On Fri, Mar 13, 2009 at 10:00:07AM -0700, Pallipadi, Venkatesh wrote: > On Fri, 2009-03-13 at 09:25 -0700, Nick Piggin wrote: > > Why not just use another vm_flag for 2.6.29 and earlier (we have 2 left), > > then do some flag rationalisation in 2.6.30? > > > > But I do like just avoiding these games entirely and encoding it > > explicitly. > > > > Totally agree that this is ugly. But, I didn't like taking away one bit > for such a light usage. VM_INSERTPAGE is hardly ever used in the kernel > (wherever it is used, it is checked along with VM_PFNMAP). That is a concern, but I think it makes the patch more intrusive than a simple fix that might possibly introduce a bug. So for 2.6.29 we have spare flags so can use one of those for the fix. > OK. I will resend the patch with taking away one of the available bits. > But, I feel we need to do some multiplexing here, in a clean manner, for > future. That would always be appreciated. There are probably several flags that can go away (eg. VM_RESERVED, VM_INSERTPAGE), others that can be merged (eg nommu flags with mmu flags), others that can be one scalar value rather than several binary values etc. Also, if worst comes to worst, I think we could probably make vm_flags 64-bit on 32-bit compiles to get another 32 bits. Of course I'd rather not until no other options left for reducing flags ;) -- To unsubscribe from this list: send the line "unsubscribe linux-tip-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html