Skip to contentSkip to conversation

serial/sysrq: Add MAGIC_SYSRQ_SERIAL_SEQUENCE

PATCH In review Dmitry Safonov 6 participants 22 messages Jan 9, 2020 – Jan 15, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

cover serial/sysrq: Add MAGIC_SYSRQ_SERIAL_SEQUENCE

Magic sysrq has proven for Arista usecases to be useful for debugging issues in field, over serial line when the switch is in such bad state that it can't accept network connections anymore.

Unfortunately, having sysrq always enabled doesn't work for some embedded boards that tend to generate garbage on serial line (including BREAKs). Since commit 732dbf3a6104 ("serial: do not accept sysrq characters via serial port"), it's possible to keep sysrq enabled, but over serial line.

Add a way to enable sysrq on a uart, where currently it can be constantly either on or off (CONFIG_MAGIC_SYSRQ_SERIAL). While doing so, cleanup __sysrq_enabled and serial_core header file.

Sending against -next tree as it's based on removing SUPPORT_SYSRQ ifdeffery [1].

[1]: https://lkml.kernel.org/r/[email protected]

Cc Greg Kroah-Hartman <[email protected]>
Cc Jiri Slaby <[email protected]>
Cc Vasiliy Khoruzhick <[email protected]>

Thanks,
Dmitry

Dmitry Safonov (3):
serial_core: Move sysrq functions from header file sysctl/sysrq: Remove __sysrq_enabled copy serial/sysrq: Add MAGIC_SYSRQ_SERIAL_SEQUENCE

 drivers/tty/serial/serial_core.c | 123 +++++++++++++++++++++++++++++++
 drivers/tty/sysrq.c              |   7 ++
 include/linux/serial_core.h      |  86 ++-------------------
 include/linux/sysrq.h            |   1 +
 kernel/sysctl.c                  |  41 ++++++-----
 lib/Kconfig.debug                |   8 ++
 6 files changed, 167 insertions(+), 99 deletions(-)
Signature · 1 line
2.24.1
#1 lore raw

PATCH 1/3 serial_core: Move sysrq functions from header file

It's not worth to have them in every serial driver and I'm about to add another helper function.

---
 drivers/tty/serial/serial_core.c | 83 +++++++++++++++++++++++++++++++
 include/linux/serial_core.h      | 84 ++------------------------------
 2 files changed, 88 insertions(+), 79 deletions(-)
drivers/tty/serial/serial_core.c +83 -0
index 7c2782785736..6ac9dfed3423 100644
@@ -3081,6 +3081,89 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
+int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch)
+{
+	if (!IS_ENABLED(CONFIG_MAGIC_SYSRQ_SERIAL))
+		return 0;
+
+	if (!port->has_sysrq || !port->sysrq)
+		return 0;
+
+	if (ch && time_before(jiffies, port->sysrq)) {
+		handle_sysrq(ch);
+		port->sysrq = 0;
+		return 1;
+	}
+	port->sysrq = 0;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(uart_handle_sysrq_char);
+
+int uart_prepare_sysrq_char(struct uart_port *port, unsigned int ch)
+{
+	if (!IS_ENABLED(CONFIG_MAGIC_SYSRQ_SERIAL))
+		return 0;
+
+	if (!port->has_sysrq || !port->sysrq)
+		return 0;
+
+	if (ch && time_before(jiffies, port->sysrq)) {
+		port->sysrq_ch = ch;
+		port->sysrq = 0;
+		return 1;
+	}
+	port->sysrq = 0;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(uart_prepare_sysrq_char);
+
+void uart_unlock_and_check_sysrq(struct uart_port *port, unsigned long irqflags)
+{
+	int sysrq_ch;
+
+	if (!port->has_sysrq) {
+		spin_unlock_irqrestore(&port->lock, irqflags);
+		return;
+	}
+
+	sysrq_ch = port->sysrq_ch;
+	port->sysrq_ch = 0;
+
+	spin_unlock_irqrestore(&port->lock, irqflags);
+
+	if (sysrq_ch)
+		handle_sysrq(sysrq_ch);
+}
+EXPORT_SYMBOL_GPL(uart_unlock_and_check_sysrq);
+
+/*
+ * We do the SysRQ and SAK checking like this...
+ */
+int uart_handle_break(struct uart_port *port)
+{
+	struct uart_state *state = port->state;
+
+	if (port->handle_break)
+		port->handle_break(port);
+
+	if (port->has_sysrq) {
+		if (port->cons && port->cons->index == port->line) {
+			if (!port->sysrq) {
+				port->sysrq = jiffies + HZ*5;
+				return 1;
+			}
+			port->sysrq = 0;
+		}
+	}
+
+	if (port->flags & UPF_SAK)
+		do_SAK(state->port.tty);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(uart_handle_break);
+
 EXPORT_SYMBOL(uart_write_wakeup);
 EXPORT_SYMBOL(uart_register_driver);
 EXPORT_SYMBOL(uart_unregister_driver);
include/linux/serial_core.h +5 -79

