As part of the efforts to eliminate the problematic f_op->mmap callback, a new callback - f_op->mmap_prepare was provided. While we are converting these callbacks, we must deal with 'stacked' filesystems and drivers - those which in their own f_op->mmap callback invoke an inner f_op->mmap callback. To accomodate for this, a compatibility layer is provided that, via vfs_mmap(), detects if f_op->mmap_prepare is provided and if so, generates a vm_area_desc containing the VMA's metadata and invokes the call. So far, we have provided desc->file equal to vma->vm_file. However this is not necessarily valid, especially in the case of stacked drivers which wish to assign a new file after the inner hook is invoked. To account for this, we adjust vm_area_desc to have both file and vm_file fields. The .vm_file field is strictly set to vma->vm_file (or in the case of a new mapping, what will become vma->vm_file). However, .file is set to whichever file vfs_mmap() is invoked with when using the compatibilty layer. Therefore, if the VMA's file needs to be updated in .mmap_prepare, desc->vm_file should be assigned, whilst desc->file should be read. No current f_op->mmap_prepare users assign desc->file so this is safe to do. This makes the .mmap_prepare callback in the context of a stacked filesystem or driver completely consistent with the existing .mmap implementations. While we're here, we do a few small cleanups, and ensure that we const-ify things correctly in the vm_area_desc struct to avoid hooks accidentally trying to assign fields they should not. v2: * Refer to 'stacked' mmap callers as per Brauner. * Updated comments etc. as per Liam. * Add new vm_file field intended to be mutable as a result of discussion with Liam. * Made desc->file a 'const struct file *const' pointer so a user won't mistakenly assign to it. * While we're here - Made desc->mm const (and updated the one case where desc->mm being const matters - mlock_future_ok(), invoked by secretmem_mmap_prepare()) and also set const for assignment to ensure hook user doesn't incorrectly consider this field to be updateable. * Separated changes to VMA descriptor logic into separate commit. * Consistently refer to 'filesystems' not 'file systems'. * Rearranged checks in set_vma_from_desc() putting the possibly desc->xxx field first as this seems clearer. * Fix the kerneldoc error for __compat_vma_mmap_prepare(). * Update comments as necessary. v1: https://lore.kernel.org/all/20250902104533.222730-1-lorenzo.stoakes@xxxxxxxxxx/ Lorenzo Stoakes (2): mm: specify separate file and vm_file params in vm_area_desc mm: do not assume file == vma->vm_file in compat_vma_mmap_prepare() include/linux/fs.h | 2 ++ include/linux/mm_types.h | 5 +-- mm/internal.h | 4 +-- mm/mmap.c | 2 +- mm/util.c | 52 ++++++++++++++++++++++++-------- mm/vma.c | 5 +-- mm/vma.h | 28 +++-------------- tools/testing/vma/vma_internal.h | 36 +++++++++++++++------- 8 files changed, 79 insertions(+), 55 deletions(-) -- 2.50.1
