Hi Michael
Hello Yang Xu,
Thank you for your patch. Please see my comments below.
On Mon, 3 Aug 2020 at 04:38, Yang Xu <xuyang2018.jy@xxxxxxxxxxxxxx> wrote:
Signed-off-by: Yang Xu <xuyang2018.jy@xxxxxxxxxxxxxx>
---
man4/loop.4 | 35 +++++++++++++++++++++++++++++++++++
1 file changed, 35 insertions(+)
diff --git a/man4/loop.4 b/man4/loop.4
index 2dcbac807..abf722a25 100644
--- a/man4/loop.4
+++ b/man4/loop.4
@@ -177,6 +177,41 @@ This value must be a power of two in the range
otherwise, an
.B EINVAL
error results.
+.TP
+.BR LOOP_CONFIGURE " (since Linux 5.8)"
+.\" commit 3448914e8cc550ba792d4ccc74471d1ca4293aae
+Setup and configure all loop device parameters using the (third)
+.BR ioctl (2)
+argument at once.
+This argument is a pointer to
+.I loop_config
+structure, defined in
+.I <linux/loop.h>
+as:
+.IP
+.in +4n
+.EX
+struct loop_config {
+ __u32 fd;
+ __u32 block_size;
+ struct loop_info64 info;
+ __u64 __reserved[8];
+};
+.EE
+.in
+.IP
+In addition to doing what
+.BR LOOP_SET_STATUS
+can do,
+.BR LOOP_CONFIGURE
+can also be used to set the correct block size immediately
+by setting loop_config.block_size. Explicitly request
+direct I/O mode by setting
+.BR LO_FLAGS_DIRECT_IO
+in loop_config.info.lo_flags. Explicitly request read-only
+mode by setting
+.BR LO_FLAGS_READ_ONLY
+in loop_config.info.lo_flags.
The last two sentences (starting "Explicitly request") are hard for me
to understand. Could you expand the wording here a little please to
clarify?
In drivers/block/loop.c code, set status function(loop_set_status)
doesn't handle LO_FLAGS_READ_ONLY flag and ingore it. it is the same for
LO_FLAGS_DIRECT_IO flag. "Explicitly request" is compared with
LOOP_SET_STATUS ioctl.
Best Regards
Yang Xu
Thanks,
Michael
.PP
Since Linux 2.6, there are two new
.BR ioctl (2)
--
2.23.0
