Re: [PATCH v3 05/13] epoll: offload polling to a work in case of epfd polled from userspace

[Eric Wong]

Roman Penyaev <rpenyaev@xxxxxxx> wrote:
> diff --git a/fs/eventpoll.c b/fs/eventpoll.c
> index 81da4571f1e0..9d3905c0afbf 100644
> --- a/fs/eventpoll.c
> +++ b/fs/eventpoll.c
> @@ -44,6 +44,7 @@
>  #include <linux/seq_file.h>
>  #include <linux/compat.h>
>  #include <linux/rculist.h>
> +#include <linux/workqueue.h>
>  #include <net/busy_poll.h>
>  
>  /*
> @@ -185,6 +186,9 @@ struct epitem {
>  
>  	/* The structure that describe the interested events and the source fd */
>  	struct epoll_event event;
> +
> +	/* Work for offloading event callback */
> +	struct work_struct work;
>  };
>  
>  /*

Can we avoid the size regression for existing epoll users?

> @@ -2547,12 +2601,6 @@ static int __init eventpoll_init(void)
>  	ep_nested_calls_init(&poll_safewake_ncalls);
>  #endif
>  
> -	/*
> -	 * We can have many thousands of epitems, so prevent this from
> -	 * using an extra cache line on 64-bit (and smaller) CPUs
> -	 */
> -	BUILD_BUG_ON(sizeof(void *) <= 8 && sizeof(struct epitem) > 128);
> -
>  	/* Allocates slab cache used to allocate "struct epitem" items */
>  	epi_cache = kmem_cache_create("eventpoll_epi", sizeof(struct epitem),
>  			0, SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT, NULL);

Perhaps a "struct uepitem" transparent union and separate slab cache.