Hi John,
On Mon, Jun 25, 2018 at 07:15:29PM +0200, John Crispin wrote:
> diff --git a/arch/mips/ath79/common.c b/arch/mips/ath79/common.c
> index cd6055f9e7a0..fc3438150b3e 100644
> --- a/arch/mips/ath79/common.c
> +++ b/arch/mips/ath79/common.c
> @@ -110,7 +110,7 @@ void ath79_device_reset_set(u32 mask)
> else if (soc_is_qca956x() || soc_is_tp9343())
> reg = QCA956X_RESET_REG_RESET_MODULE;
> else
> - BUG();
> + panic("Reset register not defined for this SOC");
>
> spin_lock_irqsave(&ath79_device_reset_lock, flags);
> t = ath79_reset_rr(reg);
> @@ -142,7 +142,7 @@ void ath79_device_reset_clear(u32 mask)
> else if (soc_is_qca956x() || soc_is_tp9343())
> reg = QCA956X_RESET_REG_RESET_MODULE;
> else
> - BUG();
> + panic("Reset register not defined for this SOC");
>
> spin_lock_irqsave(&ath79_device_reset_lock, flags);
> t = ath79_reset_rr(reg);
This was rejected when first submitted nearly back in 2013:
https://patchwork.linux-mips.org/patch/5742/
...and indeed I agree with Ralf's message from back then that it seems
odd for GCC to now be recognizing that code beyond a BUG() is
unreachable.
Could you specify which version(s) of GCC are problematic?
Thanks,
Paul
