本文转载自微信公众号「相遇Linux」,作者JeffXie。转载本文请联系相遇Linux公众号。
我们都知道linux内核中的代码非常精妙,但是有些代码由于历史原因,一个函数为了兼容处理各种情况,到最后可能变得非常糟糕,到处都是goto 和if,看的想跳楼(宋老师口头禅^_^)
如果在系统中读一个文件时会调用
generic_file_buffered_read
这个函数的功能是把磁盘中的数据读到page之后,或者直接获取cache中的page,然后调用copy_page_to_iter把page拷贝到用户层的buffer中。
一天寂静的下午,得空,打开电脑,准备仔细研究一下这个函数,发现这个函数的注释上面就写明了:
- * This is really ugly. But the goto's actually try to clarify some
- * of the logic when it comes to error handling etc.
仔细看了一下代码,果然ugly的不像话,到处都是跳转和判断,令人眩晕,而且整个函数达到300行左右(原谅我看了注释才斗胆这样讲:-) ),发现要是把这个函数看下去,今天一整天的心情都不会好了(当时看的是Linux5.10的代码)
- ssize_t generic_file_buffered_read(struct kiocb *iocb,
- struct iov_iter *iter, ssize_t written)
- {
- find_page:
- if (fatal_signal_pending(current)) {
- error = -EINTR;
- goto out;
- }
- error = wait_on_page_locked_killable(page);
- if (unlikely(error))
- goto readpage_error;
- if (PageUptodate(page))
- goto page_ok;
- if (inode->i_blkbits == PAGE_SHIFT ||
- !mapping->a_ops->is_partially_uptodate)
- goto page_not_up_to_date;
- /* pipes can't handle partially uptodate pages */
- if (unlikely(iov_iter_is_pipe(iter)))
- goto page_not_up_to_date;
- if (!trylock_page(page))
- goto page_not_up_to_date;
- /* Did it get truncated before we got the lock? */
- if (!page->mapping)
- goto page_not_up_to_date_locked;
- if (!mapping->a_ops->is_partially_uptodate(page,
- offset, iter->count))
- goto page_not_up_to_date_locked;
- unlock_page(page);
- }
于是就想内核社区这么多牛人,他们整天盯着这些代码,肯定很多人早已经注意到了,于是想去看看有没有人提交patch重构这个函数:
- ./scripts/get_maintainer.pl mm/filemap.c
- linux-kernel@vger.kernel.org (open list)
然后我就在下面网址中搜索generic_file_buffered_read,果然在10月25号(我看代码那天在11月1号前后),就有人发了相关patch:
- https://lore.kernel.org/lkml/
然后迫不及待查看patch,并把整个patch 下载下来:
这里推荐一个工具,使用b4工具
- https://git.kernel.org/pub/scm/utils/b4/b4.git
可以直接从
- https://lore.kernel.org
获取原始格式的patch,便于自己git am之后测试。
- # b4 am https://lore.kernel.org/lkml/20201025212949.602194-1-kent.overstreet@gmail.com
- v2_20201025_kent_overstreet_generic_file_buffered_read_improvements.cover
- v2_20201025_kent_overstreet_generic_file_buffered_read_improvements.mbx
然后直接 git am ,非常方便,这样就打上了lore.kernel.org上提交的patch.
- git am v2_20201025_kent_overstreet_generic_file_buffered_read_improvements.mbx
- 提示:在git am之前,可以提前git apply --check 一下
- # gitlogdate -3
- fc5608fc9917 2020-10-25 Kent Overstreet fs: generic_file_buffered_read() now uses find_get_pages_contig
- 3bcadc3306be 2020-10-25 Kent Overstreet fs: Break generic_file_buffered_read up into multiple functions
- 3650b228f83a 2020-10-25 Linus Torvalds Linux 5.10-rc1
- alias gitlogdate='git log --pretty=format:"%h%x09%ad%x09%an%x09%s" --date=short'
打了这个patch之后,generic_file_buffered_read变成了这个样子:
- ssize_t generic_file_buffered_read(struct kiocb *iocb,
- struct iov_iter *iter, ssize_t written)
- {
- ..
- pg_nr = generic_file_buffered_read_get_pages(iocb, iter,
- pages, nr_pages);
- ...
- for (i = 0; i < pg_nr; i++) {
- copied = copy_page_to_iter(pages[i], offset, bytes, iter);
- }
而且
- generic_file_buffered_read_get_pages
也非常之清晰:
- static int generic_file_buffered_read_get_pages(struct kiocb *iocb,
- struct iov_iter *iter,
- struct page **pages,
- unsigned int nr)
- nr_got = find_get_pages_contig(mapping, index, nr, pages);
- if (nr_got)
- goto got_pages;
- if (iocb->ki_flags & IOCB_NOIO)
- return -EAGAIN;
- page_cache_sync_readahead(mapping, ra, filp, index, last_index - index);
- nr_got = find_get_pages_contig(mapping, index, nr, pages);
- if (nr_got)
- goto got_pages;
- ...
- }
看完generic_file_buffered_read()之后,整个下午的心情都好多了。