> Following the Device DAX call path confirmed that the condition was
> reachable. Device DAX range validation checks the range size, but not
> its start address. An unaligned range can therefore reach
> memmap_init_zone_device() after the device binds, and a subsequent write
> to a userspace mapping may trigger a kernel panic. With the help of an
> LLM, I was able to reproduce the failure.
>
> The automatic resize path has a related problem. It can split an aligned
> size request across arbitrary free gaps, consume an unaligned fragment,
> and leave the resize partially applied when a later allocation fails.
>
> This series validates both the start and size of every Device DAX range
> and makes automatic resize account only for usable aligned space. It
> also fixes the mapping sysfs attribute to propagate validation failures
> instead of reporting success without allocating the requested range.
>
> Muchun Song (2):
> dax/bus: fix Device DAX range alignment validation
> dax/bus: fix mapping attribute error reporting
>
> drivers/dax/bus.c | 144 +++++++++++++++++++++++++++++++++++-----------
> 1 file changed, 109 insertions(+), 35 deletions(-)
>
>
> base-commit: 1a1de54f7369cd2b5bac0f265910e60ad3a6b4c3
This commit does not exist in Linus's tree. Can you please base the series on
top of the latest 7.3-rc tag so sashiko can look over it? Thanks!
