Re: [DTrace-devel] [PATCH 12/19] Increase size of BPF probes map

[Sam James <sam@gentoo.org>]

"eugene.loh--- via DTrace-devel" <dtrace-devel@oss.oracle.com> writes:

> From: Eugene Loh <eugene.loh@oracle.com>
>
> We are going to support discovery of new probes after dtrace is
> launched.  For example, there could be a pid or USDT probe with a
> wildcard pid specification that could be matched by a process that
> has not yet started.  Or, a process could load a shared library.
>
> This means the probe map will have to grow.  For now, just wire in
> some headroom, as ugly a hack as this is.
>
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
> ---
>  libdtrace/dt_bpf.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/libdtrace/dt_bpf.c b/libdtrace/dt_bpf.c
> index 70803e3c..75e48962 100644
> --- a/libdtrace/dt_bpf.c
> +++ b/libdtrace/dt_bpf.c
> @@ -920,7 +920,7 @@ gmap_create_probes(dtrace_hdl_t *dtp)
>  
>  	fd = create_gmap(dtp, "probes", BPF_MAP_TYPE_HASH, sizeof(uint32_t),
>  			 sizeof(dt_bpf_probe_t),
> -			 dt_list_length(&dtp->dt_enablings));
> +			 dt_list_length(&dtp->dt_enablings) + 1000);

I think this might merit a FIXME comment. Also, maybe a static_assert
(or a #define for now with a comment above it saying it's a temporary
hack) to ensure it matches the + 1000 from 'Create the BPF usdt_prids map'?

>  	if (fd == -1)
>  		return -1;