Wire O_DIRECT also to Uncached I/O (#17218)

Before Direct I/O was implemented, I've implemented lighter version
I called Uncached I/O.  It uses normal DMU/ARC data path with some
optimizations, but evicts data from caches as soon as possible and
reasonable.  Originally I wired it only to a primarycache property,
but now completing the integration all the way up to the VFS.

While Direct I/O has the lowest possible memory bandwidth usage,
it also has a significant number of limitations.  It require I/Os
to be page aligned, does not allow speculative prefetch, etc.  The
Uncached I/O does not have those limitations, but instead require
additional memory copy, though still one less than regular cached
I/O.  As such it should fill the gap in between.  Considering this
I've disabled annoying EINVAL errors on misaligned requests, adding
a tunable for those who wants to test their applications.

To pass the information between the layers I had to change a number
of APIs.  But as side effect upper layers can now control not only
the caching, but also speculative prefetch.  I haven't wired it to
VFS yet, since it require looking on some OS specifics.  But while
there I've implemented speculative prefetch of indirect blocks for
Direct I/O, controllable via all the same mechanisms.

Signed-off-by:	Alexander Motin <mav@FreeBSD.org>
Sponsored by:	iXsystems, Inc.
Fixes #17027
Reviewed-by: Rob Norris <robn@despairlabs.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>

tests/zfs-tests/include/tunables.cfg

@@ -107,6 +107,7 @@ VOL_USE_BLK_MQ			UNSUPPORTED			zvol_use_blk_mq
 BCLONE_ENABLED			bclone_enabled			zfs_bclone_enabled
 BCLONE_WAIT_DIRTY		bclone_wait_dirty		zfs_bclone_wait_dirty
 DIO_ENABLED			dio_enabled			zfs_dio_enabled
+DIO_STRICT			dio_strict			zfs_dio_strict
 XATTR_COMPAT			xattr_compat			zfs_xattr_compat
 ZEVENT_LEN_MAX			zevent.len_max			zfs_zevent_len_max
 ZEVENT_RETAIN_MAX		zevent.retain_max		zfs_zevent_retain_max

tests/zfs-tests/tests/functional/direct/dio_unaligned_block.ksh

@@ -40,8 +40,10 @@
 
 verify_runnable "global"
 
+log_must save_tunable DIO_STRICT
 function cleanup
 {
+	restore_tunable DIO_STRICT
 	zfs set recordsize=$rs $TESTPOOL/$TESTFS
 	zfs set direct=standard $TESTPOOL/$TESTFS
 	log_must rm -f $tmp_file
@@ -61,6 +63,13 @@ file_size=$((rs * 8))
 
 log_must stride_dd -i /dev/urandom -o $tmp_file -b $file_size -c 1
 
+log_must set_tunable32 DIO_STRICT 0
+log_must zfs set direct=standard $TESTPOOL/$TESTFS
+# sub-pagesize direct writes/read will always pass if not strict.
+log_must stride_dd -i /dev/urandom -o $tmp_file -b 512 -c 8 -D
+log_must stride_dd -i $tmp_file -o /dev/null -b 512 -c 8 -d
+
+log_must set_tunable32 DIO_STRICT 1
 log_must zfs set direct=standard $TESTPOOL/$TESTFS
 # sub-pagesize direct writes/read will always fail if direct=standard.
 log_mustnot stride_dd -i /dev/urandom -o $tmp_file -b 512 -c 8 -D

tests/zfs-tests/tests/functional/stat/statx_dioalign.ksh

@@ -48,6 +48,7 @@ TESTDS=${TESTPOOL}/${TESTFS}
 TESTFILE=${TESTDIR}/${TESTFILE0}
 
 log_must save_tunable DIO_ENABLED
+log_must save_tunable DIO_STRICT
 typeset recordsize_saved=$(get_prop recordsize $TESTDS)
 typeset direct_saved=$(get_prop direct $TESTDS)
 
@@ -57,6 +58,7 @@ function cleanup
 	zfs set recordsize=$recordsize_saved $TESTDS
 	zfs set direct=$direct_saved $TESTDS
 	restore_tunable DIO_ENABLED
+	restore_tunable DIO_STRICT
 }
 log_onexit cleanup
 
@@ -154,6 +156,7 @@ for krs in 4 8 16 32 64 128 256 512 ; do
 done
 
 # reset for write tests
+log_must set_tunable32 DIO_STRICT 1
 log_must zfs set recordsize=16K $TESTDS
 log_must zfs set direct=standard $TESTDS
 
@@ -173,4 +176,12 @@ log_must zpool sync
 assert_dioalign $TESTFILE $PAGE_SIZE 16384
 log_mustnot dd if=/dev/urandom of=$TESTFILE bs=1024 count=256 oflag=direct
 
+# same again, but without strict, which should succeed.
+log_must set_tunable32 DIO_STRICT 0
+log_must rm -f $TESTFILE
+log_must touch $TESTFILE
+log_must zpool sync
+assert_dioalign $TESTFILE $PAGE_SIZE 16384
+log_must dd if=/dev/urandom of=$TESTFILE bs=1024 count=256 oflag=direct
+
 log_pass $CLAIM