On Tue, Aug 27, 2024 at 05:53:30PM -0500, Segher Boessenkool wrote:
> On Tue, Aug 27, 2024 at 11:08:19AM -0700, Eric Biggers wrote:
> > On Thu, Aug 22, 2024 at 09:13:13AM +0200, Christophe Leroy wrote:
> > > With the current implementation, __cvdso_getrandom_data() calls
> > > memset(), which is unexpected in the VDSO.
> > >
> > > Rewrite opaque data initialisation to avoid memset().
> > >
> > > Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxxxxxx>
> > > ---
> > > lib/vdso/getrandom.c | 15 ++++++++++-----
> > > 1 file changed, 10 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/lib/vdso/getrandom.c b/lib/vdso/getrandom.c
> > > index cab153c5f9be..4a56f45141b4 100644
> > > --- a/lib/vdso/getrandom.c
> > > +++ b/lib/vdso/getrandom.c
> > > @@ -4,6 +4,7 @@
> > > */
> > >
> > > #include <linux/minmax.h>
> > > +#include <linux/array_size.h>
> > > #include <vdso/datapage.h>
> > > #include <vdso/getrandom.h>
> > > #include <vdso/unaligned.h>
> > > @@ -74,11 +75,15 @@ __cvdso_getrandom_data(const struct vdso_rng_data *rng_info, void *buffer, size_
> > > u32 counter[2] = { 0 };
> > >
> > > if (unlikely(opaque_len == ~0UL && !buffer && !len && !flags)) {
> > > - *(struct vgetrandom_opaque_params *)opaque_state = (struct vgetrandom_opaque_params) {
> > > - .size_of_opaque_state = sizeof(*state),
> > > - .mmap_prot = PROT_READ | PROT_WRITE,
> > > - .mmap_flags = MAP_DROPPABLE | MAP_ANONYMOUS
> > > - };
> > > + struct vgetrandom_opaque_params *params = opaque_state;
> > > + int i;
> > > +
> > > + params->size_of_opaque_state = sizeof(*state);
> > > + params->mmap_prot = PROT_READ | PROT_WRITE;
> > > + params->mmap_flags = MAP_DROPPABLE | MAP_ANONYMOUS;
> > > + for (i = 0; i < ARRAY_SIZE(params->reserved); i++)
> > > + params->reserved[i] = 0;
> > > +
> > > return 0;
> > > }
> >
> > Is there a compiler flag that could be used to disable the generation of calls
> > to memset?
>
> -fno-tree-loop-distribute-patterns . But, as always, read up on it, see
> what it actually does (and how it avoids your problem, and mostly: learn
> what the actual problem *was*!)
This might help with various loops, but it doesn't help with the matter
that this patch fixes, which is struct initialization. I just tried it
with the arm64 patch to no avail.
