Hi,
On Wed, 5 Jul 2017 00:19:28 +0200 stefan.bruens@rwth-aachen.de wrote:
On Dienstag, 4. Juli 2017 19:05:25 CEST Pantelis Antoniou wrote:
Hi Marek,
On Sat, 2017-07-01 at 16:11 +0200, Marek Vasut wrote:
On 06/30/2017 06:23 PM, Pantelis Antoniou wrote:
Introduce an overlay based method for constructing a base DT blob to pass to the kernel.
Both canned and runtime feature selection is supported.
Signed-off-by: Pantelis Antoniou pantelis.antoniou@konsulko.com
[...]
@@ -1475,29 +1482,58 @@ int fit_conf_get_node(const void *fit, const char *conf_uname)> > debug("Found default configuration: '%s'\n", conf_uname);
}
- s = strchr(conf_uname, '#');
- if (s) {
len = s - conf_uname;conf_uname_copy = malloc(len + 1);if (!conf_uname_copy) {debug("Can't allocate uname copy: '%s'\n",conf_uname);return -ENOMEM;}memcpy(conf_uname_copy, conf_uname, len);Is that like strdup() here ?
No. The space allocated is not the full string, just the span until the next #.
You could use strndup() to copy only the first n characters.
or strsep() to tokenize (a copy of) the original string. IMO using strdup() and strsep() makes the intent of this function clearer, than the current code.
Lothar Waßmann