90 lines not on this page · raw message

Signature · 1 line
2.24.1
#2 lore raw

Re: [PATCH-next 1/3] serial_core: Move sysrq functions from header file

Quoted 8 lines · Dmitry Safonov

It's not worth to have them in every serial driver and I'm about to add another helper function.

---
 drivers/tty/serial/serial_core.c | 83 +++++++++++++++++++++++++++++++
 include/linux/serial_core.h      | 84 ++------------------------------
 2 files changed, 88 insertions(+), 79 deletions(-)

This is a nice cleanup, so I took this patch now, but the last two patches still need a bit of rework. Please do that and rebase, no need to send this patch again.

thanks,

greg k-h

#3 lore raw

PATCH 2/3 sysctl/sysrq: Remove __sysrq_enabled copy

In reply to Dmitry Safonov

Many embedded boards have a disconnected TTL level serial which can generate some garbage that can lead to spurious false sysrq detects.

Currently, sysrq can be either completely disabled for serial console or always disabled (with CONFIG_MAGIC_SYSRQ_SERIAL), since
commit 732dbf3a6104 ("serial: do not accept sysrq characters via serial port")

At Arista, we have such boards that can generate BREAK and random garbage. While disabling sysrq for serial console would solve the problem with spurious false sysrq triggers, it's also desirable to have a way to enable sysrq back.

Having the way to enable sysrq was beneficial to debug lockups with a manual investigation in field and on the other side preventing false sysrq detections.

As a preparation to add sysrq_toggle_support() call into uart, remove a private copy of sysrq_enabled from sysctl - it should reflect the actual status of sysrq.

Furthermore, the private copy isn't correct already in case sysrq_always_enabled is true. So, remove __sysrq_enabled and use a getter-helper for sysrq enabled status.

Cc Iurii Zaikin <[email protected]>
Cc Jiri Slaby <[email protected]>
Cc Luis Chamberlain <[email protected]>
Cc Kees Cook <[email protected]>
---
 drivers/tty/sysrq.c   |  7 +++++++
 include/linux/sysrq.h |  1 +
 kernel/sysctl.c       | 41 ++++++++++++++++++++++-------------------
 3 files changed, 30 insertions(+), 19 deletions(-)
drivers/tty/sysrq.c +7 -0
index f724962a5906..ef3e78967146 100644
@@ -73,6 +73,13 @@ static bool sysrq_on_mask(int mask)
 	       (sysrq_enabled & mask);
 }
 
+int sysrq_get_mask(void)
+{
+	if (sysrq_always_enabled)
+		return 1;
+	return sysrq_enabled;
+}
+
 static int __init sysrq_always_enabled_setup(char *str)
 {
 	sysrq_always_enabled = true;
include/linux/sysrq.h +1 -0
index 8c71874e8485..4a0b351fa2d3 100644
@@ -50,6 +50,7 @@ int unregister_sysrq_key(int key, struct sysrq_key_op *op);
 struct sysrq_key_op *__sysrq_get_key_op(int key);
 
 int sysrq_toggle_support(int enable_mask);
+int sysrq_get_mask(void);
 
 #else
 
kernel/sysctl.c +22 -19

60 lines not on this page · raw message

Signature · 1 line
2.24.1
#4 lore raw

Re: [PATCH-next 2/3] sysctl/sysrq: Remove __sysrq_enabled copy

Quoted 3 lines

Furthermore, the private copy isn't correct already in case sysrq_always_enabled is true. So, remove __sysrq_enabled and use a getter-helper for sysrq enabled status.

Since the old way was known to be unreliable, the new way looks like a great candidate for a KUnit test.

Off topic: I wonder if Magic Sysrq could be extended with reasonable effort to be triggered by a sequence of keystrokes which would be less likely to be generated by a noisy serial.

#5 lore raw

Re: [PATCH-next 2/3] sysctl/sysrq: Remove __sysrq_enabled copy

Hi Iurii,

Quoted 3 lines · Iurii Zaikin

Off topic: I wonder if Magic Sysrq could be extended with reasonable effort to be triggered by a sequence of keystrokes which would be less likely to be generated by a noisy serial.

