On Fri, Mar 22, 2019 at 03:06:28PM +0100, Arnd Bergmann wrote:
> When CONFIG_PCI is disabled, clang gets confused about the
> control flow of the switch() statement always ending up
> in the default case, and warns:
>
> sound/pci/hda/patch_ca0132.c:7558:6: error: variable 'fw_entry' is used
> uninitialized whenever 'if' condition is false
> [-Werror,-Wsometimes-uninitialized]
> if (!spec->alt_firmware_present) {
> ^~~~~~~~~~~~~~~~~~~~~~~~~~~
> sound/pci/hda/patch_ca0132.c:7565:42: note: uninitialized use occurs here
> dsp_os_image = (struct dsp_image_seg *)(fw_entry->data);
> ^~~~~~~~
> sound/pci/hda/patch_ca0132.c:7558:2: note: remove the 'if' if its condition
> is always true
> if (!spec->alt_firmware_present) {
> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> sound/pci/hda/patch_ca0132.c:7521:33: note: initialize the variable
> 'fw_entry' to silence this warning
> const struct firmware *fw_entry;
> ^
> = NULL
>
> Adding an explicit check for CONFIG_PCI avoids the issue.
> Unfortunately this is not very intuitive here.
>
> Link: https://bugs.llvm.org/show_bug.cgi?id=41197#c1
> Signed-off-by: Arnd Bergmann <[email protected]>
Reviewed-by: Nathan Chancellor <[email protected]>
> ---
> Any suggestions for other workarounds appreciated. If you can think
> of a better fix, please treat this as a reported-by:
> ---
> sound/pci/hda/patch_ca0132.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/sound/pci/hda/patch_ca0132.c b/sound/pci/hda/patch_ca0132.c
> index 29882bda7632..415b16b7db70 100644
> --- a/sound/pci/hda/patch_ca0132.c
> +++ b/sound/pci/hda/patch_ca0132.c
> @@ -7555,7 +7555,7 @@ static bool ca0132_download_dsp_images(struct hda_codec
> *codec)
> * Use default ctefx.bin if no alt firmware is detected, or if none
> * exists for your particular codec.
> */
> - if (!spec->alt_firmware_present) {
> + if (!IS_ENABLED(CONFIG_PCI) || !spec->alt_firmware_present) {
> codec_dbg(codec, "Default firmware selected.");
> if (request_firmware(&fw_entry, EFX_FILE,
> codec->card->dev) != 0)
> --
> 2.20.0
>
