[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <e6ac0f5d-61ba-3525-10b9-db2db9549464@huawei.com>
Date: Thu, 6 May 2021 10:15:49 +0800
From: Chao Yu <yuchao0@...wei.com>
To: Jaegeuk Kim <jaegeuk@...nel.org>
CC: <linux-f2fs-devel@...ts.sourceforge.net>,
<linux-kernel@...r.kernel.org>, <chao@...nel.org>
Subject: Re: [PATCH v2] f2fs: compress: remove unneed check condition
On 2021/5/4 22:42, Jaegeuk Kim wrote:
> Hi Chao,
>
> I split this into two patches along with upstreamed change.
>
> https://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git/log/?h=dev-test
Jaegeuk, it's better, thanks, :)
Thanks,
>
> Thanks,
>
> On 04/27, Chao Yu wrote:
>> This patch changes as below:
>> - remove unneeded check condition in __cluster_may_compress()
>> - rename __cluster_may_compress() to cluster_has_invalid_data() for
>> better readability
>> - add cp_error check in f2fs_write_compressed_pages() like we did
>> in f2fs_write_single_data_page()
>>
>> Signed-off-by: Chao Yu <yuchao0@...wei.com>
>> ---
>> v2:
>> - rename function for better readability
>> - add cp_error check in f2fs_write_compressed_pages()
>> fs/f2fs/compress.c | 22 +++++++++++-----------
>> 1 file changed, 11 insertions(+), 11 deletions(-)
>>
>> diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c
>> index 6e46a00c1930..53f78befed8f 100644
>> --- a/fs/f2fs/compress.c
>> +++ b/fs/f2fs/compress.c
>> @@ -888,9 +888,8 @@ bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index)
>> return is_page_in_cluster(cc, index);
>> }
>>
>> -static bool __cluster_may_compress(struct compress_ctx *cc)
>> +static bool cluster_has_invalid_data(struct compress_ctx *cc)
>> {
>> - struct f2fs_sb_info *sbi = F2FS_I_SB(cc->inode);
>> loff_t i_size = i_size_read(cc->inode);
>> unsigned nr_pages = DIV_ROUND_UP(i_size, PAGE_SIZE);
>> int i;
>> @@ -898,18 +897,13 @@ static bool __cluster_may_compress(struct compress_ctx *cc)
>> for (i = 0; i < cc->cluster_size; i++) {
>> struct page *page = cc->rpages[i];
>>
>> - f2fs_bug_on(sbi, !page);
>> -
>> - if (unlikely(f2fs_cp_error(sbi)))
>> - return false;
>> - if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
>> - return false;
>> + f2fs_bug_on(F2FS_I_SB(cc->inode), !page);
>>
>> /* beyond EOF */
>> if (page->index >= nr_pages)
>> - return false;
>> + return true;
>> }
>> - return true;
>> + return false;
>> }
>>
>> static int __f2fs_cluster_blocks(struct compress_ctx *cc, bool compr)
>> @@ -985,7 +979,7 @@ static bool cluster_may_compress(struct compress_ctx *cc)
>> return false;
>> if (unlikely(f2fs_cp_error(F2FS_I_SB(cc->inode))))
>> return false;
>> - return __cluster_may_compress(cc);
>> + return !cluster_has_invalid_data(cc);
>> }
>>
>> static void set_cluster_writeback(struct compress_ctx *cc)
>> @@ -1232,6 +1226,12 @@ static int f2fs_write_compressed_pages(struct compress_ctx *cc,
>> loff_t psize;
>> int i, err;
>>
>> + /* we should bypass data pages to proceed the kworkder jobs */
>> + if (unlikely(f2fs_cp_error(sbi))) {
>> + mapping_set_error(cc->rpages[0]->mapping, -EIO);
>> + goto out_free;
>> + }
>> +
>> if (IS_NOQUOTA(inode)) {
>> /*
>> * We need to wait for node_write to avoid block allocation during
>> --
>> 2.29.2
> .
>
Powered by blists - more mailing lists