Messages in this thread | | | Subject | Re: [PATCH v15 07/26] x86/mm: Remove _PAGE_DIRTY_HW from kernel RO pages | From | "Yu, Yu-cheng" <> | Date | Mon, 7 Dec 2020 09:11:38 -0800 |
| |
On 12/7/2020 8:36 AM, Borislav Petkov wrote: > On Tue, Nov 10, 2020 at 08:21:52AM -0800, Yu-cheng Yu wrote: >> Kernel read-only PTEs are setup as _PAGE_DIRTY_HW. Since these become >> shadow stack PTEs, remove the dirty bit. > > This commit message is laconic to say the least. You need to start > explaining what you're doing because everytime I look at a patch of > yours, I'm always grepping the SDM and looking forward in the patchset, > trying to rhyme up what that is all about. > > Like for this one. I had to fast-forward to the next patch where all > that is explained. But this is not how review works - each patch's > commit message needs to be understandable on its own because when > they land upstream, they're not in a patchset like here. And review > should be done in the order the patches are numbered - not by jumping > back'n'forth. > > So please think of the readers of your patches when writing those commit > messages. Latter are *not* write-only and not unimportant. > > And those readers haven't spent copious amounts of time on the > technology so being more verbose and explaining things is a Good > Thing(tm). Don't worry about explaining too much - better too much than > too little. > > And last but not least, having understandable and properly written > commit messages increases the chances of your patches landing upstream > considerably. > > Thx. >
Thanks for your feedback. I will improve the commit logs.
-- Yu-cheng
| |