Messages in this thread | | | From | "Kirill A. Shutemov" <> | Subject | Re: [PATCH 1/3] mm: generalize VM_BUG_ON() macros | Date | Thu, 2 Oct 2014 17:11:33 +0300 (EEST) |
| |
Andrew Morton wrote: > On Wed, 1 Oct 2014 14:31:59 +0300 "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> wrote: > > > This patch makes VM_BUG_ON() to accept one to three arguments after the > > condition. Any of these arguments can be page, vma or mm. VM_BUG_ON() > > will dump info about the argument using appropriate dump_* function. > > > > It's intended to replace separate VM_BUG_ON_PAGE(), VM_BUG_ON_VMA(), > > VM_BUG_ON_MM() and allows additional use-cases like: > > > > VM_BUG_ON(cond, vma, page); > > VM_BUG_ON(cond, vma, src_page, dst_page); > > VM_BUG_ON(cond, mm, src_vma, dst_vma); > > ... > > I can't say I'm a fan of this. We don't do this sort of thing anywhere > else in the kernel and passing different types to the same thing in > different places is unusual and exceptional. We gain very little from > this so why bother?
We had bug like this: lkml.kernel.org/r/53F487EB.7070703@oracle.com where it's useful to see more than one structure dumped: vma + page in this case.
We can keep inventing new macros: VM_BUG_ON_PAGE_AND_VM() for the case. But why not have one to rule them all? ;)
-- Kirill A. Shutemov
| |