[PATCH v2 08/10] Avoid double free in startup_inferior

[Simon Marchi]

On 2020-09-03 8:29 p.m., Kamil Rytarowski wrote:
> Do not free the last execd pathname as it will be used in
> prepare_resume_reply(), after attaching a client side.

Ok, so this function returns to its caller the last waitstatus.  So indeed we
want to clean up all the watstatus objects except the last one, which we hand
over to the caller.

> gdb/ChangeLog:
> 
> 	* fork-inferior.c (startup_inferior): Avoid double free.
> ---
>  gdb/ChangeLog           | 4 ++++
>  gdb/nat/fork-inferior.c | 5 ++++-
>  2 files changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/gdb/ChangeLog b/gdb/ChangeLog
> index b96e7bf08e8..1013f6a0b3c 100644
> --- a/gdb/ChangeLog
> +++ b/gdb/ChangeLog
> @@ -1,3 +1,7 @@
> +2020-09-04  Kamil Rytarowski  <n54@gmx.com>
> +
> +	* fork-inferior.c (startup_inferior): Avoid double free.
> +
>  2020-09-04  Kamil Rytarowski  <n54@gmx.com>
> 
>  	* netbsd-nat.h (netbsd_nat::qxfer_siginfo): Add.
> diff --git a/gdb/nat/fork-inferior.c b/gdb/nat/fork-inferior.c
> index 1185ef8998b..94ab0b9cbc2 100644
> --- a/gdb/nat/fork-inferior.c
> +++ b/gdb/nat/fork-inferior.c
> @@ -526,7 +526,10 @@ startup_inferior (process_stratum_target *proc_target, pid_t pid, int ntraps,
> 
>  	  case TARGET_WAITKIND_EXECD:
>  	    /* Handle EXEC signals as if they were SIGTRAP signals.  */
> -	    xfree (ws.value.execd_pathname);
> +	    /* Do not free the last execd pathname as it will be used in
> +	       prepare_resume_reply(), after attaching a client side.  */

Since this is common code, let's just write it in a gdbserver or gdb agnostic way.

I'd suggest something like: Free the exec'ed pathname, but only if this isn't the
waitstatus we are returning to the caller.

Simon