[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1cb50014-8678-40de-bca3-8a33555ec70e@huawei.com>
Date: Wed, 16 Oct 2024 20:36:34 +0800
From: Kefeng Wang <wangkefeng.wang@...wei.com>
To: Baolin Wang <baolin.wang@...ux.alibaba.com>, <akpm@...ux-foundation.org>,
<hughd@...gle.com>
CC: <willy@...radead.org>, <david@...hat.com>, <linux-mm@...ck.org>,
<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/2] mm: shmem: improve the tmpfs large folio read
performance
On 2024/10/16 18:09, Baolin Wang wrote:
> The tmpfs has already supported the PMD-sized large folios, but the tmpfs
> read operation still performs copying at the PAGE SIZE granularity, which
> is unreasonable. This patch changes to copy data at the folio granularity,
> which can improve the read performance, as well as changing to use folio
> related functions.
>
> Use 'fio bs=64k' to read a 1G tmpfs file populated with 2M THPs, and I can
> see about 20% performance improvement, and no regression with bs=4k.
> Before the patch:
> READ: bw=10.0GiB/s
>
> After the patch:
> READ: bw=12.0GiB/s
>
> Signed-off-by: Baolin Wang <baolin.wang@...ux.alibaba.com>
> ---
> mm/shmem.c | 22 ++++++++++++----------
> 1 file changed, 12 insertions(+), 10 deletions(-)
>
> diff --git a/mm/shmem.c b/mm/shmem.c
> index edab02a26aac..7e79b6a96da0 100644
> --- a/mm/shmem.c
> +++ b/mm/shmem.c
> @@ -3108,13 +3108,12 @@ static ssize_t shmem_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> ssize_t retval = 0;
>
> index = iocb->ki_pos >> PAGE_SHIFT;
> - offset = iocb->ki_pos & ~PAGE_MASK;
>
> for (;;) {
> struct folio *folio = NULL;
> - struct page *page = NULL;
> unsigned long nr, ret;
> loff_t end_offset, i_size = i_size_read(inode);
> + size_t fsize;
>
> if (unlikely(iocb->ki_pos >= i_size))
> break;
> @@ -3128,8 +3127,9 @@ static ssize_t shmem_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> if (folio) {
> folio_unlock(folio);
>
> - page = folio_file_page(folio, index);
> - if (PageHWPoison(page)) {
> + if (folio_test_hwpoison(folio) ||
> + (folio_test_large(folio) &&
> + folio_test_has_hwpoisoned(folio))) {
> folio_put(folio);
> error = -EIO;
> break;
> @@ -3147,7 +3147,12 @@ static ssize_t shmem_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> break;
> }
> end_offset = min_t(loff_t, i_size, iocb->ki_pos + to->count);
> - nr = min_t(loff_t, end_offset - iocb->ki_pos, PAGE_SIZE - offset);
> + if (folio)
> + fsize = folio_size(folio);
> + else
> + fsize = PAGE_SIZE;
> + offset = iocb->ki_pos & (fsize - 1);
> + nr = min_t(loff_t, end_offset - iocb->ki_pos, fsize - offset);
>
> if (folio) {
> /*
> @@ -3156,7 +3161,7 @@ static ssize_t shmem_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> * before reading the page on the kernel side.
> */
We'd better to update all the comment from page to folio.
> if (mapping_writably_mapped(mapping))
> - flush_dcache_page(page);
> + flush_dcache_folio(folio);
> /*
> * Mark the page accessed if we read the beginning.
> */
> @@ -3166,9 +3171,8 @@ static ssize_t shmem_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> * Ok, we have the page, and it's up-to-date, so
> * now we can copy it to user space...
> */
> - ret = copy_page_to_iter(page, offset, nr, to);
> + ret = copy_folio_to_iter(folio, offset, nr, to);
> folio_put(folio);
> -
> } else if (user_backed_iter(to)) {
> /*
> * Copy to user tends to be so well optimized, but
> @@ -3186,8 +3190,6 @@ static ssize_t shmem_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
> }
>
> retval += ret;
> - offset += ret;
> - offset &= ~PAGE_MASK;
> iocb->ki_pos += ret;
> index = iocb->ki_pos >> PAGE_SHIFT;
>
Powered by blists - more mailing lists