That's what I do in 3/3 patch from the series:
https://lkml.kernel.org/r/[email protected]

Or do you mean something else?

Thanks,
Dmitry

#6 lore raw

Re: [PATCH-next 2/3] sysctl/sysrq: Remove __sysrq_enabled copy

In reply to Dmitry Safonov
Quoted 119 lines · Dmitry Safonov

Many embedded boards have a disconnected TTL level serial which can generate some garbage that can lead to spurious false sysrq detects.

Currently, sysrq can be either completely disabled for serial console or always disabled (with CONFIG_MAGIC_SYSRQ_SERIAL), since
commit 732dbf3a6104 ("serial: do not accept sysrq characters via serial port")

At Arista, we have such boards that can generate BREAK and random garbage. While disabling sysrq for serial console would solve the problem with spurious false sysrq triggers, it's also desirable to have a way to enable sysrq back.

Having the way to enable sysrq was beneficial to debug lockups with a manual investigation in field and on the other side preventing false sysrq detections.

As a preparation to add sysrq_toggle_support() call into uart, remove a private copy of sysrq_enabled from sysctl - it should reflect the actual status of sysrq.

Furthermore, the private copy isn't correct already in case sysrq_always_enabled is true. So, remove __sysrq_enabled and use a getter-helper for sysrq enabled status.

Cc Iurii Zaikin <[email protected]>
Cc Jiri Slaby <[email protected]>
Cc Luis Chamberlain <[email protected]>
Cc Kees Cook <[email protected]>
---
 drivers/tty/sysrq.c   |  7 +++++++
 include/linux/sysrq.h |  1 +
 kernel/sysctl.c       | 41 ++++++++++++++++++++++-------------------
 3 files changed, 30 insertions(+), 19 deletions(-)
drivers/tty/sysrq.c +7 -0
index f724962a5906..ef3e78967146 100644
@@ -73,6 +73,13 @@ static bool sysrq_on_mask(int mask)
 	       (sysrq_enabled & mask);
 }
 
+int sysrq_get_mask(void)
+{
+	if (sysrq_always_enabled)
+		return 1;
+	return sysrq_enabled;
+}
+
 static int __init sysrq_always_enabled_setup(char *str)
 {
 	sysrq_always_enabled = true;
include/linux/sysrq.h +1 -0
index 8c71874e8485..4a0b351fa2d3 100644
@@ -50,6 +50,7 @@ int unregister_sysrq_key(int key, struct sysrq_key_op *op);
 struct sysrq_key_op *__sysrq_get_key_op(int key);
 
 int sysrq_toggle_support(int enable_mask);
+int sysrq_get_mask(void);
 
 #else
 
kernel/sysctl.c +11 -19
index d396aaaf19a3..6ddb4d7df0e1 100644
@@ -229,25 +229,8 @@ static int proc_dopipe_max_size(struct ctl_table *table, int write,
 		void __user *buffer, size_t *lenp, loff_t *ppos);
 
 #ifdef CONFIG_MAGIC_SYSRQ
-/* Note: sysrq code uses its own private copy */
-static int __sysrq_enabled = CONFIG_MAGIC_SYSRQ_DEFAULT_ENABLE;
-
 static int sysrq_sysctl_handler(struct ctl_table *table, int write,
-				void __user *buffer, size_t *lenp,
-				loff_t *ppos)
-{
-	int error;
-
-	error = proc_dointvec(table, write, buffer, lenp, ppos);
-	if (error)
-		return error;
-
-	if (write)
-		sysrq_toggle_support(__sysrq_enabled);
-
-	return 0;
-}
-
+			void __user *buffer, size_t *lenp, loff_t *ppos);
 #endif
 
 static struct ctl_table kern_table[];
@@ -747,7 +730,7 @@ static struct ctl_table kern_table[] = {
 #ifdef CONFIG_MAGIC_SYSRQ
 	{
 		.procname	= "sysrq",
-		.data		= &__sysrq_enabled,
+		.data		= NULL,
 		.maxlen		= sizeof (int),
 		.mode		= 0644,
 		.proc_handler	= sysrq_sysctl_handler,
@@ -2844,6 +2827,26 @@ static int proc_dostring_coredump(struct ctl_table *table, int write,
 }
 #endif
 
+#ifdef CONFIG_MAGIC_SYSRQ
+static int sysrq_sysctl_handler(struct ctl_table *table, int write,
+				void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+	int tmp, ret;
+
+	tmp = sysrq_get_mask();
+
+	ret = __do_proc_dointvec(&tmp, table, write, buffer,

kernel/sysctl.c static int proc_dostring_coredump(struct ctl_table *table, int write, · line 2848

