Merge drm/drm-fixes into drm-misc-fixes

Backmerging to get fixes from v6.18-rc6.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>

5 files changed, 117 insertions, 16 deletions

diff --git a/fs/xfs/libxfs/xfs_rtgroup.h b/fs/xfs/libxfs/xfs_rtgroup.h
index d36a6ae0abe5..d4fcf591e63d 100644
--- a/fs/xfs/libxfs/xfs_rtgroup.h
+++ b/fs/xfs/libxfs/xfs_rtgroup.h
@@ -50,6 +50,12 @@ struct xfs_rtgroup {
 		uint8_t			*rtg_rsum_cache;
 		struct xfs_open_zone	*rtg_open_zone;
 	};
+
+	/*
+	 * Count of outstanding GC operations for zoned XFS.  Any RTG with a
+	 * non-zero rtg_gccount will not be picked as new GC victim.
+	 */
+	atomic_t		rtg_gccount;
 };
 
 /*
diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
index ee49f20875af..6917de832191 100644
--- a/fs/xfs/xfs_discard.c
+++ b/fs/xfs/xfs_discard.c
@@ -726,8 +726,10 @@ xfs_trim_rtgroup_extents(
 			break;
 		}
 
-		if (!tr.queued)
+		if (!tr.queued) {
+			kfree(tr.extents);
 			break;
+		}
 
 		/*
 		 * We hand the extent list to the discard function here so the
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index d3f6e3e42a11..490e12cb99be 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -1091,6 +1091,29 @@ const struct iomap_ops xfs_zoned_direct_write_iomap_ops = {
 };
 #endif /* CONFIG_XFS_RT */
 
+#ifdef DEBUG
+static void
+xfs_check_atomic_cow_conversion(
+	struct xfs_inode		*ip,
+	xfs_fileoff_t			offset_fsb,
+	xfs_filblks_t			count_fsb,
+	const struct xfs_bmbt_irec	*cmap)
+{
+	struct xfs_iext_cursor		icur;
+	struct xfs_bmbt_irec		cmap2 = { };
+
+	if (xfs_iext_lookup_extent(ip, ip->i_cowfp, offset_fsb, &icur, &cmap2))
+		xfs_trim_extent(&cmap2, offset_fsb, count_fsb);
+
+	ASSERT(cmap2.br_startoff == cmap->br_startoff);
+	ASSERT(cmap2.br_blockcount == cmap->br_blockcount);
+	ASSERT(cmap2.br_startblock == cmap->br_startblock);
+	ASSERT(cmap2.br_state == cmap->br_state);
+}
+#else
+# define xfs_check_atomic_cow_conversion(...)	((void)0)
+#endif
+
 static int
 xfs_atomic_write_cow_iomap_begin(
 	struct inode		*inode,
@@ -1102,9 +1125,10 @@ xfs_atomic_write_cow_iomap_begin(
 {
 	struct xfs_inode	*ip = XFS_I(inode);
 	struct xfs_mount	*mp = ip->i_mount;
-	const xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
-	xfs_fileoff_t		end_fsb = xfs_iomap_end_fsb(mp, offset, length);
-	xfs_filblks_t		count_fsb = end_fsb - offset_fsb;
+	const xfs_fileoff_t	offset_fsb = XFS_B_TO_FSBT(mp, offset);
+	const xfs_fileoff_t	end_fsb = XFS_B_TO_FSB(mp, offset + length);
+	const xfs_filblks_t	count_fsb = end_fsb - offset_fsb;
+	xfs_filblks_t		hole_count_fsb;
 	int			nmaps = 1;
 	xfs_filblks_t		resaligned;
 	struct xfs_bmbt_irec	cmap;
@@ -1130,7 +1154,7 @@ xfs_atomic_write_cow_iomap_begin(
 		return -EAGAIN;
 
 	trace_xfs_iomap_atomic_write_cow(ip, offset, length);
-
+retry:
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 
 	if (!ip->i_cowfp) {
@@ -1141,14 +1165,22 @@ xfs_atomic_write_cow_iomap_begin(
 	if (!xfs_iext_lookup_extent(ip, ip->i_cowfp, offset_fsb, &icur, &cmap))
 		cmap.br_startoff = end_fsb;
 	if (cmap.br_startoff <= offset_fsb) {
+		if (isnullstartblock(cmap.br_startblock))
+			goto convert_delay;
+
+		/*
+		 * cmap could extend outside the write range due to previous
+		 * speculative preallocations.  We must trim cmap to the write
+		 * range because the cow fork treats written mappings to mean
+		 * "write in progress".
+		 */
 		xfs_trim_extent(&cmap, offset_fsb, count_fsb);
 		goto found;
 	}
 
-	end_fsb = cmap.br_startoff;
-	count_fsb = end_fsb - offset_fsb;
+	hole_count_fsb = cmap.br_startoff - offset_fsb;
 
-	resaligned = xfs_aligned_fsb_count(offset_fsb, count_fsb,
+	resaligned = xfs_aligned_fsb_count(offset_fsb, hole_count_fsb,
 			xfs_get_cowextsz_hint(ip));
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 
@@ -1169,8 +1201,10 @@ xfs_atomic_write_cow_iomap_begin(
 	if (!xfs_iext_lookup_extent(ip, ip->i_cowfp, offset_fsb, &icur, &cmap))
 		cmap.br_startoff = end_fsb;
 	if (cmap.br_startoff <= offset_fsb) {
-		xfs_trim_extent(&cmap, offset_fsb, count_fsb);
 		xfs_trans_cancel(tp);
+		if (isnullstartblock(cmap.br_startblock))
+			goto convert_delay;
+		xfs_trim_extent(&cmap, offset_fsb, count_fsb);
 		goto found;
 	}
 
@@ -1182,7 +1216,7 @@ xfs_atomic_write_cow_iomap_begin(
 	 * atomic writes to that same range will be aligned (and don't require
 	 * this COW-based method).
 	 */
-	error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb,
+	error = xfs_bmapi_write(tp, ip, offset_fsb, hole_count_fsb,
 			XFS_BMAPI_COWFORK | XFS_BMAPI_PREALLOC |
 			XFS_BMAPI_EXTSZALIGN, 0, &cmap, &nmaps);
 	if (error) {
@@ -1195,21 +1229,43 @@ xfs_atomic_write_cow_iomap_begin(
 	if (error)
 		goto out_unlock;
 
+	/*
+	 * cmap could map more blocks than the range we passed into bmapi_write
+	 * because of EXTSZALIGN or adjacent pre-existing unwritten mappings
+	 * that were merged.  Trim cmap to the original write range so that we
+	 * don't convert more than we were asked to do for this write.
+	 */
+	xfs_trim_extent(&cmap, offset_fsb, count_fsb);
+
 found:
 	if (cmap.br_state != XFS_EXT_NORM) {
-		error = xfs_reflink_convert_cow_locked(ip, offset_fsb,
-				count_fsb);
+		error = xfs_reflink_convert_cow_locked(ip, cmap.br_startoff,
+				cmap.br_blockcount);
 		if (error)
 			goto out_unlock;
 		cmap.br_state = XFS_EXT_NORM;
+		xfs_check_atomic_cow_conversion(ip, offset_fsb, count_fsb,
+				&cmap);
 	}
 
-	length = XFS_FSB_TO_B(mp, cmap.br_startoff + cmap.br_blockcount);
-	trace_xfs_iomap_found(ip, offset, length - offset, XFS_COW_FORK, &cmap);
+	trace_xfs_iomap_found(ip, offset, length, XFS_COW_FORK, &cmap);
 	seq = xfs_iomap_inode_sequence(ip, IOMAP_F_SHARED);
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags, IOMAP_F_SHARED, seq);
 
+convert_delay:
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	error = xfs_bmapi_convert_delalloc(ip, XFS_COW_FORK, offset, iomap,
+			NULL);
+	if (error)
+		return error;
+
+	/*
+	 * Try the lookup again, because the delalloc conversion might have
+	 * turned the COW mapping into unwritten, but we need it to be in
+	 * written state.
+	 */
+	goto retry;
 out_unlock:
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return error;
diff --git a/fs/xfs/xfs_zone_alloc.c b/fs/xfs/xfs_zone_alloc.c
index 23cdab4515bb..ef7a931ebde5 100644
--- a/fs/xfs/xfs_zone_alloc.c
+++ b/fs/xfs/xfs_zone_alloc.c
@@ -246,6 +246,14 @@ xfs_zoned_map_extent(
 	 * If a data write raced with this GC write, keep the existing data in
 	 * the data fork, mark our newly written GC extent as reclaimable, then
 	 * move on to the next extent.
+	 *
+	 * Note that this can also happen when racing with operations that do
+	 * not actually invalidate the data, but just move it to a different
+	 * inode (XFS_IOC_EXCHANGE_RANGE), or to a different offset inside the
+	 * inode (FALLOC_FL_COLLAPSE_RANGE / FALLOC_FL_INSERT_RANGE).  If the
+	 * data was just moved around, GC fails to free the zone, but the zone
+	 * becomes a GC candidate again as soon as all previous GC I/O has
+	 * finished and these blocks will be moved out eventually.
 	 */
 	if (old_startblock != NULLFSBLOCK &&
 	    old_startblock != data.br_startblock)
@@ -607,7 +615,7 @@ xfs_select_open_zone_mru(
 	lockdep_assert_held(&zi->zi_open_zones_lock);
 
