On Fri, 2021-01-15 at 14:15 -0800, James Bottomley wrote:

On Fri, 2021-01-15 at 13:23 -0800, Joe Perches wrote:

quoted

In commit e28bf1f03b01 ("RDMA: Convert various random sprintf sysfs
_show
uses to sysfs_emit") I mistakenly used len = sysfs_emit_at to
overwrite
the last trailing space of potentially multiple entry output.

The length of the last sysfs_emit_at call is 1 and it should instead
be
ignored.  Do so.

Fixes: e28bf1f03b01 ("RDMA: Convert various random sprintf sysfs
_show uses to sysfs_emit")

Reported-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Signed-off-by: Joe Perches <joe@perches.com>
---
 drivers/infiniband/hw/usnic/usnic_ib_sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/hw/usnic/usnic_ib_sysfs.c

b/drivers/infiniband/hw/usnic/usnic_ib_sysfs.c
index e59615a4c9d9..fc077855b46c 100644

--- a/drivers/infiniband/hw/usnic/usnic_ib_sysfs.c
+++ b/drivers/infiniband/hw/usnic/usnic_ib_sysfs.c

@@ -231,7 +231,7 @@ static ssize_t summary_show(struct

usnic_ib_qp_grp *qp_grp, char *buf)
 		}
 	}
 

-	len = sysfs_emit_at(buf, len, "\n");
+	sysfs_emit_at(buf, len, "\n");	/* Overwrite the last
trailing space */

len is the offset of where the next character gets written, isn't it?
so if you're overwriting the last character emitted into buf, shouldn't
the offset point at that character rather than one beyond it?  So

sysfs_emit_at(buf, len - 1, "\n");	/* Overwrite the last trailing
space */

<sigh> quite right, thanks for catching yet another braino from me today.

I'll step away from the keyboard and go back to doing other things today
after submitting a V2 with a more typical style without the silly
backspace/newline by trimming the trailing space from the formats...