[dpdk-dev] eal / malloc : alignment parameter check failing due to changes in rte_is_power_of_2
Commit Message
In commit 2fc8d6d the behaviour of function rte_is_power_of_2 was
changed to not return true for 0. memzone_reserve_aligned_thread_unsafe
and rte_malloc_socket both make the assumption that for align = 0
!rte_is_power_of_2(align) will return false. This patch adds a check
that align parameter is non-zero before doing the power of 2 check
Signed-off-by: Declan Doherty <declan.doherty@intel.com>
---
lib/librte_eal/common/eal_common_memzone.c | 2 +-
lib/librte_malloc/rte_malloc.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
Comments
> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Declan Doherty
> Sent: Friday, January 16, 2015 3:11 PM
> To: dev@dpdk.org
> Subject: [dpdk-dev] [PATCH] eal / malloc : alignment parameter check failing
> due to changes in rte_is_power_of_2
>
> In commit 2fc8d6d the behaviour of function rte_is_power_of_2 was
> changed to not return true for 0. memzone_reserve_aligned_thread_unsafe
> and rte_malloc_socket both make the assumption that for align = 0
> !rte_is_power_of_2(align) will return false. This patch adds a check
> that align parameter is non-zero before doing the power of 2 check
>
> Signed-off-by: Declan Doherty <declan.doherty@intel.com>
Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
2015-01-16 15:10, Declan Doherty:
> In commit 2fc8d6d the behaviour of function rte_is_power_of_2 was
> changed to not return true for 0. memzone_reserve_aligned_thread_unsafe
> and rte_malloc_socket both make the assumption that for align = 0
> !rte_is_power_of_2(align) will return false. This patch adds a check
> that align parameter is non-zero before doing the power of 2 check
>
> Signed-off-by: Declan Doherty <declan.doherty@intel.com>
[...]
> - if (!rte_is_power_of_2(align)) {
> + if (align ? !rte_is_power_of_2(align) : 0) {
[...]
> - if (size == 0 || !rte_is_power_of_2(align))
> + if (size == 0 || align ? !rte_is_power_of_2(align) : 0)
I don't understand why you write "align ? !rte_is_power_of_2(align) : 0"
instead of the more readable "align && !rte_is_power_of_2(align)" ?
Pablo acked it so I guess there is something obvious I'm missing.
> -----Original Message-----
> From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com]
> Sent: Friday, January 16, 2015 5:02 PM
> To: Doherty, Declan
> Cc: dev@dpdk.org; De Lara Guarch, Pablo
> Subject: Re: [dpdk-dev] [PATCH] eal / malloc : alignment parameter check failing
> due to changes in rte_is_power_of_2
>
> 2015-01-16 15:10, Declan Doherty:
> > In commit 2fc8d6d the behaviour of function rte_is_power_of_2 was
> > changed to not return true for 0. memzone_reserve_aligned_thread_unsafe
> > and rte_malloc_socket both make the assumption that for align = 0
> > !rte_is_power_of_2(align) will return false. This patch adds a check
> > that align parameter is non-zero before doing the power of 2 check
> >
> > Signed-off-by: Declan Doherty <declan.doherty@intel.com>
> [...]
> > - if (!rte_is_power_of_2(align)) {
> > + if (align ? !rte_is_power_of_2(align) : 0) {
> [...]
> > - if (size == 0 || !rte_is_power_of_2(align))
> > + if (size == 0 || align ? !rte_is_power_of_2(align) : 0)
>
> I don't understand why you write "align ? !rte_is_power_of_2(align) : 0"
> instead of the more readable "align && !rte_is_power_of_2(align)" ?
>
> Pablo acked it so I guess there is something obvious I'm missing.
>
> --
> Thomas
No there's nothing you're missing, this is just the way I saw the logic, if align is none
zero, then test the power of 2 condition otherwise return 0. I have no problem with
your suggestion in you prefer that, at the end of the day the logic test works out equivalent.
2015-01-18 19:26, Doherty, Declan:
> From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com]
> > 2015-01-16 15:10, Declan Doherty:
> > > In commit 2fc8d6d the behaviour of function rte_is_power_of_2 was
> > > changed to not return true for 0. memzone_reserve_aligned_thread_unsafe
> > > and rte_malloc_socket both make the assumption that for align = 0
> > > !rte_is_power_of_2(align) will return false. This patch adds a check
> > > that align parameter is non-zero before doing the power of 2 check
> > >
> > > Signed-off-by: Declan Doherty <declan.doherty@intel.com>
> > [...]
> > > - if (!rte_is_power_of_2(align)) {
> > > + if (align ? !rte_is_power_of_2(align) : 0) {
> > [...]
> > > - if (size == 0 || !rte_is_power_of_2(align))
> > > + if (size == 0 || align ? !rte_is_power_of_2(align) : 0)
There is an operator precedence bug here.
Parens are needed after ||.
> > I don't understand why you write "align ? !rte_is_power_of_2(align) : 0"
> > instead of the more readable "align && !rte_is_power_of_2(align)" ?
> >
> > Pablo acked it so I guess there is something obvious I'm missing.
>
> No there's nothing you're missing, this is just the way I saw the logic, if align is none
> zero, then test the power of 2 condition otherwise return 0. I have no problem with
> your suggestion in you prefer that, at the end of the day the logic test works out equivalent.
So I change to the simpler && form and I add parentheses
after || to correctly check size==0.
Applied with above changes.
Thanks
@@ -156,7 +156,7 @@ memzone_reserve_aligned_thread_unsafe(const char *name, size_t len,
}
/* if alignment is not a power of two */
- if (!rte_is_power_of_2(align)) {
+ if (align ? !rte_is_power_of_2(align) : 0) {
RTE_LOG(ERR, EAL, "%s(): Invalid alignment: %u\n", __func__,
align);
rte_errno = EINVAL;
@@ -75,7 +75,7 @@ rte_malloc_socket(const char *type, size_t size, unsigned align, int socket_arg)
void *ret;
/* return NULL if size is 0 or alignment is not power-of-2 */
- if (size == 0 || !rte_is_power_of_2(align))
+ if (size == 0 || align ? !rte_is_power_of_2(align) : 0)
return NULL;
if (socket_arg == SOCKET_ID_ANY)