[PATCH] MIPS: asm/debug_ll_ns16550.h: add nop in delay slot
Antony Pavlov
antonynpavlov at gmail.com
Tue Sep 8 07:39:53 EDT 2020
On Tue, 8 Sep 2020 14:25:05 +0800
Du Huanpeng <u74147 at gmail.com> wrote:
Hi!
Adding nop is not necessary.
Please see full text of debug_ll_outc_a0:
.macro debug_ll_outc_a0
#ifdef CONFIG_DEBUG_LL
.set push
.set reorder <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
la t0, DEBUG_LL_UART_ADDR
201: lbu t1, UART_LSR(t0) /* get line status */
andi t1, t1, UART_LSR_THRE /* check for transmitter empty */
beqz t1, 201b /* try again */
sb a0, UART_THR(t0) /* write the character */
.set pop
#endif /* CONFIG_DEBUG_LL */
.endm
There is '.set reorder' so assembler has insert nop after beqz automatically.
I have compiled barebox for qemu-malta with DEBUG_LL enabled.
barebox$ mips-linux-gnu-objdump -D arch/mips/boards/qemu-malta/lowlevel.pbl.o | grep -A 10 "_nmon_outc_a0>:$"
00000314 <_nmon_outc_a0>:
314: 3c08b800 lui t0,0xb800
318: 350803f8 ori t0,t0,0x3f8
31c: 91090005 lbu t1,5(t0)
320: 31290020 andi t1,t1,0x20
324: 1120fffd beqz t1,31c <_nmon_outc_a0+0x8>
328: 00000000 nop <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
32c: a1040000 sb a0,0(t0)
330: 03e00008 jr ra
334: 00000000 nop
> In the debug_ll_outc_a0 macro, a0 is written to the uart data
> register many times in a delay slot when the uart is not ready
> to send data. insert a nop to avoid this.
>
> Signed-off-by: Du Huanpeng <u74147 at gmail.com>
> ---
> arch/mips/include/asm/debug_ll_ns16550.h | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/arch/mips/include/asm/debug_ll_ns16550.h b/arch/mips/include/asm/debug_ll_ns16550.h
> index df58c4c..cf666d1 100644
> --- a/arch/mips/include/asm/debug_ll_ns16550.h
> +++ b/arch/mips/include/asm/debug_ll_ns16550.h
> @@ -88,6 +88,7 @@ static inline void PUTC_LL(char ch)
> 201: lbu t1, UART_LSR(t0) /* get line status */
> andi t1, t1, UART_LSR_THRE /* check for transmitter empty */
> beqz t1, 201b /* try again */
> + nop
>
> sb a0, UART_THR(t0) /* write the character */
>
> --
> 2.7.4
>
>
> _______________________________________________
> barebox mailing list
> barebox at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/barebox
--
Best regards,
Antony Pavlov
More information about the barebox
mailing list