Re: [PATCH] stapsdt: fix out-of-bounds access when provider name is not specified

Reviewed-by: Eugene Loh <[email protected]>

I'm assuming no test is practical here... it all depends on what garbage 
one finds in front of pdp->prv?

On 8/29/25 14:46, Kris Van Hees wrote:
> Signed-off-by: Kris Van Hees <[email protected]>
> ---
>   libdtrace/dt_pid.c | 6 +++++-
>   1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/libdtrace/dt_pid.c b/libdtrace/dt_pid.c
> index 4af9141aa..ffc52132f 100644
> --- a/libdtrace/dt_pid.c
> +++ b/libdtrace/dt_pid.c
> @@ -1518,10 +1518,14 @@ dt_pid_create_stapsdt_probes(dtrace_probedesc_t *pdp, dtrace_hdl_t *dtp, dt_pcb_
>   	dt_proc_t *dpr = NULL;
>   	const char *pidstr;
>   	pid_t pid;
> +	size_t len = strlen(pdp->prv);
> +
> +	if (len == 0)
> +		return 0;
>   
>   	assert(pcb != NULL);
>   
> -	pidstr = &pdp->prv[strlen(pdp->prv)];
> +	pidstr = &pdp->prv[len];
>   
>   	while (isdigit(*(pidstr - 1)))
>   		pidstr--;