block: Use logical block size for alignment

O_DIRECT requires buffer addresses to be aligned to the backend
device's logical block size. The existing bounce buffer logic in
execute_async() hardcodes SECTOR_SIZE (512) for the alignment check
and bounce buffer allocation. This is insufficient for devices with
a 4096-byte logical block size, where misaligned buffers cause
-EINVAL from the host kernel.

Add an alignment() method to the AsyncIo trait that returns the
backend's logical block size, defaulting to SECTOR_SIZE. The three
raw I/O backends (io_uring, AIO, synchronous) probe the device
topology via DiskTopology::probe() at creation time and return the
actual logical block size. All image format backends would simply
use the default value of 512 bytes since their underlying are
not block devices.

execute_async() now queries disk_image.alignment() instead of using
the hardcoded SECTOR_SIZE

Fixes: #7720

Signed-off-by: Saravanan D <saravanand@crusoe.ai>
This commit is contained in:
Saravanan D
2026-02-18 14:23:03 -08:00
committed by Rob Bradford
parent 272fa624ef
commit 00c05f4761
5 changed files with 55 additions and 20 deletions
+8 -6
View File
@@ -436,6 +436,7 @@ impl Request {
let sector = self.sector;
let request_type = self.request_type;
let offset = (sector << SECTOR_SHIFT) as libc::off_t;
let alignment = disk_image.alignment();
let mut iovecs: SmallVec<[libc::iovec; DEFAULT_DESCRIPTOR_VEC_SIZE]> =
SmallVec::with_capacity(self.data_descriptors.len());
@@ -466,14 +467,15 @@ impl Request {
assert!(origin_ptr.len() >= data_len);
let origin_ptr = origin_ptr.ptr_guard();
// Verify the buffer alignment.
// In case it's not properly aligned, an intermediate buffer is
// created with the correct alignment, and a copy from/to the
// origin buffer is performed, depending on the type of operation.
let iov_base = if (origin_ptr.as_ptr() as u64).is_multiple_of(SECTOR_SIZE) {
// O_DIRECT requires buffer addresses to be aligned to the
// backend device's logical block size. In case it's not properly
// aligned, an intermediate buffer is created with the correct
// alignment, and a copy from/to the origin buffer is performed,
// depending on the type of operation.
let iov_base = if (origin_ptr.as_ptr() as u64).is_multiple_of(alignment) {
origin_ptr.as_ptr() as *mut libc::c_void
} else {
let layout = Layout::from_size_align(data_len, SECTOR_SIZE as usize).unwrap();
let layout = Layout::from_size_align(data_len, alignment as usize).unwrap();
// SAFETY: layout has non-zero size
let aligned_ptr = unsafe { alloc_zeroed(layout) };
if aligned_ptr.is_null() {