CVE-2026-98216

In the Linux kernel, the following vulnerability has been resolved:

IB/hfi1: Fix the PIO CRED credit-return mmap

hfi1 file mmap()'s PIO CRED case must hand user space the single credit-return page that holds this context's entry. That page is the second or third page of the per-node credit-return allocation once the hardware send context index reaches 64 or 128, so the failure below is intermittent: when the entry lands on the first page the offset is zero and everything works.

Two things are wrong.

First, cr page offset is a byte offset but .va is a struct credit return *, so adding it is pointer arithmetic and scales the offset by sizeof(struct credit return) == 64. memvirt then lands 256 KiB or 512 KiB past a 10240-byte allocation. With an IOMMU translating, that address is inside the vmalloc range but in no vm area, so dma mmap coherent() -> iommu dma mmap() finds no pages, vmalloc to pfn() returns page to pfn(NULL), and remap pfn range() installs a frame above MAXPHYADDR. The first user read then takes:

psm2 ep open pr: Corrupted page table at address 7a14d007e000 PGD 800000013886a067 P4D 800000013886a067 PUD 13886b067 PMD 13886c067 PTE 800049168e911235 Oops: Bad pagetable: 000d [#1] SMP PTI

Second, and still wrong once the arithmetic is corrected, dma mmap coherent() describes a whole coherent buffer and selects the page within it with vma->vm pgoff. Offsetting cpu addr has no effect: for a vmap'd allocation iommu dma mmap() uses cpu addr only to locate the vm area and then maps pages[vm pgoff], which hfi1 file mmap() has just set to 0. User space therefore always receives the first credit-return page, every credit read is for the wrong context, and send PIO stalls forever.

Use the DMA API as intended: pass the base of the allocation with its full length and select the page with vm pgoff. A separate length is needed because memlen must keep describing the VMA for the existing size check. The dma-direct path stays correct as well, since dma direct mmap() adds the same vm pgoff to the base pfn.

Tested on a Dell T7610 (Xeon E5-2650 v2, Intel IOMMU in DMA-FQ mode) against a Threadripper PRO 3995WX peer, both Omni-Path 100. Before this change psm2 ep open() Oopses the kernel; with only the arithmetic corrected psm2 ep open() succeeds but any transfer that uses send PIO hangs, PSM2 SDMA=2 (send PIO disabled) completing normally while PSM2 SDMA=0 (send PIO only) hangs every time. With this change send PIO, send DMA and the default mixed mode all work.