Messages in this thread | | | Date | Mon, 18 Mar 2013 12:14:45 -0700 | From | "H. Peter Anvin" <> | Subject | Re: [PATCH] x86: mm: accurate the comments for STEP_SIZE_SHIFT macro |
| |
On 03/18/2013 12:13 PM, Yinghai Lu wrote: >> >> No, it doesn't. This is C, not elementary school Now I'm really bothered. >> >> The comment doesn't say *why* (PUD_SHIFT-PMD_SHIFT)/2 or any other >> variant is correct, furthermore I suspect that the +1 is misplaced. >> However, what is really needed is: >> >> 1. Someone needs to explain what the logic should be and why, and >> 2. replace the macro with a symbolic macro, not with a constant and a >> comment explaining, incorrectly, how that value was derived. > > yes, we should find out free_mem_size instead to decide next step size. > > But that will come out page table size estimation problem again. >
Sorry, that comment is double nonsense for someone who isn't intimately familiar with the code, and it sounds like it is just plain wrong.
Instead, try to explain why 5 is the correct value in the current code and how it is (or should be!) derived.
-hpa
| |