mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: luoxueqin <luoxueqin@kylinos.cn>
To: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: pavel@kernel.org, lenb@kernel.org, linux-pm@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] PM: hibernate: make compression threads configurable via kernel parameter
Date: Thu, 28 Aug 2025 11:29:52 +0800	[thread overview]
Message-ID: <e0968be3-e33a-4e94-ab8f-215cdaa2be17@kylinos.cn> (raw)
In-Reply-To: <CAJZ5v0gD8RkKG8+6MneaDkxndS-oAm8a1AswEDP1w8HVCGZDdw@mail.gmail.com>


在 2025/8/26 19:43, Rafael J. Wysocki 写道:
> On Tue, Aug 26, 2025 at 11:19 AM Xueqin Luo <luoxueqin@kylinos.cn> wrote:
>> A new kernel parameter 'cmp_threads=' is introduced to
>> allow tuning the number of compression/decompression threads at boot.
> And why is it useful/needed?
The number of compression/decompression threads directly impacts 
hibernate and resume time.
In our tests(averaged over 10 runs):
     cmp_threads   hibernate(s)   resume(s)
             3                           12.14          18.86
             4                           12.28          17.48
             5                           11.09          16.77
             6                           11.08          16.44
With 5–6 threads, resume latency improves by ~12% compared to 3 threads. 
But on low-core systems,
more threads may cause contention. Making it configurable allows 
integrators to balance performance
  and CPU usage across different hardware without recompiling the kernel.
>> Signed-off-by: Xueqin Luo <luoxueqin@kylinos.cn>
>> ---
>>   kernel/power/swap.c | 22 +++++++++++++++++-----
>>   1 file changed, 17 insertions(+), 5 deletions(-)
>>
>> diff --git a/kernel/power/swap.c b/kernel/power/swap.c
>> index ad13c461b657..43280e08a4ad 100644
>> --- a/kernel/power/swap.c
>> +++ b/kernel/power/swap.c
>> @@ -520,7 +520,8 @@ static int swap_writer_finish(struct swap_map_handle *handle,
>>   #define CMP_SIZE       (CMP_PAGES * PAGE_SIZE)
>>
>>   /* Maximum number of threads for compression/decompression. */
>> -#define CMP_THREADS    3
>> +#define CMP_MAX_THREADS        12
>> +static int cmp_threads = 3
>>
>>   /* Minimum/maximum number of pages for read buffering. */
>>   #define CMP_MIN_RD_PAGES       1024
>> @@ -585,8 +586,8 @@ struct crc_data {
>>          wait_queue_head_t go;                     /* start crc update */
>>          wait_queue_head_t done;                   /* crc update done */
>>          u32 *crc32;                               /* points to handle's crc32 */
>> -       size_t *unc_len[CMP_THREADS];             /* uncompressed lengths */
>> -       unsigned char *unc[CMP_THREADS];          /* uncompressed data */
>> +       size_t *unc_len[CMP_MAX_THREADS];             /* uncompressed lengths */
>> +       unsigned char *unc[CMP_MAX_THREADS];          /* uncompressed data */
>>   };
>>
>>   /*
>> @@ -703,7 +704,7 @@ static int save_compressed_image(struct swap_map_handle *handle,
>>           * footprint.
>>           */
>>          nr_threads = num_online_cpus() - 1;
>> -       nr_threads = clamp_val(nr_threads, 1, CMP_THREADS);
>> +       nr_threads = clamp_val(nr_threads, 1, cmp_threads);
>>
>>          page = (void *)__get_free_page(GFP_NOIO | __GFP_HIGH);
>>          if (!page) {
>> @@ -1223,7 +1224,7 @@ static int load_compressed_image(struct swap_map_handle *handle,
>>           * footprint.
>>           */
>>          nr_threads = num_online_cpus() - 1;
>> -       nr_threads = clamp_val(nr_threads, 1, CMP_THREADS);
>> +       nr_threads = clamp_val(nr_threads, 1, cmp_threads);
>>
>>          page = vmalloc(array_size(CMP_MAX_RD_PAGES, sizeof(*page)));
>>          if (!page) {
>> @@ -1667,3 +1668,14 @@ static int __init swsusp_header_init(void)
>>   }
>>
>>   core_initcall(swsusp_header_init);
>> +
>> +static int __init cmp_threads_setup(char *str)
>> +{
>> +       int rc = kstrtouint(str, 0, &cmp_threads);
>> +       if (rc)
>> +               return rc;
>> +       return 1;
>> +
>> +}
>> +
>> +__setup("cmp_threads=", cmp_threads_setup);
>> --
>> 2.43.0
>>

  reply	other threads:[~2025-08-28  3:29 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-08-26  9:19 Xueqin Luo
2025-08-26 11:43 ` Rafael J. Wysocki
2025-08-28  3:29   ` luoxueqin [this message]
2025-09-04 19:37     ` Rafael J. Wysocki

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=e0968be3-e33a-4e94-ab8f-215cdaa2be17@kylinos.cn \
    --to=luoxueqin@kylinos.cn \
    --cc=lenb@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=pavel@kernel.org \
    --cc=rafael@kernel.org \
    /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

all inboxes | Powered by JetHome®