From: Xin Hao <xhao@linux.alibaba.com>
To: SeongJae Park <sj@kernel.org>
Cc: sjpark@amazon.de, akpm@linux-foundation.org, linux-mm@kvack.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH] mm/damon/dbgfs: Optimize target_ids interface write operation
Date: Fri, 22 Oct 2021 00:42:50 +0800 [thread overview]
Message-ID: <3fecc78c-44ed-368a-d64f-6e809020bd77@linux.alibaba.com> (raw)
In-Reply-To: <20211021094543.1846-1-sj@kernel.org>
On 2021/10/21 下午5:45, SeongJae Park wrote:
> Hello Xin,
>
> On Thu, 21 Oct 2021 16:56:11 +0800 Xin Hao <xhao@linux.alibaba.com> wrote:
>
>> When writing some pids to target_ids interface, calling scanf()
>> to get 'id' may be failed. If the value of '*nr_ids' is 0 at this time,
>> there is no need to return 'ids' here, we just need to release it and
>> return NULL pointer to improve related code operation efficiency.
> Thank you for the patch! But, I don't think this patch makes sense, because
> the case (*nr_ids == 0) means not error but an ask to remove previously set
> target ids. For example, it works as below now:
>
> # echo 42 > target_ids
> # cat target_ids
> 42
> # echo > target_ids
> # cat target_ids
>
> #
>
> But, with your patch, the behavior will be changed as below:
>
> # echo 42 > target_ids
> # cat target_ids
> 42
> # echo > target_ids
> bash: echo: write error: Cannot allocate memory
> # cat target_ids
> 42
> #
Hi Park:
I did not use the lastest code to test, so I ignored that it calls
'damon_set_targets'
to clear previously set target ids, Sorry.
I think if we want only want to clear the previously set target
ids, this codes are all no need to execute, except call
'damon_set_targets(ctx, NULL, 0);'
----------------------------------------------------------
if (id_is_pid) {
for (i = 0; i < nr_targets; i++) {
targets[i] = (unsigned long)find_get_pid(
(int)targets[i]);
if (!targets[i]) {
dbgfs_put_pids(targets, i);
ret = -EINVAL;
goto free_targets_out;
}
}
}
mutex_lock(&ctx->kdamond_lock);
if (ctx->kdamond) {
if (id_is_pid)
dbgfs_put_pids(targets, nr_targets);
ret = -EBUSY;
goto unlock_out;
}
. . .
/* Configure the context for the address space type */
if (id_is_pid)
damon_va_set_primitives(ctx);
else
damon_pa_set_primitives(ctx);
ret = damon_set_targets(ctx, targets, nr_targets);
if (ret) {
if (id_is_pid)
dbgfs_put_pids(targets, nr_targets);
} else {
ret = count;
>
> Also, this patch makes the DAMON selftest fails as below:
>
> $ sudo make -C tools/testing/selftests/damon run_tests
> make: Entering directory '/home/sjpark/linux/tools/testing/selftests/damon'
> TAP version 13
> 1..2
> # selftests: damon: debugfs_attrs.sh
> # ./debugfs_attrs.sh: line 11: echo: write error: Invalid argument
> # ./debugfs_attrs.sh: line 11: echo: write error: Invalid argument
> # ./debugfs_attrs.sh: line 11: echo: write error: Invalid argument
> # ./debugfs_attrs.sh: line 11: echo: write error: Invalid argument
> # ./debugfs_attrs.sh: line 11: echo: write error: Invalid argument
> # ./debugfs_attrs.sh: line 11: echo: write error: Cannot allocate memory
> # writing abc 2 3 to /sys/kernel/debug/damon/target_ids doesn't return 0
> # expected because: the file allows wrong input
> not ok 1 selftests: damon: debugfs_attrs.sh # exit=1
>
> If I'm missing something, please let me know.
I will post a new patch to fix the problems you encountered in the above
test, thank you so much!
>
>
> Thanks,
> SJ
>
>> Signed-off-by: Xin Hao <xhao@linux.alibaba.com>
>> ---
>> mm/damon/dbgfs.c | 12 ++++++++----
>> 1 file changed, 8 insertions(+), 4 deletions(-)
>>
>> diff --git a/mm/damon/dbgfs.c b/mm/damon/dbgfs.c
>> index a02cf6bee8e8..2d77bf579ffb 100644
>> --- a/mm/damon/dbgfs.c
>> +++ b/mm/damon/dbgfs.c
>> @@ -308,21 +308,25 @@ static unsigned long *str_to_target_ids(const char *str, ssize_t len,
>> unsigned long *ids;
>> const int max_nr_ids = 32;
>> unsigned long id;
>> - int pos = 0, parsed, ret;
>> + int pos = 0, parsed;
>>
>> *nr_ids = 0;
>> ids = kmalloc_array(max_nr_ids, sizeof(id), GFP_KERNEL);
>> if (!ids)
>> return NULL;
>> while (*nr_ids < max_nr_ids && pos < len) {
>> - ret = sscanf(&str[pos], "%lu%n", &id, &parsed);
>> - pos += parsed;
>> - if (ret != 1)
>> + if (sscanf(&str[pos], "%lu%n", &id, &parsed) != 1)
>> break;
>> + pos += parsed;
>> ids[*nr_ids] = id;
>> *nr_ids += 1;
>> }
>>
>> + if (!*nr_ids) {
>> + kfree(ids);
>> + return NULL;
>> + }
>> +
>> return ids;
>> }
>>
>> --
>> 2.31.0
--
Best Regards!
Xin Hao
prev parent reply other threads:[~2021-10-21 16:43 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-21 8:56 Xin Hao
2021-10-21 9:45 ` SeongJae Park
2021-10-21 16:42 ` Xin Hao [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3fecc78c-44ed-368a-d64f-6e809020bd77@linux.alibaba.com \
--to=xhao@linux.alibaba.com \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=sj@kernel.org \
--cc=sjpark@amazon.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox