Fix abd leak, kmem_free correct size of abd_t

Fix a leak of abd_t that manifested mostly when using
raidzN with at least as many columns as N (e.g. a
four-disk raidz2 but not a three-disk raidz2).
Sufficiently heavy raidz use would eventually run a system
out of memory.

Additionally:

* Switch abd_cache arena to FIRSTFIT, which empirically
improves perofrmance.

* Make abd_chunk_cache more performant and debuggable.

* Allocate the abd_zero_buf from abd_chunk_cache rather
than the heap.

* Don't try to reap non-existent qcaches in abd_cache arena.

* KM_PUSHPAGE->KM_SLEEP when allocating chunks from their
own arena

Reviewed-by: Matthew Ahrens <[email protected]>
Reviewed-by: Alexander Motin <[email protected]>
Signed-off-by: Jorgen Lundman <[email protected]>
Co-authored-by: Sean Doran <[email protected]>
Closes #12295

1 files changed, 6 insertions, 3 deletions

diff --git a/module/os/freebsd/zfs/abd_os.c b/module/os/freebsd/zfs/abd_os.c
index 47adc2278..4f5b33d94 100644
--- a/module/os/freebsd/zfs/abd_os.c
+++ b/module/os/freebsd/zfs/abd_os.c
@@ -374,14 +374,17 @@ abd_alloc_for_io(size_t size, boolean_t is_metadata)
 }
 
 abd_t *
-abd_get_offset_scatter(abd_t *abd, abd_t *sabd, size_t off)
+abd_get_offset_scatter(abd_t *abd, abd_t *sabd, size_t off,
+    size_t size)
 {
 	abd_verify(sabd);
 	ASSERT3U(off, <=, sabd->abd_size);
 
 	size_t new_offset = ABD_SCATTER(sabd).abd_offset + off;
-	uint_t chunkcnt = abd_scatter_chunkcnt(sabd) -
-	    (new_offset / zfs_abd_chunk_size);
+	size_t chunkcnt = abd_chunkcnt_for_bytes(
+	    (new_offset % zfs_abd_chunk_size) + size);
+
+	ASSERT3U(chunkcnt, <=, abd_scatter_chunkcnt(sabd));
 
 	/*
 	 * If an abd struct is provided, it is only the minimum size.  If we