On Fri, 2025-02-14 at 21:11 +0000, Matthew Wilcox wrote: > On Tue, Feb 04, 2025 at 04:02:48PM -0800, Viacheslav Dubeyko wrote: > > This patch implements refactoring of ceph_submit_write() > > and also it solves the second issue. > > > > Signed-off-by: Viacheslav Dubeyko <Slava.Dubeyko@xxxxxxx> > > This kind of giant refactoring to solve a bug is a really bad idea. > First, it's going to need to be backported to older kernels. How far > back? You need to identify that with a Fixes: line. > > It's also really hard to review and know whether it's right. You might > have introduced several new bugs while doing it. In general, bugfixes > first, refactor later. I *think* this means we can do without 1/7 of > the patches I resent earlier today, but it's really hard to be sure. I really would like not to do refactoring here. But the function is huge and logic is pretty complicated. It's hard to follow the logic of the method. Refactoring is doing only of moving the existing code into smaller functions and nothing more. I can not to do the refactoring and simply add the fixes but function will be bigger and more complex, from my point of view. Thanks, Slava.
