engines/mmap: fix full/limited prep logic

The engine supports a "full" mode where the entire range for a job is mapped,
and a "limited" mode that tries to map less memory. This fixes a bug and
implements the following simplified logic:

- If we are on a 64-bit architecture, or if io_size <= mmap_map_size, then we
  map the entire io_size region, and re-use for all IO.
- Otherwise, we map the range for each individual IOs.

Previously, we would fallback to limited mode in case of 32-bit overflow of the
job size. That isn't strict enough as we want to stay below mmap_map_size for a
single mapping. This fixes that.

Also, in limited mode, this code previously tried to map min(io_size,
mmap_map_size) bytes, starting at the offset. This could allow a future IO use
the same mapping if happens at a higher offset. I think it's better to keep it
simple and get the consistent performance of always having one mmap per IO in
limited mode.

Signed-off-by: Geert Jansen <[email protected]>

Author: geertj

engines/mmap.c

@@ -15,7 +15,7 @@
 #include "../verify.h"
 
 /*
- * Limits us to 1GiB of mapped files in total
+ * Limits us to 1GiB of mapped files in total on 32-bit architectures
  */
 #define MMAP_TOTAL_SZ	(1 * 1024 * 1024 * 1024UL)
 
@@ -157,11 +157,8 @@ static int fio_mmapio_prep_limited(struct thread_data *td, struct io_u *io_u)
 		return EIO;
 	}
 
-	fmd->mmap_sz = mmap_map_size;
-	if (fmd->mmap_sz  > f->io_size)
-		fmd->mmap_sz = f->io_size;
-
 	fmd->mmap_off = io_u->offset;
+	fmd->mmap_sz = io_u->buflen;
 
 	return fio_mmap_file(td, f, fmd->mmap_sz, fmd->mmap_off);
 }
@@ -177,8 +174,8 @@ static int fio_mmapio_prep_full(struct thread_data *td, struct io_u *io_u)
 
 	if (fio_file_partial_mmap(f))
 		return EINVAL;
-	if (io_u->offset != (size_t) io_u->offset ||
-	    f->io_size != (size_t) f->io_size) {
+
+	if (sizeof(size_t) < 8 && f->io_size > mmap_map_size) {
 		fio_file_set_partial_mmap(f);
 		return EINVAL;
 	}