From: "Pali Rohár" <pali@kernel.org>
To: Stefan Roese <sr@denx.de>
Cc: "Marek Behún" <marek.behun@nic.cz>, u-boot@lists.denx.de
Subject: [PATCH] serial: a37xx: Reset whole UART when changing parent clock from TBG to XTAL
Date: Mon, 15 Nov 2021 13:48:06 +0100 [thread overview]
Message-ID: <20211115124806.16122-1-pali@kernel.org> (raw)
Sometimes UART stops transmitting characters after UART clock is changed
back to XTAL. In this state UART fifo is always full. Kernel during early
boot wants to print output on UART and is waiting for non-empty UART fifo.
Which leads to CPU hangup without any (debug) output on UART.
Marvell Armada 3700 Functional Specifications says that for programming
fractional divisor registers it is required to disable UART, enable
loopback mode, reset fifos, program registers, disable loopback mode,
release reset of fifos and enable UART.
But these steps do not fix above mentioned issue that UART hangup. Also
gating UART clock does not help. And even resetting UART state machines do
not help.
Experiments showed that UART fifo is unblocked after board is being reset
(during board reset UART HW transmit UART fifo even CPU is not executing
kernel/bootloader anymore).
And another experiments showed that same workaround can be achieved also
by external reset of UART HW (without need to reset board).
So do not implement any of "Marvell recommended" steps from Functional
Specifications as they do not work. And rather prior changing parent clock
back to XTAL, do external reset of UART HW. This operation also resets all
UART registers, so basically it also sets UART clock to default, which is
XTAL. It is unknown why UART hangups and enters such broken state.
Signed-off-by: Pali Rohár <pali@kernel.org>
---
drivers/serial/serial_mvebu_a3700.c | 13 ++++++++++++-
1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/drivers/serial/serial_mvebu_a3700.c b/drivers/serial/serial_mvebu_a3700.c
index 6bca8e4b7e2d..93e4d38d34c7 100644
--- a/drivers/serial/serial_mvebu_a3700.c
+++ b/drivers/serial/serial_mvebu_a3700.c
@@ -9,6 +9,7 @@
#include <serial.h>
#include <asm/io.h>
#include <asm/arch/cpu.h>
+#include <mach/soc.h>
struct mvebu_plat {
void __iomem *base;
@@ -213,6 +214,7 @@ static int mvebu_serial_remove(struct udevice *dev)
u32 new_oversampling;
u32 oversampling;
u32 d1, d2;
+ u32 nb_rst;
/*
* Switch UART base clock back to XTAL because older Linux kernel
@@ -260,12 +262,22 @@ static int mvebu_serial_remove(struct udevice *dev)
return 0;
}
+ /* wait until TX empty */
while (!(readl(base + UART_STATUS_REG) & UART_STATUS_TX_EMPTY))
;
+ /* external reset of UART via North Bridge Peripheral */
+ nb_rst = readl(MVEBU_REGISTER(0x12400));
+ writel(nb_rst & ~BIT(3), MVEBU_REGISTER(0x12400));
+ writel(nb_rst | BIT(3), MVEBU_REGISTER(0x12400));
+
+ /* set baudrate and oversampling */
writel(new_divider, base + UART_BAUD_REG);
writel(new_oversampling, base + UART_POSSR_REG);
+ /* No Parity, 1 Stop */
+ writel(0, base + UART_CTRL_REG);
+
return 0;
}
@@ -305,7 +317,6 @@ U_BOOT_DRIVER(serial_mvebu) = {
#ifdef CONFIG_DEBUG_MVEBU_A3700_UART
#include <debug_uart.h>
-#include <mach/soc.h>
static inline void _debug_uart_init(void)
{
--
2.20.1
next reply other threads:[~2021-11-15 12:48 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-15 12:48 Pali Rohár [this message]
2021-11-30 6:36 ` [PATCH] serial: a37xx: Reset whole UART when changing parent clock from TBG to XTAL Stefan Roese
2021-11-30 13:38 ` Stefan Roese
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20211115124806.16122-1-pali@kernel.org \
--to=pali@kernel.org \
--cc=marek.behun@nic.cz \
--cc=sr@denx.de \
--cc=u-boot@lists.denx.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox