On 2025/2/18 22:17, Chuck Lever wrote:
> On 2/18/25 4:16 AM, Yunsheng Lin wrote:
>> On 2025/2/17 22:20, Chuck Lever wrote:
>>> On 2/17/25 7:31 AM, Yunsheng Lin wrote:
>>>> As mentioned in [1], it seems odd to check NULL elements in
>>>> the middle of page bulk allocating,
>>>
>>> I think I requested that check to be added to the bulk page allocator.
>>>
>>> When sending an RPC reply, NFSD might release pages in the middle of
>>
>> It seems there is no usage of the page bulk allocation API in fs/nfsd/
>> or fs/nfs/, which specific fs the above 'NFSD' is referring to?
>
> NFSD is in fs/nfsd/, and it is the major consumer of
> net/sunrpc/svc_xprt.c.
>
>
>>> the rq_pages array, marking each of those array entries with a NULL
>>> pointer. We want to ensure that the array is refilled completely in this
>>> case.
>>>
>>
>> I did some researching, it seems you requested that in [1]?
>> It seems the 'holes are always at the start' for the case in that
>> discussion too, I am not sure if the case is referring to the caller
>> in net/sunrpc/svc_xprt.c? If yes, it seems caller can do a better
>> job of bulk allocating pages into a whole array sequentially without
>> checking NULL elements first before doing the page bulk allocation
>> as something below:
>>
>> +++ b/net/sunrpc/svc_xprt.c
>> @@ -663,9 +663,10 @@ static bool svc_alloc_arg(struct svc_rqst *rqstp)
>> pages = RPCSVC_MAXPAGES;
>> }
>>
>> - for (filled = 0; filled < pages; filled = ret) {
>> - ret = alloc_pages_bulk(GFP_KERNEL, pages, rqstp->rq_pages);
>> - if (ret > filled)
>> + for (filled = 0; filled < pages; filled += ret) {
>> + ret = alloc_pages_bulk(GFP_KERNEL, pages - filled,
>> + rqstp->rq_pages + filled);
>> + if (ret)
>> /* Made progress, don't sleep yet */
>> continue;
>>
>> @@ -674,7 +675,7 @@ static bool svc_alloc_arg(struct svc_rqst *rqstp)
>> set_current_state(TASK_RUNNING);
>> return false;
>> }
>> - trace_svc_alloc_arg_err(pages, ret);
>> + trace_svc_alloc_arg_err(pages, filled);
>> memalloc_retry_wait(GFP_KERNEL);
>> }
>> rqstp->rq_page_end = &rqstp->rq_pages[pages];
>>
>>
>> 1. https://lkml.iu.edu/hypermail/linux/kernel/2103.2/09060.html
>
> I still don't see what is broken about the current API.
As mentioned in [1], the page bulk alloc API before this patch may
have some space for improvement from performance and easy-to-use
perspective as the most existing calllers of page bulk alloc API
are trying to bulk allocate the page for the whole array sequentially.
1. https://lore.kernel.org/all/c9950a79-7bcb-41c2-a59e-af315dc6d7ff@xxxxxxxxxx/
>
> Anyway, any changes in svc_alloc_arg() will need to be run through the
> upstream NFSD CI suite before they are merged.
Is there any web link pointing to the above NFSD CI suite, so that I can
test it if removing assumption of populating only NULL elements is indeed
possible?
Look more closely, it seems svc_rqst_release_pages()/svc_rdma_save_io_pages()
does set rqstp->rq_respages[i] to NULL based on rqstp->rq_next_page,
and the original code before using the page bulk alloc API does seem to only
allocate page for NULL elements as can see from the below patch：
https://lore.kernel.org/all/20210325114228.27719-8-mgorman@xxxxxxxxxxxxxxxxxxx/T/#u
The clearing of rqstp->rq_respages[] to NULL does seems sequentially, is it
possible to only pass NULL elements in rqstp->rq_respages[] to alloc_pages_bulk()
so that bulk alloc API does not have to do the NULL checking and use the array only
as output parameter?
>
>
