CVE-2026-46325

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

RDMA/rxe: Fix iova-to-va conversion for MR page sizes != PAGE_SIZE

The current implementation incorrectly handles memory regions (MRs) with

page sizes different from the system PAGE_SIZE. The core issue is that

rxesetpage() is called with mr->page_size step increments, but the

page_list stores individual struct page pointers, each representing

PAGE_SIZE of memory.

ibsgto_page() has ensured that when i>=1 either

a) SG[i-1].dmaend and SG[i].dmaaddr are contiguous

or

b) SG[i-1].dmaend and SG[i].dmaaddr are mr->page_size aligned.

This leads to incorrect iova-to-va conversion in scenarios:

1. pagesize < PAGESIZE (e.g., MR: 4K, system: 64K):

   ibmr->iova = 0x181800

   sg[0]: dma_addr=0x181800, len=0x800

   sg[1]: dma_addr=0x173000, len=0x1000

   Access iova = 0x181800 + 0x810 = 0x182010

   Expected VA: 0x173010 (second SG, offset 0x10)

   Before fix:

• • index = (0x182010 >> 12) - (0x181800 >> 12) = 1
  • page_offset = 0x182010 & 0xFFF = 0x10
  • xarray[1] stores system page base 0x170000
  • Resulting VA: 0x170000 + 0x10 = 0x170010 (wrong)

1. pagesize > PAGESIZE (e.g., MR: 64K, system: 4K):

   ibmr->iova = 0x18f800

   sg[0]: dma_addr=0x18f800, len=0x800

   sg[1]: dma_addr=0x170000, len=0x1000

   Access iova = 0x18f800 + 0x810 = 0x190010

   Expected VA: 0x170010 (second SG, offset 0x10)

   Before fix:

• • index = (0x190010 >> 16) - (0x18f800 >> 16) = 1
  • page_offset = 0x190010 & 0xFFFF = 0x10
  • xarray[1] stores system page for dma_addr 0x170000
  • Resulting VA: system page of 0x170000 + 0x10 = 0x170010 (wrong)

Yi Zhang reported a kernel panic[1] years ago related to this defect.

Solution:

1. Replace xarray with pre-allocated rxemrpage array for sequential

   indexing (all MR page indices are contiguous)

1. Each rxemrpage stores both struct page* and offset within the

   system page

1. Handle MR pagesize != PAGESIZE relationships:

• pagesize > PAGESIZE: Split MR pages into multiple system pages
• pagesize <= PAGESIZE: Store offset within system page

1. Add boundary checks and compatibility validation

This ensures correct iova-to-va conversion regardless of MR page size

and system PAGE_SIZE relationship, while improving performance through

array-based sequential access.

Tests on 4K and 64K PAGE_SIZE hosts:

• rdma-core/pytests

  $ ./build/bin/runtests.py  --dev eth0rxe

• blktest:

  $ TIMEOUT=30 QUICKRUN=1 USERXE=1 NVMET_TRTYPES=rdma ./check nvme srp rnbd

[1] https://lore.kernel.org/all/CAHj4cs9XRqE25jyVw9rj9YugffLn5+f=1znaBEnu1usLOciD+g@mail.gmail.com/T/