Re: [UNIKRAFT PATCH v2 1/2] lib/ukallocpool: Always provide `uk_alloc_availmem()`, `uk_alloc_maxalloc()`

[Simon Kuenzer <simon.kuenzer@xxxxxxxxx>]

Hi Cezar,

thanks a lot for your review.

On 03.03.21 21:10, Cezar Craciunoiu wrote:
> Hey Simon,
>
> This patch looks good to me.
>
> Just one small thing. Why do these two functions return a signed value if 
> obj_len is unsigned?
This is coming from the API and for the case to return error values as 
negative numbers, e.g., -ENOTSUP.
Thanks,

Simon

> Cezar
>
> Reviewed-by: Cezar Craciunoiu <cezar.craciunoiu@xxxxxxxxx 
> [mailto:cezar.craciunoiu@xxxxxxxxx]>
> On 2020-12-22 14:57:11, Simon Kuenzer <simon.kuenzer@xxxxxxxxx> wrote:
> Adopts to the changed meaning of `CONFIG_LIBUKALLLC_IFSTATS` and provide
> `uk_alloc_availmem()` and `uk_alloc_maxalloc()` always.
>
> Signed-off-by: Simon Kuenzer
> ---
> lib/ukallocpool/pool.c | 13 ++++++++++---
> 1 file changed, 10 insertions(+), 3 deletions(-)
>
> diff --git a/lib/ukallocpool/pool.c b/lib/ukallocpool/pool.c
> index 390f6b92..0ff24c38 100644
> --- a/lib/ukallocpool/pool.c
> +++ b/lib/ukallocpool/pool.c
> @@ -208,14 +208,19 @@ void uk_allocpool_return_batch(struct uk_allocpool *p,
> _prepend_free_obj(p, obj[i]);
> }
>
> -#if CONFIG_LIBUKALLOC_IFSTATS
> static ssize_t pool_availmem(struct uk_alloc *a)
> {
> struct uk_allocpool *p = ukalloc2pool(a);
>
> - return (size_t) p->free_obj_count * p->obj_len;
> + return (ssize_t) (p->free_obj_count * p->obj_len);
> +}
> +
> +static ssize_t pool_maxalloc(struct uk_alloc *a)
> +{
> + struct uk_allocpool *p = ukalloc2pool(a);
> +
> + return (ssize_t) p->obj_len;
> }
> -#endif
>
> size_t uk_allocpool_reqmem(unsigned int obj_count, size_t obj_len,
> size_t obj_align)
> @@ -300,6 +305,8 @@ out:
> pool_free,
> pool_posix_memalign,
> uk_memalign_compat,
> + pool_maxalloc,
> + pool_availmem,
> NULL);
> #if CONFIG_LIBUKALLOC_IFSTATS
> p->self.availmem = pool_availmem;
> --
> 2.20.1