[PATCH RFC] mm/readahead: improve randread performance with readahead disabled

[PATCH RFC] mm/readahead: improve randread performance with readahead disabled

[Yu Kuai]

From: Yu Kuai <yukuai3@huawei.com>

We have a workload of random 4k-128k read on a HDD, from iostat we observed
that average request size is 256k+ and bandwidth is 100MB+, this is because
readahead waste lots of disk bandwidth. Hence we disable readahead and
performance from user side is indeed much better(2x+), however, from
iostat we observed request size is just 4k and bandwidth is just around
40MB.

Then we do a simple dd test and found out if readahead is disabled,
page_cache_sync_ra() will force to read one page at a time, and this
really doesn't make sense because we can just issue user requested size
request to disk.

Fix this problem by removing the limit to read one page at a time from
page_cache_sync_ra(), this way the random read workload can get better
performance with readahead disabled.

PS: I'm not sure if I miss anything, so this version is RFC
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
---
 mm/readahead.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/mm/readahead.c b/mm/readahead.c
index 20d36d6b055e..1df85ccba575 100644
--- a/mm/readahead.c
+++ b/mm/readahead.c
@@ -561,13 +561,21 @@ void page_cache_sync_ra(struct readahead_control *ractl,
 	 * Even if readahead is disabled, issue this request as readahead
 	 * as we'll need it to satisfy the requested range. The forced
 	 * readahead will do the right thing and limit the read to just the
-	 * requested range, which we'll set to 1 page for this case.
+	 * requested range.
 	 */
-	if (!ra->ra_pages || blk_cgroup_congested()) {
+	if (blk_cgroup_congested()) {
 		if (!ractl->file)
 			return;
+		/*
+		 * If the cgroup is congested, ensure to do at least 1 page of
+		 * readahead to make progress on the read.
+		 */
 		req_count = 1;
 		do_forced_ra = true;
+	} else if (!ra->ra_pages) {
+		if (!ractl->file)
+			return;
+		do_forced_ra = true;
 	}
 
 	/* be dumb */
-- 
2.39.2

Re: [PATCH RFC] mm/readahead: improve randread performance with readahead disabled

[Yu Kuai]

Hi,

在 2025/07/01 19:08, Yu Kuai 写道:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> We have a workload of random 4k-128k read on a HDD, from iostat we observed
> that average request size is 256k+ and bandwidth is 100MB+, this is because
> readahead waste lots of disk bandwidth. Hence we disable readahead and
> performance from user side is indeed much better(2x+), however, from
> iostat we observed request size is just 4k and bandwidth is just around
> 40MB.
> 
> Then we do a simple dd test and found out if readahead is disabled,
> page_cache_sync_ra() will force to read one page at a time, and this
> really doesn't make sense because we can just issue user requested size
> request to disk.
> 
> Fix this problem by removing the limit to read one page at a time from
> page_cache_sync_ra(), this way the random read workload can get better
> performance with readahead disabled.
> 
> PS: I'm not sure if I miss anything, so this version is RFC
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> ---
>   mm/readahead.c | 12 ++++++++++--
>   1 file changed, 10 insertions(+), 2 deletions(-)
> 

Friendly ping ...

> diff --git a/mm/readahead.c b/mm/readahead.c
> index 20d36d6b055e..1df85ccba575 100644
> --- a/mm/readahead.c
> +++ b/mm/readahead.c
> @@ -561,13 +561,21 @@ void page_cache_sync_ra(struct readahead_control *ractl,
>   	 * Even if readahead is disabled, issue this request as readahead
>   	 * as we'll need it to satisfy the requested range. The forced
>   	 * readahead will do the right thing and limit the read to just the
> -	 * requested range, which we'll set to 1 page for this case.
> +	 * requested range.
>   	 */
> -	if (!ra->ra_pages || blk_cgroup_congested()) {
> +	if (blk_cgroup_congested()) {
>   		if (!ractl->file)
>   			return;
> +		/*
> +		 * If the cgroup is congested, ensure to do at least 1 page of
> +		 * readahead to make progress on the read.
> +		 */
>   		req_count = 1;
>   		do_forced_ra = true;
> +	} else if (!ra->ra_pages) {
> +		if (!ractl->file)
> +			return;
> +		do_forced_ra = true;
>   	}
>   
>   	/* be dumb */
> 

Re: [PATCH RFC] mm/readahead: improve randread performance with readahead disabled

[Yu Kuai]

Hi,

在 2025/07/14 9:42, Yu Kuai 写道:
> Hi,
> 
> 在 2025/07/01 19:08, Yu Kuai 写道:
>> From: Yu Kuai <yukuai3@huawei.com>
>>
>> We have a workload of random 4k-128k read on a HDD, from iostat we 
>> observed
>> that average request size is 256k+ and bandwidth is 100MB+, this is 
>> because
>> readahead waste lots of disk bandwidth. Hence we disable readahead and
>> performance from user side is indeed much better(2x+), however, from
>> iostat we observed request size is just 4k and bandwidth is just around
>> 40MB.
>>
>> Then we do a simple dd test and found out if readahead is disabled,
>> page_cache_sync_ra() will force to read one page at a time, and this
>> really doesn't make sense because we can just issue user requested size
>> request to disk.
>>
>> Fix this problem by removing the limit to read one page at a time from
>> page_cache_sync_ra(), this way the random read workload can get better
>> performance with readahead disabled.
>>
>> PS: I'm not sure if I miss anything, so this version is RFC
>> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
>> ---
>>   mm/readahead.c | 12 ++++++++++--
>>   1 file changed, 10 insertions(+), 2 deletions(-)
>>
> 
> Friendly ping ...
> 

Friendly ping again ...

>> diff --git a/mm/readahead.c b/mm/readahead.c
>> index 20d36d6b055e..1df85ccba575 100644
>> --- a/mm/readahead.c
>> +++ b/mm/readahead.c
>> @@ -561,13 +561,21 @@ void page_cache_sync_ra(struct readahead_control 
>> *ractl,
>>        * Even if readahead is disabled, issue this request as readahead
>>        * as we'll need it to satisfy the requested range. The forced
>>        * readahead will do the right thing and limit the read to just the
>> -     * requested range, which we'll set to 1 page for this case.
>> +     * requested range.
>>        */
>> -    if (!ra->ra_pages || blk_cgroup_congested()) {
>> +    if (blk_cgroup_congested()) {
>>           if (!ractl->file)
>>               return;
>> +        /*
>> +         * If the cgroup is congested, ensure to do at least 1 page of
>> +         * readahead to make progress on the read.
>> +         */
>>           req_count = 1;
>>           do_forced_ra = true;
>> +    } else if (!ra->ra_pages) {
>> +        if (!ractl->file)
>> +            return;
>> +        do_forced_ra = true;
>>       }
>>       /* be dumb */
>>
> 
> .
>