Sorry for the late review... Ah! On Tue, Jun 11, 2019 at 02:26:25PM +0200, Takashi Iwai wrote: > @@ -354,7 +454,12 @@ module_param_string(path, fw_path_para, > sizeof(fw_path_para), 0644); > MODULE_PARM_DESC(path, "customized firmware image search path with a higher > priority than default path"); > > static int > -fw_get_filesystem_firmware(struct device *device, struct fw_priv *fw_priv) > +fw_get_filesystem_firmware(struct device *device, struct fw_priv *fw_priv, > + const char *suffix, > + int (*decompress)(struct device *dev, > + struct fw_priv *fw_priv, > + size_t in_size, > + const void *in_buffer))
I *think* this could be cleaner, I'll elaborate below. > @@ -645,7 +768,13 @@ _request_firmware(const struct firmware **firmware_p, > const char *name, > if (ret <= 0) /* error or already assigned */ > goto out; > > - ret = fw_get_filesystem_firmware(device, fw->priv); > + ret = fw_get_filesystem_firmware(device, fw->priv, "", NULL); > +#ifdef CONFIG_FW_LOADER_COMPRESS > + if (ret == -ENOENT) > + ret = fw_get_filesystem_firmware(device, fw->priv, ".xz", > + fw_decompress_xz); > +#endif Hrm, and let more #ifdef'ery. And so if someone wants to add bzip, we'd add yet-another if else on the return value of this call... and yet more #ifdefs. We already have a list of paths supported. It seems what we need instead is a list of supported suffixes, and a respective structure which then has its set of callbacks for posthandling. This way, this could all be handled inside fw_get_filesystem_firmware() neatly, and we can just strive towards avoiding #ifdef'ery. Since I'm late to review, this could be done in the future, but I do think something along these lines would make the code more maintainable and extensible. Luis
