Fix bug caused by rounding in vdev_raidz_asize_to_psize

When an allocation is happening on a raidz vdev, the number of sectors
allocated is rounded up to a multiple of nparity + 1. If this results in
the allocation spilling into an extra row, then the corresponding call
to vdev_raidz_asize_to_psize will incorrectly assume that parity sectors
were allocated for that spilled row, even though no data is stored
there.

If we determine that happened, we need to subtract out those extra
sectors before performing the rest of the capacity calculation.

Signed-off-by: Paul Dagnelie <[email protected]>
Sponsored-by: Klara, Inc.
Sponsored-by: Wasabi Technology, Inc.

Author: pcd1193182

module/zfs/vdev_raidz.c

@@ -2257,6 +2257,14 @@ vdev_raidz_asize_to_psize(vdev_t *vd, uint64_t asize, uint64_t txg)
 	ASSERT0(asize % (1 << ashift));
 
 	psize = (asize >> ashift);
+	/*
+	 * If the roundup to nparity + 1 caused us to spill into a new row, we
+	 * need to ignore that row entirely (since it can't store data or
+	 * parity).
+	 */
+	uint64_t rows = psize / cols;
+	psize = psize - (rows * cols) <= nparity ? rows * cols : psize;
+	/*  Subtract out parity sectors for each row storing data. */
 	psize -= nparity * DIV_ROUND_UP(psize, cols);
 	psize <<= ashift;
 

module/zfs/zio.c

@@ -3239,6 +3239,7 @@ zio_write_gang_block(zio_t *pio, metaslab_class_t *mc)
 
 		uint64_t psize = allocated ? MIN(resid, allocated_size) :
 		    min_size;
+		ASSERT3U(psize, >=, min_size);
 
 		zio_t *cio = zio_write(zio, spa, txg, bp, has_data ?
 		    abd_get_offset(pio->io_abd, pio->io_size - resid) : NULL,