On 3/31/20 8:33 PM, xiakaixu1987@xxxxxxxxx wrote:
> From: Kaixu Xia <kaixuxia@xxxxxxxxxxx>
>
> The trace event xfs_dquot_dqalloc does not depend on the
> value uq, so remove the condition, and trace quota allocations
> for all quota types.
>
> Signed-off-by: Kaixu Xia <kaixuxia@xxxxxxxxxxx>
> ---
> v2:
> - don't move the tracepoint higher in the function.
>
> fs/xfs/xfs_qm.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
> index 0b09096..43df596 100644
> --- a/fs/xfs/xfs_qm.c
> +++ b/fs/xfs/xfs_qm.c
> @@ -1714,8 +1714,7 @@ struct xfs_qm_isolate {
> pq = xfs_qm_dqhold(ip->i_pdquot);
> }
> }
> - if (uq)
> - trace_xfs_dquot_dqalloc(ip);
> + trace_xfs_dquot_dqalloc(ip);
>
> xfs_iunlock(ip, lockflags);
> if (O_udqpp)
I'm not sure, but do we need something like following description in the
commit log ?
"Prior to commit 0b1b213fcf3a8 ("xfs: event tracing support") xfs
tracing had
xfs_dqtrace_entry_ino() which was dependent on the user quota "uq" , the new
API introduced by the above mentioned commit is not dependent on the uq,
hence remove the check for uq."
Irrespective of that this looks good to me :-
Reviewed-by : Chaitanya Kulkarni <chaitanya.kulkarni@xxxxxxx>
