diff mbox series

[v3,1/1] tty: serial: Use fifo in 8250 console driver

Message ID 20211222112831.1968392-2-wander@redhat.com
State Superseded
Headers show
Series [v3,1/1] tty: serial: Use fifo in 8250 console driver | expand

Commit Message

Wander Lairson Costa Dec. 22, 2021, 11:28 a.m. UTC
Note: I am using a small test app + driver located at [0] for the
problem description. serco is a driver whose write function dispatches
to the serial controller. sertest is a user-mode app that writes n bytes
to the serial console using the serco driver.

While investigating a bug in the RHEL kernel, I noticed that the serial
console throughput is way below the configured speed of 115200 bps in
a HP Proliant DL380 Gen9. I was expecting something above 10KB/s, but
I got 2.5KB/s.

$ time ./sertest -n 2500 /tmp/serco

real    0m0.997s
user    0m0.000s
sys     0m0.997s

With the help of the function tracer, I then noticed the serial
controller was taking around 410us seconds to dispatch one single byte:

$ trace-cmd record -p function_graph -g serial8250_console_write \
   ./sertest -n 1 /tmp/serco

$ trace-cmd report

            |  serial8250_console_write() {
 0.384 us   |    _raw_spin_lock_irqsave();
 1.836 us   |    io_serial_in();
 1.667 us   |    io_serial_out();
            |    uart_console_write() {
            |      serial8250_console_putchar() {
            |        wait_for_xmitr() {
 1.870 us   |          io_serial_in();
 2.238 us   |        }
 1.737 us   |        io_serial_out();
 4.318 us   |      }
 4.675 us   |    }
            |    wait_for_xmitr() {
 1.635 us   |      io_serial_in();
            |      __const_udelay() {
 1.125 us   |        delay_tsc();
 1.429 us   |      }
...
...
...
 1.683 us   |      io_serial_in();
            |      __const_udelay() {
 1.248 us   |        delay_tsc();
 1.486 us   |      }
 1.671 us   |      io_serial_in();
 411.342 us |    }

In another machine, I measured a throughput of 11.5KB/s, with the serial
controller taking between 80-90us to send each byte. That matches the
expected throughput for a configuration of 115200 bps.

This patch changes the serial8250_console_write to use the 16550 fifo
if available. In my benchmarks I got around 25% improvement in the slow
machine, and no performance penalty in the fast machine.

Signed-off-by: Wander Lairson Costa <wander@redhat.com>
---
 drivers/tty/serial/8250/8250_port.c | 61 ++++++++++++++++++++++++++---
 1 file changed, 55 insertions(+), 6 deletions(-)

Comments

Sebastian Andrzej Siewior Jan. 25, 2022, 1:07 p.m. UTC | #1
On 2021-12-22 08:28:30 [-0300], Wander Lairson Costa wrote:
> Note: I am using a small test app + driver located at [0] for the
> problem description. serco is a driver whose write function dispatches
> to the serial controller. sertest is a user-mode app that writes n bytes
> to the serial console using the serco driver.
> 
> While investigating a bug in the RHEL kernel, I noticed that the serial
> console throughput is way below the configured speed of 115200 bps in
> a HP Proliant DL380 Gen9. I was expecting something above 10KB/s, but
> I got 2.5KB/s.
> 
> $ time ./sertest -n 2500 /tmp/serco
> 
> real    0m0.997s
> user    0m0.000s
> sys     0m0.997s
> 
> With the help of the function tracer, I then noticed the serial
> controller was taking around 410us seconds to dispatch one single byte:

was this the HW access or did this include the wait-for-fifo empty?

> $ trace-cmd record -p function_graph -g serial8250_console_write \
>    ./sertest -n 1 /tmp/serco
> 
> $ trace-cmd report
> 
>             |  serial8250_console_write() {
>  0.384 us   |    _raw_spin_lock_irqsave();
>  1.836 us   |    io_serial_in();
>  1.667 us   |    io_serial_out();
>             |    uart_console_write() {
>             |      serial8250_console_putchar() {
>             |        wait_for_xmitr() {
>  1.870 us   |          io_serial_in();
>  2.238 us   |        }
>  1.737 us   |        io_serial_out();
>  4.318 us   |      }
>  4.675 us   |    }
>             |    wait_for_xmitr() {
>  1.635 us   |      io_serial_in();
>             |      __const_udelay() {
>  1.125 us   |        delay_tsc();
>  1.429 us   |      }
> ...
> ...
> ...
>  1.683 us   |      io_serial_in();
>             |      __const_udelay() {
>  1.248 us   |        delay_tsc();
>  1.486 us   |      }
>  1.671 us   |      io_serial_in();
>  411.342 us |    }

So this includes waiting for empty slot. It is wait_for_xmitr() only.

