summaryrefslogtreecommitdiff
path: root/apps/plugin.c
diff options
context:
space:
mode:
authorDave Chapman <dave@dchapman.com>2007-05-07 22:30:23 +0000
committerDave Chapman <dave@dchapman.com>2007-05-07 22:30:23 +0000
commitf4a9dab4d161a8f59800656417baf38e6f163f9d (patch)
treeab2b89c90480e1dbae5c83ece53a3ab7c116adf0 /apps/plugin.c
parent173b611921db2d353d977dc617a8288345f5e6ed (diff)
downloadrockbox-f4a9dab4d161a8f59800656417baf38e6f163f9d.zip
rockbox-f4a9dab4d161a8f59800656417baf38e6f163f9d.tar.gz
rockbox-f4a9dab4d161a8f59800656417baf38e6f163f9d.tar.bz2
rockbox-f4a9dab4d161a8f59800656417baf38e6f163f9d.tar.xz
Use a separate thread in test_codec, with the same (IRAM) stack as the main codec thread. Add a function to the plugin API to steal the codec thread, which test_codec copies and then restores. Now libmad can be benchmarked.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@13347 a1c6a512-1295-4272-9138-f99709370657
Diffstat (limited to 'apps/plugin.c')
-rw-r--r--apps/plugin.c1
1 files changed, 1 insertions, 0 deletions
diff --git a/apps/plugin.c b/apps/plugin.c
index 9e17c23..a7832ef 100644
--- a/apps/plugin.c
+++ b/apps/plugin.c
@@ -493,6 +493,7 @@ static const struct plugin_api rockbox_api = {
codec_load_file,
get_metadata,
get_codec_filename,
+ steal_codec_stack,
#endif
};