Claude review: dma-buf: set SB_I_NOEXEC on the pseudo filesystem

[Claude Code Review Bot <claude-review@example.com>]

Overall Series Review

Subject: dma-buf: set SB_I_NOEXEC on the pseudo filesystem
Author: John Hubbard <jhubbard@nvidia.com>
Patches: 3
Reviewed: 2026-05-25T21:11:46.873174

---

This is a single-patch fix, not a series. The patch is **correct, well-motivated, and straightforward**. It fixes a real regression introduced by commit `1e7ab6f67824` ("anon_inode: rework assertions") in 6.17, where `path_noexec()` now warns (via `VFS_WARN_ON_ONCE`) when an `S_ANON_INODE` inode is encountered on a superblock that lacks `SB_I_NOEXEC`. The dma-buf pseudo filesystem uses `alloc_anon_inode()` (via `init_pseudo`) but never set `SB_I_NOEXEC`, triggering the warning on any dma-buf mmap.

The fix follows the established pattern used by every other pseudo filesystem in the kernel that issues anonymous inodes (`anon_inodes.c`, `aio.c`, `secretmem.c`, `guest_memfd.c`, `pidfs.c`, `nsfs.c`, etc.) — all set `fc->s_iflags |= SB_I_NOEXEC` in their `init_fs_context` callback.

The `Fixes:` tag and `Cc: stable` are appropriate since this is a regression fix for a new assertion.

No concerns with this patch.

---
Generated by Claude Code Patch Reviewer