> In another machine, I measured a throughput of 11.5KB/s, with the serial
> controller taking between 80-90us to send each byte. That matches the
> expected throughput for a configuration of 115200 bps.
> 
> This patch changes the serial8250_console_write to use the 16550 fifo
> if available. In my benchmarks I got around 25% improvement in the slow
> machine, and no performance penalty in the fast machine.

Either the HW is slow on starting to work, or…

What I noticed now in -rc1 is this output during boot:

|[    6.370196] ACPI: button: Power Button [PWRF]
|[    6.443501] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
|[0I 15
|      [0I 15
|            [No
|               [ld
|                  [a2
|                     [a20tm
|                           [a2nct
|                                 [s
|[s
|[s
|[s
…
|[sk65,
|      [rt
|         [Pe
|            [a
|               [    6.873611] ata1: SATA link down (SStatus 0 SControl 300)
|[    6.879680] ata3: SATA link down (SStatus 0 SControl 300)

The kernel buffer reports here:

|[    6.370196] ACPI: button: Power Button [PWRF]
|[    6.443501] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
|[    6.450643] 00:03: ttyS0 at I/O 0x3f8 (irq = 4, base_baud = 115200) is a 16550A
|[    6.451625] 00:04: ttyS1 at I/O 0x2f8 (irq = 3, base_baud = 115200) is a 16550A
|[    6.453808] Non-volatile memory driver v1.3
|[    6.475688] loop: module loaded
|[    6.476401] ahci 0000:00:1f.2: version 3.0
|[    6.487238] ahci 0000:00:1f.2: AHCI 0001.0300 32 slots 6 ports 6 Gbps 0x3f impl SATA mode

I did remove the last few lines but it appears that since the
initialisation of the port some of the lines got lost.

Do you see the same?

> Signed-off-by: Wander Lairson Costa <wander@redhat.com>

Sebastian
Wander Costa Jan. 25, 2022, 1:15 p.m. UTC | #2
On Tue, Jan 25, 2022 at 10:07 AM Sebastian Andrzej Siewior
<bigeasy@linutronix.de> wrote:
>
> On 2021-12-22 08:28:30 [-0300], Wander Lairson Costa wrote:
> > Note: I am using a small test app + driver located at [0] for the
> > problem description. serco is a driver whose write function dispatches
> > to the serial controller. sertest is a user-mode app that writes n bytes
> > to the serial console using the serco driver.
> >
> > While investigating a bug in the RHEL kernel, I noticed that the serial
> > console throughput is way below the configured speed of 115200 bps in
> > a HP Proliant DL380 Gen9. I was expecting something above 10KB/s, but
> > I got 2.5KB/s.
> >
> > $ time ./sertest -n 2500 /tmp/serco
> >
> > real    0m0.997s
> > user    0m0.000s
> > sys     0m0.997s
> >
> > With the help of the function tracer, I then noticed the serial
> > controller was taking around 410us seconds to dispatch one single byte:
>
> was this the HW access or did this include the wait-for-fifo empty?
>
> > $ trace-cmd record -p function_graph -g serial8250_console_write \
> >    ./sertest -n 1 /tmp/serco
> >
> > $ trace-cmd report
> >
> >             |  serial8250_console_write() {
> >  0.384 us   |    _raw_spin_lock_irqsave();
> >  1.836 us   |    io_serial_in();
> >  1.667 us   |    io_serial_out();
> >             |    uart_console_write() {
> >             |      serial8250_console_putchar() {
> >             |        wait_for_xmitr() {
> >  1.870 us   |          io_serial_in();
> >  2.238 us   |        }
> >  1.737 us   |        io_serial_out();
> >  4.318 us   |      }
> >  4.675 us   |    }
> >             |    wait_for_xmitr() {
> >  1.635 us   |      io_serial_in();
> >             |      __const_udelay() {
> >  1.125 us   |        delay_tsc();
> >  1.429 us   |      }
> > ...
> > ...
> > ...
> >  1.683 us   |      io_serial_in();
> >             |      __const_udelay() {
> >  1.248 us   |        delay_tsc();
> >  1.486 us   |      }
> >  1.671 us   |      io_serial_in();
> >  411.342 us |    }
>
> So this includes waiting for empty slot. It is wait_for_xmitr() only.
>
> > In another machine, I measured a throughput of 11.5KB/s, with the serial
> > controller taking between 80-90us to send each byte. That matches the
> > expected throughput for a configuration of 115200 bps.
> >
> > This patch changes the serial8250_console_write to use the 16550 fifo
> > if available. In my benchmarks I got around 25% improvement in the slow
> > machine, and no performance penalty in the fast machine.
>
> Either the HW is slow on starting to work, or…
>
> What I noticed now in -rc1 is this output during boot:
>
> |[    6.370196] ACPI: button: Power Button [PWRF]
> |[    6.443501] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
> |[0I 15
> |      [0I 15
> |            [No
> |               [ld
> |                  [a2
> |                     [a20tm
> |                           [a2nct
> |                                 [s
> |[s
> |[s
> |[s
> …
> |[sk65,
> |      [rt
> |         [Pe
> |            [a
> |               [    6.873611] ata1: SATA link down (SStatus 0 SControl 300)
> |[    6.879680] ata3: SATA link down (SStatus 0 SControl 300)
>
> The kernel buffer reports here:
>
> |[    6.370196] ACPI: button: Power Button [PWRF]
> |[    6.443501] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
> |[    6.450643] 00:03: ttyS0 at I/O 0x3f8 (irq = 4, base_baud = 115200) is a 16550A
> |[    6.451625] 00:04: ttyS1 at I/O 0x2f8 (irq = 3, base_baud = 115200) is a 16550A
> |[    6.453808] Non-volatile memory driver v1.3
> |[    6.475688] loop: module loaded
> |[    6.476401] ahci 0000:00:1f.2: version 3.0
> |[    6.487238] ahci 0000:00:1f.2: AHCI 0001.0300 32 slots 6 ports 6 Gbps 0x3f impl SATA mode
>
> I did remove the last few lines but it appears that since the
> initialisation of the port some of the lines got lost.
>
> Do you see the same?

