Re: [PATCH] xfs: simplify the flags setting in xfs_qm_scall_quotaon

[kaixuxia <xiakaixu1987@gmail.com>]

On 2020/4/11 11:18, Darrick J. Wong wrote:
> On Sat, Apr 11, 2020 at 09:09:57AM +0800, kaixuxia wrote:
>>
>> On 2020/4/10 22:51, Darrick J. Wong wrote:
>>> On Fri, Apr 10, 2020 at 04:57:04PM +0800, xiakaixu1987@gmail.com wrote:
>>>> From: Kaixu Xia <kaixuxia@tencent.com>
>>>>
>>>> Simplify the setting of the flags value, and only consider
>>>> quota enforcement stuff here.
>>>>
>>>> Signed-off-by: Kaixu Xia <kaixuxia@tencent.com>
>>>> ---
>>>>  fs/xfs/xfs_qm_syscalls.c | 6 +++---
>>>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>>>
>>>> diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c
>>>> index 5d5ac65..944486f 100644
>>>> --- a/fs/xfs/xfs_qm_syscalls.c
>>>> +++ b/fs/xfs/xfs_qm_syscalls.c
>>>> @@ -357,11 +357,11 @@
>>>
>>> No idea which function this is.  diff -p, please.
>>
>> Yeah, the changed function is xfs_qm_scall_quotaon().
>> Anyway, the result of diff -p as follows,
> 
> That was a request to generate your patches with diff -Naurp.

Got it, I will resend all the patches that have not been applied by
the xfs for-next branch with a single patchset.

> 
> --D
> 
>> *** fs/xfs/xfs_qm_syscalls.c	Sat Apr 11 08:32:03 2020
>> --- /tmp/xfs_qm_syscalls.c	Sat Apr 11 08:31:51 2020
>> *************** xfs_qm_scall_quotaon(
>> *** 357,367 ****
>>   	int		error;
>>   	uint		qf;
>>   
>>   	/*
>> ! 	 * Switching on quota accounting must be done at mount time,
>> ! 	 * only consider quota enforcement stuff here.
>>   	 */
>> ! 	flags &= XFS_ALL_QUOTA_ENFD;
>>   
>>   	if (flags == 0) {
>>   		xfs_debug(mp, "%s: zero flags, m_qflags=%x",
>> --- 357,367 ----
>>   	int		error;
>>   	uint		qf;
>>   
>> + 	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
>>   	/*
>> ! 	 * Switching on quota accounting must be done at mount time.
>>   	 */
>> ! 	flags &= ~(XFS_ALL_QUOTA_ACCT);
>>   
>>   	if (flags == 0) {
>>   		xfs_debug(mp, "%s: zero flags, m_qflags=%x",
>>
>>>
>>> Also, please consider putting all these minor cleanups into a single
>>> patchset, it's a lot easier (for me) to track and land one series than
>>> it is to handle a steady trickle of single patches.
>> Yeah, got it. Should I resend all of the patches that have been
>> reviewed or just resend the last two patches with a single patchset?
>>
>> The patches that have been reviewed as follows,
>> xfs: trace quota allocations for all quota types
>> xfs: combine two if statements with same condition
>> xfs: check if reserved free disk blocks is needed
>> xfs: remove unnecessary variable udqp from xfs_ioctl_setattr
>>
>> The last two patches that have not been reviewed as follow,
>> xfs: remove unnecessary assertion from xfs_qm_vop_create_dqattach
>> xfs: simplify the flags setting in xfs_qm_scall_quotaon
>>
>>>
>>> --D
>>>
>>>>  	int		error;
>>>>  	uint		qf;
>>>>  
>>>> -	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
>>>>  	/*
>>>> -	 * Switching on quota accounting must be done at mount time.
>>>> +	 * Switching on quota accounting must be done at mount time,
>>>> +	 * only consider quota enforcement stuff here.
>>>>  	 */
>>>> -	flags &= ~(XFS_ALL_QUOTA_ACCT);
>>>> +	flags &= XFS_ALL_QUOTA_ENFD;
>>>>  
>>>>  	if (flags == 0) {
>>>>  		xfs_debug(mp, "%s: zero flags, m_qflags=%x",
>>>> -- 
>>>> 1.8.3.1
>>>>
>>
>> -- 
>> kaixuxia

-- 
kaixuxia