[PATCH][RFC] 8250_dw: unregister dw8250_set_termios for rk3188 chip
Heiko Stuebner
heiko at sntech.de
Thu Nov 23 07:11:12 PST 2017
Hi,
Am Donnerstag, 23. November 2017, 22:56:22 CET schrieb Shuyu Wei:
> On Tue, Nov 21, 2017 at 05:33:47PM +0000, Ed Blake wrote:
> > On 21/11/17 17:17, Shuyu Wei wrote:
> > >
> > >> What baud rate is being set, and what rate does clk_round_rate() return
> > >> in dw8250_set_termios()?
> > > I'm using baud 115200, clk_round_rate() returns 1843200.
> >
> > So dw8250_set_termios() requests a rate of 1843200 (baud * 16) from
> > clk_round_rate(), which returns the exact frequency requested. If it
> > really is capable of being set to that frequency, I can't see why this
> > wouldn't work. Or clk_round_rate() is incorrectly reporting that it's
> > capable of that rate, in which case the clock driver needs fixing.
> >
> > Do you have any way of verifying the actual clock rate going into the UART?
>
> Sorry, I know little about the clock system, and no equipment to do that,
> maybe Heiko knows.
>
> The /sys/kernel/debug/clk/clk_summary is attached below, might be helpful.
>
>
> clock enable_cnt prepare_cnt rate accuracy phase
> ----------------------------------------------------------------------------------------
> xin32k 0 0 32768 0 0
> xin24m 12 12 24000000 0 0
> timer6 1 1 24000000 0 0
> timer5 0 0 24000000 0 0
> timer4 0 0 24000000 0 0
> timer3 1 1 24000000 0 0
> timer2 0 0 24000000 0 0
> pll_gpll 1 1 891000000 0 0
> gpll 2 2 891000000 0 0
> i2s_src 0 0 891000000 0 0
> i2s0_pre 0 0 891000000 0 0
> i2s0_frac 0 0 44550000 0 0
> spdif_pre 0 0 891000000 0 0
> aclk_cpu_pre 3 3 297000000 0 0
> hclk_cpu_pre 2 2 148500000 0 0
> hclk_ahb2apb 2 2 74250000 0 0
> pclk_uart1 1 1 74250000 0 0
> pclk_uart0 1 1 74250000 0 0
> hclk_cpu 2 2 148500000 0 0
> hclk_imem1 0 0 148500000 0 0
> hclk_imem0 0 0 148500000 0 0
> hclk_rga 0 0 148500000 0 0
> hclk_ipp 0 0 148500000 0 0
> hclk_cif0 0 0 148500000 0 0
> hclk_lcdc1 0 0 148500000 0 0
> hclk_lcdc0 0 0 148500000 0 0
> hclk_vio_bus 0 0 148500000 0 0
> hclk_cpubus 1 1 148500000 0 0
> hclk_spdif 1 2 148500000 0 0
> hclk_i2s0 0 0 148500000 0 0
> hclk_rom 0 0 148500000 0 0
> pclk_cpu_pre 1 1 37125000 0 0
> atclk_cpu 0 0 37125000 0 0
> trace 0 0 37125000 0 0
> atclk 0 0 37125000 0 0
> pclk_cpu 4 8 37125000 0 0
> pclk_timer3 1 1 37125000 0 0
> pclk_pmu 0 0 37125000 0 0
> pclk_grf 0 0 37125000 0 0
> pclk_dbg 0 0 37125000 0 0
> pclk_ddrpubl 0 0 37125000 0 0
> pclk_ddrupctl 0 0 37125000 0 0
> pclk_tzpc 0 0 37125000 0 0
> pclk_efuse 0 0 37125000 0 0
> pclk_gpio2 0 1 37125000 0 0
> pclk_gpio1 0 1 37125000 0 0
> pclk_gpio0 2 1 37125000 0 0
> pclk_i2c1 0 2 37125000 0 0
> pclk_i2c0 0 0 37125000 0 0
> pclk_timer0 1 1 37125000 0 0
> pclk_pwm01 0 2 37125000 0 0
> aclk_cpu 2 2 297000000 0 0
> aclk_strc_sys 0 0 297000000 0 0
> aclk_intmem 0 0 297000000 0 0
> aclk_dma1 1 1 297000000 0 0
> gpll_armclk 1 1 891000000 0 0
> gpll_ddr 0 0 891000000 0 0
> hsadc_src 0 0 89100000 0 0
> sclk_hsadc_out 0 0 89100000 0 0
> sclk_hsadc 0 0 89100000 0 0
> hsadc_frac 0 0 4455000 0 0
> uart_src 0 0 891000000 0 0
> uart3_pre 0 0 891000000 0 0
> uart3_frac 0 0 44550000 0 0
> uart2_pre 0 0 891000000 0 0
> uart2_frac 0 0 44550000 0 0
> uart1_pre 0 0 891000000 0 0
> uart1_frac 0 0 44550000 0 0
> uart0_pre 0 0 891000000 0 0
> uart0_frac 0 0 44550000 0 0
> ..............
> more output is omitted.
you actually omitted the output part where sclk_uart2 is actually shown :-) .
On my rk3188 radxarock with a kernel build this morning from
the middle of this merge-window, the relevant part of the clock-tree
looks like the following and my serial console works like a charm:
xin24m 6 6 24000000 0 0
[...]
pll_gpll 1 1 594000000 0 0
gpll 5 5 594000000 0 0
[...]
uart_src 1 1 594000000 0 0
uart3_pre 0 0 594000000 0 0
uart3_frac 0 0 29700000 0 0
uart2_pre 1 1 594000000 0 0
uart2_frac 1 1 1843200 0 0
sclk_uart2 1 1 1843200 0 0
[ ^^ the important clock]
In your dump the sclk_uart2 clock is not muxed to the uart2_frac clock
but to something else but that part is missing from you dump.
So clk_round_rate is definitly correct in that it can reach this rate
using the fractional divider and also can sucessfully set this in the
clock framework.
Can you show where sclk_uart2 is for you please, as I guess your dump
is with the settermios patch disabled, right?
Thanks
Heiko
More information about the Linux-rockchip
mailing list