lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <eaf268ed-cd1b-107c-3092-36df4277fa0f@linux.alibaba.com>
Date:   Wed, 8 Dec 2021 23:13:34 +0800
From:   Xin Hao <xhao@...ux.alibaba.com>
To:     SeongJae Park <sj@...nel.org>
Cc:     akpm@...ux-foundation.org, shuah@...nel.org,
        brendanhiggins@...gle.com, linux-mm@...ck.org,
        linux-kselftest@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 02/11] mm/damon/dbgfs: Remove an unnecessary error message

Hi SeongJae:

On 12/8/21 8:49 PM, SeongJae Park wrote:
> On Wed, 8 Dec 2021 14:29:40 +0800 Xin Hao <xhao@...ux.alibaba.com> wrote:
>
> Hi Xin,
>
>> Hi park:
>>
>> On 12/1/21 11:04 PM, SeongJae Park wrote:
>>> When wrong scheme action is requested via the debugfs interface, DAMON
>>> prints an error message.  Because the function returns error code, this
>>> is not really needed.  Because the code path is triggered by the user
>>> specified input, this can result in kernel log mistakenly being messy.
>> Completely correct, but there will also be a problem that users can’t
>> quickly locate where the problem is,
>>
>> Especially too many parameters need to be written into the interface.
>>
>> I think it is necessary to add some debugging methods to help users find
>> the error without polluting the kernel log.
>>
>> And i have an idea, like this:
>>
>> in dbgfs, add a last_cmd_stat interface.
>>
>>       # echo "1 2 1 2 1 2  1 2 1 2 100 ..."  > schemes
>>
>>       #  cat last_cmd_stat
>>
>>       #  wrong action 100
>>
>> In this way, on the one hand, it will not pollute the kernel log, on the
>> other hand, it will help users find  the cause of the operation
>> interface error.
>>
>> Park, how do you think of about this idea, if ok, i will send a patch.
> Thank you always for your great suggestions and efforts!  BTW, I prefer to be
> called with my first name ;)
Ha-Ha, Sorry!
>
> I want DAMON kernel code to be as simple and small as possible, while putting
> fancy but complicated features for user conveniences in user space tools like
> DAMO[1].  In other words, I hope the DAMON debugfs interface to be used as an
> interface for such user space tools, not an interface for human hands.
Ok, I know what you mean.
>
> IMHO, implementing the feature you proposed in the kernel could make the code
> slightly bigger, while it can easily implemented in user space.  I therefore
> think the feature would be better to be implemented in user space.  If you
> could send a pull request of the feature for DAMO, it would be so great.

Ok,  i will do it,  But there's a problem here, If the user does not use 
the DAMO tools to operate  the dbgfs interface,

the operation interface error will still hard to find the cause of errors.


>
> [1] https://github.com/awslabs/damo
>
>
> Thanks,
> SJ
>
>>> To avoid the case, this commit removes the message
>>>
>>> Fixes: af122dd8f3c0 ("mm/damon/dbgfs: support DAMON-based Operation Schemes")
>>> Signed-off-by: SeongJae Park <sj@...nel.org>
>>> ---
>>>    mm/damon/dbgfs.c | 4 +---
>>>    1 file changed, 1 insertion(+), 3 deletions(-)
>>>
>>> diff --git a/mm/damon/dbgfs.c b/mm/damon/dbgfs.c
>>> index 4bf4204444ab..5b628990ae6e 100644
>>> --- a/mm/damon/dbgfs.c
>>> +++ b/mm/damon/dbgfs.c
>>> @@ -210,10 +210,8 @@ static struct damos **str_to_schemes(const char *str, ssize_t len,
>>>    				&wmarks.low, &parsed);
>>>    		if (ret != 18)
>>>    			break;
>>> -		if (!damos_action_valid(action)) {
>>> -			pr_err("wrong action %d\n", action);
>>> +		if (!damos_action_valid(action))
>>>    			goto fail;
>>> -		}
>>>    
>>>    		if (min_sz > max_sz || min_nr_a > max_nr_a || min_age > max_age)
>>>    			goto fail;
>> -- 
>> Best Regards!
>> Xin Hao
>>
-- 
Best Regards!
Xin Hao

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