Messages in this thread |  | | Date | Sat, 6 Dec 2025 16:47:58 +0000 | | From | Lorenzo Stoakes <> | | Subject | Re: [PATCH] mm: avoid use of BIT() macro for initialising VMA flags |
| |
On Fri, Dec 05, 2025 at 09:49:40PM +0000, David Laight wrote: > On Fri, 5 Dec 2025 19:18:56 +0000 > Lorenzo Stoakes <lorenzo.stoakes@oracle.com> wrote: > > > On Fri, Dec 05, 2025 at 06:43:42PM +0000, David Laight wrote: > > > On Fri, 5 Dec 2025 17:50:37 +0000 > > > Lorenzo Stoakes <lorenzo.stoakes@oracle.com> wrote: > > > > > > > Commit 2b6a3f061f11 ("mm: declare VMA flags by bit") significantly changed > > > > how VMA flags are declared, utilising an enum of VMA bit values and > > > > ifdef-fery VM_xxx flag declarations via macro. > > > > > > > > As part of this change, it uses INIT_VM_FLAG() to define VM_xxx flags from > > > > the newly introduced VMA bit numbers. > > > > > > > > However, use of this macro results in apparently unfortunate macro > > > > expansion and resulted in a performance degradation.This appears to be due > > > > to the (__force int), which is required for the sparse typechecking to > > > > work. > > > > > > Does sparse complain if you just add 0? As in: > > > #define INIT_VM_FLAG(name) BIT(VMA_ ## name ## _BIT + 0u) > > > > > > That should change the type without affecting what BIT() expands to. > > > > Thanks, checked that and unfortunately that doesn't satisfy sparse :) > > Oh - it is that __bitwise that causes grief.
Well, if a sparse build is not enabled this tag is just removed (as is __force).
> > > I don't think it's too crazy to use 1UL << here, just very frustrating (TM) > > that this is an issue. > > Did you try getting DECLARE_VMA_BIT to define both the bit number and the > bit flag and put them both into the anonymous enum? > Or are there other reasons for not doing that?
I did and we can't do that because it results in errors like 'enum constant in boolean context [-Werror=int-in-bool-context]' as clearly VM_xxx flags are used in many different contexts in the kernel many of which seem incompatible with enum constants (even though... they should be equivalent, at least in theory?)
> > > > > <insert rant about C macros here> > > Add rant about the compiler thinking anon enums are doing anything other > than defining constants.
Right yes :)
I put these in an enum in part to make life easier for tools like drgn to be able to find these values (the guys asked about this explicitly). But also it makes sense for them to be in an enum!
Really the VM_xxx flags are at least in theory a temporary hack until everything can use bit numbers... assuming I can find a way to do that without causing performance regressions :)
The perf regression here was - rather unexpected - however - I must say!
> > David > > > > > Cheers, Lorenzo >
Cheers, Lorenzo
|  |