 	list_for_each_entry_reverse(oz, &zi->zi_open_zones, oz_entry)
-		if (xfs_try_use_zone(zi, file_hint, oz, false))
+		if (xfs_try_use_zone(zi, file_hint, oz, XFS_ZONE_ALLOC_OK))
 			return oz;
 
 	cond_resched_lock(&zi->zi_open_zones_lock);
@@ -1241,8 +1249,10 @@ xfs_mount_zones(
 
 		while ((rtg = xfs_rtgroup_next(mp, rtg))) {
 			error = xfs_init_zone(&iz, rtg, NULL);
-			if (error)
+			if (error) {
+				xfs_rtgroup_rele(rtg);
 				goto out_free_zone_info;
+			}
 		}
 	}
 
diff --git a/fs/xfs/xfs_zone_gc.c b/fs/xfs/xfs_zone_gc.c
index 109877d9a6bf..4ade54445532 100644
--- a/fs/xfs/xfs_zone_gc.c
+++ b/fs/xfs/xfs_zone_gc.c
@@ -114,6 +114,8 @@ struct xfs_gc_bio {
 	/* Open Zone being written to */
 	struct xfs_open_zone		*oz;
 
+	struct xfs_rtgroup		*victim_rtg;
+
 	/* Bio used for reads and writes, including the bvec used by it */
 	struct bio_vec			bv;
 	struct bio			bio;	/* must be last */
@@ -264,6 +266,7 @@ xfs_zone_gc_iter_init(
 	iter->rec_count = 0;
 	iter->rec_idx = 0;
 	iter->victim_rtg = victim_rtg;
+	atomic_inc(&victim_rtg->rtg_gccount);
 }
 
 /*
@@ -362,6 +365,7 @@ xfs_zone_gc_query(
 
 	return 0;
 done:
+	atomic_dec(&iter->victim_rtg->rtg_gccount);
 	xfs_rtgroup_rele(iter->victim_rtg);
 	iter->victim_rtg = NULL;
 	return 0;
@@ -451,6 +455,20 @@ xfs_zone_gc_pick_victim_from(
 		if (!rtg)
 			continue;
 
+		/*
+		 * If the zone is already undergoing GC, don't pick it again.
+		 *
+		 * This prevents us from picking one of the zones for which we
+		 * already submitted GC I/O, but for which the remapping hasn't
+		 * concluded yet.  This won't cause data corruption, but
+		 * increases write amplification and slows down GC, so this is
+		 * a bad thing.
+		 */
+		if (atomic_read(&rtg->rtg_gccount)) {
+			xfs_rtgroup_rele(rtg);
+			continue;
+		}
+
 		/* skip zones that are just waiting for a reset */
 		if (rtg_rmap(rtg)->i_used_blocks == 0 ||
 		    rtg_rmap(rtg)->i_used_blocks >= victim_used) {
@@ -688,6 +706,9 @@ xfs_zone_gc_start_chunk(
 	chunk->scratch = &data->scratch[data->scratch_idx];
 	chunk->data = data;
 	chunk->oz = oz;
+	chunk->victim_rtg = iter->victim_rtg;
+	atomic_inc(&chunk->victim_rtg->rtg_group.xg_active_ref);
+	atomic_inc(&chunk->victim_rtg->rtg_gccount);
 
 	bio->bi_iter.bi_sector = xfs_rtb_to_daddr(mp, chunk->old_startblock);
 	bio->bi_end_io = xfs_zone_gc_end_io;
@@ -710,6 +731,8 @@ static void
 xfs_zone_gc_free_chunk(
 	struct xfs_gc_bio	*chunk)
 {
+	atomic_dec(&chunk->victim_rtg->rtg_gccount);
+	xfs_rtgroup_rele(chunk->victim_rtg);
 	list_del(&chunk->entry);
 	xfs_open_zone_put(chunk->oz);
 	xfs_irele(chunk->ip);
@@ -770,6 +793,10 @@ xfs_zone_gc_split_write(
 	split_chunk->oz = chunk->oz;
 	atomic_inc(&chunk->oz->oz_ref);
 
+	split_chunk->victim_rtg = chunk->victim_rtg;
+	atomic_inc(&chunk->victim_rtg->rtg_group.xg_active_ref);
+	atomic_inc(&chunk->victim_rtg->rtg_gccount);
+
 	chunk->offset += split_len;
 	chunk->len -= split_len;
 	chunk->old_startblock += XFS_B_TO_FSB(data->mp, split_len);