[PATCH] dlm: use KMEM_CACHE in dlm_memory_init

[PATCH] dlm: use KMEM_CACHE in dlm_memory_init

[Hongfu Li]

Using KMEM_CACHE() macro makes the code more concise and easy to read.

Signed-off-by: Hongfu Li <lihongfu@kylinos.cn>
---
 fs/dlm/memory.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/fs/dlm/memory.c b/fs/dlm/memory.c
index 15a8b1cee433..829cc4fe7bf1 100644
--- a/fs/dlm/memory.c
+++ b/fs/dlm/memory.c
@@ -34,8 +34,7 @@ int __init dlm_memory_init(void)
 	if (!mhandle_cache)
 		goto mhandle;
 
-	lkb_cache = kmem_cache_create("dlm_lkb", sizeof(struct dlm_lkb),
-				__alignof__(struct dlm_lkb), 0, NULL);
+	lkb_cache = KMEM_CACHE(dlm_lkb, 0);
 	if (!lkb_cache)
 		goto lkb;
 
@@ -43,14 +42,11 @@ int __init dlm_memory_init(void)
 	if (!msg_cache)
 		goto msg;
 
-	rsb_cache = kmem_cache_create("dlm_rsb", sizeof(struct dlm_rsb),
-				__alignof__(struct dlm_rsb), 0, NULL);
+	rsb_cache = KMEM_CACHE(dlm_rsb, 0);
 	if (!rsb_cache)
 		goto rsb;
 
-	cb_cache = kmem_cache_create("dlm_cb", sizeof(struct dlm_callback),
-				     __alignof__(struct dlm_callback), 0,
-				     NULL);
+	cb_cache = KMEM_CACHE(dlm_callback, 0);
 	if (!cb_cache)
 		goto cb;
 
-- 
2.25.1

Re: [PATCH] dlm: use KMEM_CACHE() in dlm_memory_init()

[Markus Elfring]

> Using KMEM_CACHE() macro makes the code more concise and easy to read.

Can the three passed name strings matter still for the identification
of the created caches from this function implementation?
https://elixir.bootlin.com/linux/v6.10-rc4/source/fs/dlm/memory.c#L27
https://elixir.bootlin.com/linux/v6.10-rc4/source/mm/slab_common.c#L362

Regards,
Markus

Re: [PATCH] dlm: use KMEM_CACHE() in dlm_memory_init()

[Alexander Aring]

Hi,

On Wed, Jun 19, 2024 at 4:54 PM Markus Elfring <Markus.Elfring@web.de> wrote:
>
> > Using KMEM_CACHE() macro makes the code more concise and easy to read.
>
> Can the three passed name strings matter still for the identification
> of the created caches from this function implementation?
> https://elixir.bootlin.com/linux/v6.10-rc4/source/fs/dlm/memory.c#L27
> https://elixir.bootlin.com/linux/v6.10-rc4/source/mm/slab_common.c#L362

probably only for "dlm_cb" that turns into "dlm_callback". The only
place for me would be /proc/slabinfo, but I usually do `cat
/proc/slabinfo | grep dlm` to get an overview. If you are very strict
and "/proc/slabinfo" is not just "debugging" only and allowed to be
screen scraped (which I don't believe) it might can indeed break some
userspace applications.

- Alex

Re: dlm: use KMEM_CACHE() in dlm_memory_init()

[Markus Elfring]

>> Can the three passed name strings matter still for the identification
>> of the created caches from this function implementation?
>> https://elixir.bootlin.com/linux/v6.10-rc4/source/fs/dlm/memory.c#L27
>> https://elixir.bootlin.com/linux/v6.10-rc4/source/mm/slab_common.c#L362
>
> probably only for "dlm_cb" that turns into "dlm_callback".
…

Will the development attention grow for deviations of passed name strings
from applied data structure identifiers?

Regards,
Markus

Re: [PATCH] dlm: use KMEM_CACHE() in dlm_memory_init()

[Hongfu Li]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=y, Size: 752 bytes --]

>>> Can the three passed name strings matter still for the identification
>>> of the created caches from this function implementation?
>>> https://elixir.bootlin.com/linux/v6.10-rc4/source/fs/dlm/memory.c#L27
>>> https://elixir.bootlin.com/linux/v6.10-rc4/source/mm/slab_common.c#L362
>>
>> probably only for "dlm_cb" that turns into "dlm_callback".
>…
>
>Will the development attention grow for deviations of passed name strings
>from applied data structure identifiers?

My initial purpose is to replace kmem_cache_create() with KMEM_CACHE().
In my view，there is no problem in unifying passed name strings and
structure identifiers.
Maybe that's wrong，I will resubmit a patch that does not change
"dlm_cb" to "dlm_callback".

thanks,
Hongfu Li

Re: dlm: use KMEM_CACHE() in dlm_memory_init()

[Markus Elfring]

>>>> https://elixir.bootlin.com/linux/v6.10-rc4/source/fs/dlm/memory.c#L27
>>>> https://elixir.bootlin.com/linux/v6.10-rc4/source/mm/slab_common.c#L362
>>>
>>> probably only for "dlm_cb" that turns into "dlm_callback".
>> …
>>
>> Will the development attention grow for deviations of passed name strings
>> from applied data structure identifiers?
>
> My initial purpose is to replace kmem_cache_create() with KMEM_CACHE().

Do you take any help from advanced analysis tools into account
for such transformation attempts?

Example for the semantic patch language (Coccinelle software):
[PATCH v2] Coccinelle: api: Add SmPL script “use_KMEM_CACHE.cocci”
https://lore.kernel.org/cocci/b08603d6-cac1-4876-a56c-30c680d5dc52@web.de/
https://sympa.inria.fr/sympa/arc/cocci/2024-02/msg00000.html


> In my view，there is no problem in unifying passed name strings and
> structure identifiers.
> Maybe that's wrong，

I suggest to take another look also at feedback which other contributors
(from your Linux software distribution) got for similar change suggestions.


> I will resubmit a patch that does not change "dlm_cb" to "dlm_callback".

Do you see further opportunities to improve the change description accordingly?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.10-rc4#n94

Regards,
Markus

Re: dlm: use KMEM_CACHE() in dlm_memory_init()

[Hongfu Li]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=y, Size: 1115 bytes --]

>> My initial purpose is to replace kmem_cache_create() with KMEM_CACHE().
>
>Do you take any help from advanced analysis tools into account
>for such transformation attempts?
>
>Example for the semantic patch language (Coccinelle software):
>[PATCH v2] Coccinelle: api: Add SmPL script “use_KMEM_CACHE.cocci”
>https://lore.kernel.org/cocci/b08603d6-cac1-4876-a56c-30c680d5dc52@web.de/
>https://sympa.inria.fr/sympa/arc/cocci/2024-02/msg00000.html
>
>
>> In my view，there is no problem in unifying passed name strings and
>> structure identifiers.
>> Maybe that's wrong，
>
>I suggest to take another look also at feedback which other contributors
>(from your Linux software distribution) got for similar change suggestions.
>
>
>> I will resubmit a patch that does not change "dlm_cb" to "dlm_callback".
>
>Do you see further opportunities to improve the change description accordingly?
>https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.10-rc4#n94

So grateful for you suggestion and I will definitely consider it.

thanks,
Hongfu Li