Re: [PATCH v4 4/5] trace-cmd: Add validation for reading and writing trace.dat files

[Steven Rostedt <rostedt@goodmis.org>]

On Fri, 26 Feb 2021 14:13:05 +0200
"Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com> wrote:

> @@ -1409,7 +1497,6 @@ struct tracecmd_output *tracecmd_get_output_handle_fd(int fd)
>  {
>  	struct tracecmd_output *handle = NULL;
>  	struct tracecmd_input *ihandle;
> -	struct tep_handle *pevent;
>  	int fd2;
>  
>  	/* Move the file descriptor to the beginning */
> @@ -1425,6 +1512,7 @@ struct tracecmd_output *tracecmd_get_output_handle_fd(int fd)
>  	ihandle = tracecmd_alloc_fd(fd2, TRACECMD_FL_LOAD_NO_PLUGINS);
>  	if (!ihandle)
>  		return NULL;
> +	tracecmd_read_headers(ihandle);
>  
>  	/* move the file descriptor to the end */
>  	if (lseek(fd, 0, SEEK_END) == (off_t)-1)
> @@ -1437,11 +1525,11 @@ struct tracecmd_output *tracecmd_get_output_handle_fd(int fd)
>  
>  	handle->fd = fd;
>  
> -	/* get endian and page size */
> -	pevent = tracecmd_get_tep(ihandle);

Actually this clean up of removing the duplicate pevent should have been a
separate patch as well. I'll take this as is (as the patches seem to be
solid), but for the future, try to keep small clean ups like this as
separate patches. One advantage is, if someone is supporting an older
version of trace-cmd, they may want to backport the clean up, but not the
feature.

-- Steve