There is another thread[1] reporting some issues with this patch.
There, this diff seems to fix the problems, could you please try and
report if it works for you too?

diff --git a/drivers/tty/serial/8250/8250_port.c
b/drivers/tty/serial/8250/8250_port.c
index 2abb3de11a48..d3a93e5d55f7 100644
--- a/drivers/tty/serial/8250/8250_port.c
+++ b/drivers/tty/serial/8250/8250_port.c
@@ -3343,7 +3343,7 @@ static void serial8250_console_fifo_write(struct
uart_8250_port *up,
 {
        int i;
        const char *end = s + count;
-       unsigned int fifosize = up->port.fifosize;
+       unsigned int fifosize = up->tx_loadsz;
        bool cr_sent = false;

        while (s != end) {
@@ -3409,8 +3409,8 @@ void serial8250_console_write(struct
uart_8250_port *up, const char *s,
        }

        use_fifo = (up->capabilities & UART_CAP_FIFO) &&
-               port->fifosize > 1 &&
-               (serial_port_in(port, UART_FCR) & UART_FCR_ENABLE_FIFO) &&
+               up->tx_loadsz > 1 &&
+               (up->fcr & UART_FCR_ENABLE_FIFO) &&
                /*
                 * After we put a data in the fifo, the controller will send
                 * it regardless of the CTS state. Therefore, only use fifo

[1] https://lore.kernel.org/all/fa42a60c-954a-acc0-3962-f00427153f78@nvidia.com/


>
> > Signed-off-by: Wander Lairson Costa <wander@redhat.com>
>
> Sebastian
>
Sebastian Andrzej Siewior Jan. 25, 2022, 2:11 p.m. UTC | #3
On 2022-01-25 10:15:04 [-0300], Wander Costa wrote:
> There is another thread[1] reporting some issues with this patch.
> There, this diff seems to fix the problems, could you please try and
> report if it works for you too?

Nope. Still there.

Sebastian
Wander Costa Jan. 25, 2022, 2:26 p.m. UTC | #4
On Tue, Jan 25, 2022 at 11:11 AM Sebastian Andrzej Siewior
<bigeasy@linutronix.de> wrote:
>
> On 2022-01-25 10:15:04 [-0300], Wander Costa wrote:
> > There is another thread[1] reporting some issues with this patch.
> > There, this diff seems to fix the problems, could you please try and
> > report if it works for you too?
>
> Nope. Still there.

I will check on my side if I see something similar. Do you still get
lines lost as well?
Sebastian Andrzej Siewior Jan. 25, 2022, 2:34 p.m. UTC | #5
On 2022-01-25 11:26:03 [-0300], Wander Costa wrote:
> I will check on my side if I see something similar. Do you still get
> lines lost as well?

Yes. The output is garbage for a while and then it fixes itself. Setting
"use_fifo = 0" works.
Thanks.

Sebastian
diff mbox series

Patch

diff --git a/drivers/tty/serial/8250/8250_port.c b/drivers/tty/serial/8250/8250_port.c
index 46e2079ad1aa..5805f18520dd 100644
--- a/drivers/tty/serial/8250/8250_port.c
+++ b/drivers/tty/serial/8250/8250_port.c
@@ -2056,10 +2056,7 @@  static void serial8250_break_ctl(struct uart_port *port, int break_state)
 	serial8250_rpm_put(up);
 }
 
-/*
- *	Wait for transmitter & holding register to empty
- */
-static void wait_for_xmitr(struct uart_8250_port *up, int bits)
+static void wait_for_lsr(struct uart_8250_port *up, int bits)
 {
 	unsigned int status, tmout = 10000;
 
@@ -2076,6 +2073,16 @@  static void wait_for_xmitr(struct uart_8250_port *up, int bits)
 		udelay(1);
 		touch_nmi_watchdog();
 	}
+}
+
+/*
+ *	Wait for transmitter & holding register to empty
+ */
+static void wait_for_xmitr(struct uart_8250_port *up, int bits)
+{
+	unsigned int tmout;
+
+	wait_for_lsr(up, bits);
 
 	/* Wait up to 1s for flow control if necessary */
 	if (up->port.flags & UPF_CONS_FLOW) {
@@ -3325,6 +3332,35 @@  static void serial8250_console_restore(struct uart_8250_port *up)
 	serial8250_out_MCR(up, UART_MCR_DTR | UART_MCR_RTS);
 }
 
+/*
+ * Print a string to the serial port using the device FIFO
+ *
+ * It sends fifosize bytes and then waits for the fifo
+ * to get empty.
+ */
+static void serial8250_console_fifo_write(struct uart_8250_port *up,
+					  const char *s, unsigned int count)
+{
+	int i;
+	const char *end = s + count;
+	unsigned int fifosize = up->port.fifosize;
+	bool cr_sent = false;
+
+	while (s != end) {
+		wait_for_lsr(up, UART_LSR_THRE);
+
+		for (i = 0; i < fifosize && s != end; ++i) {
+			if (*s == '\n' && !cr_sent) {
+				serial_out(up, UART_TX, '\r');
+				cr_sent = true;
+			} else {
+				serial_out(up, UART_TX, *s++);
+				cr_sent = false;
+			}
+		}
+	}
+}
+
 /*
  *	Print a string to the serial port trying not to disturb
  *	any possible real use of the port...
@@ -3340,7 +3376,7 @@  void serial8250_console_write(struct uart_8250_port *up, const char *s,
 	struct uart_8250_em485 *em485 = up->em485;
 	struct uart_port *port = &up->port;
 	unsigned long flags;
-	unsigned int ier;
+	unsigned int ier, use_fifo;
 	int locked = 1;
 
 	touch_nmi_watchdog();
@@ -3372,7 +3408,20 @@  void serial8250_console_write(struct uart_8250_port *up, const char *s,
 		mdelay(port->rs485.delay_rts_before_send);
 	}
 
-	uart_console_write(port, s, count, serial8250_console_putchar);
+	use_fifo = (up->capabilities & UART_CAP_FIFO) &&
+		port->fifosize > 1 &&
+		(serial_port_in(port, UART_FCR) & UART_FCR_ENABLE_FIFO) &&
+		/*
+		 * After we put a data in the fifo, the controller will send
+		 * it regardless of the CTS state. Therefore, only use fifo
+		 * if we don't use control flow.
+		 */
+		!(up->port.flags & UPF_CONS_FLOW);
+
+	if (likely(use_fifo))
+		serial8250_console_fifo_write(up, s, count);
+	else
+		uart_console_write(port, s, count, serial8250_console_putchar);
 
 	/*
 	 *	Finally, wait for transmitter to become empty