Quoted 10 lines
kernel/sysctl.c +10 -0
+			       lenp, ppos, NULL, NULL);
+	if (ret || !write)
+		return ret;
+
+	if (write)
+		sysrq_toggle_support(tmp);
+
+	return 0;
+}
+#endif

Why did you move this function down here? Can't it stay where it is and you can just fix the logic there? Now you have two different #ifdef blocks intead of just one :(

thanks,

greg k-h

#7 lore raw

Re: [PATCH-next 2/3] sysctl/sysrq: Remove __sysrq_enabled copy

Hi Greg,

Quoted 1 line · Greg Kroah-Hartman

On Thu, Jan 09, 2020 at 09:54:43PM +0000, Dmitry Safonov wrote:

[..]

Quoted 19 lines
Quoted 19 lines
kernel/sysctl.c +15 -0
@@ -2844,6 +2827,26 @@ static int proc_dostring_coredump(struct ctl_table *table, int write,
 }
 #endif
 
+#ifdef CONFIG_MAGIC_SYSRQ
+static int sysrq_sysctl_handler(struct ctl_table *table, int write,
+				void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+	int tmp, ret;
+
+	tmp = sysrq_get_mask();
+
+	ret = __do_proc_dointvec(&tmp, table, write, buffer,
+			       lenp, ppos, NULL, NULL);
+	if (ret || !write)
+		return ret;
+
+	if (write)
+		sysrq_toggle_support(tmp);
Quoted 8 lines

kernel/sysctl.c static int proc_dostring_coredump(struct ctl_table *table, int write, · line 2848

Quoted 4 lines
kernel/sysctl.c +4 -0
+
+	return 0;
+}
+#endif

Why did you move this function down here? Can't it stay where it is and you can just fix the logic there? Now you have two different #ifdef blocks intead of just one :(

Yeah, well __do_proc_dointvec() made me do it.

sysrq_sysctl_handler() declaration should be before ctl_table array of sysctls, so I couldn't remove the forward-declaration.

So, I could forward-declare __do_proc_dointvec() instead, but looking at the neighborhood, I decided to follow the file-style (there is a couple of forward-declarations before the sysctl array, some under ifdefs).

I admit that the result is imperfect and can put __do_proc_dointvec() definition before instead, no hard feelings.

Thanks,
Dmitry

#8 lore raw

Re: [PATCH-next 2/3] sysctl/sysrq: Remove __sysrq_enabled copy

Quoted 38 lines · Dmitry Safonov

Hi Greg,

Quoted 1 line · Greg Kroah-Hartman

On Thu, Jan 09, 2020 at 09:54:43PM +0000, Dmitry Safonov wrote:

[..]

Quoted 19 lines
Quoted 19 lines
kernel/sysctl.c +15 -0
@@ -2844,6 +2827,26 @@ static int proc_dostring_coredump(struct ctl_table *table, int write,
 }
 #endif
 
+#ifdef CONFIG_MAGIC_SYSRQ
+static int sysrq_sysctl_handler(struct ctl_table *table, int write,
+				void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+	int tmp, ret;
+
+	tmp = sysrq_get_mask();
+
+	ret = __do_proc_dointvec(&tmp, table, write, buffer,
+			       lenp, ppos, NULL, NULL);
+	if (ret || !write)
+		return ret;
+
+	if (write)
+		sysrq_toggle_support(tmp);
Quoted 8 lines

kernel/sysctl.c static int proc_dostring_coredump(struct ctl_table *table, int write, · line 2848

Quoted 4 lines
kernel/sysctl.c +4 -0
+
+	return 0;
+}
+#endif

Why did you move this function down here? Can't it stay where it is and you can just fix the logic there? Now you have two different #ifdef blocks intead of just one :(

Yeah, well __do_proc_dointvec() made me do it.

sysrq_sysctl_handler() declaration should be before ctl_table array of sysctls, so I couldn't remove the forward-declaration.

Quoted 6 lines

So, I could forward-declare __do_proc_dointvec() instead, but looking at the neighborhood, I decided to follow the file-style (there is a couple of forward-declarations before the sysctl array, some under ifdefs).

I admit that the result is imperfect and can put __do_proc_dointvec() definition before instead, no hard feelings.

Ah, no, nevermind, I missed that reason, sorry about that. Moving it is fine.

greg k-h

#9 lore raw

PATCH 3/3 serial/sysrq: Add MAGIC_SYSRQ_SERIAL_SEQUENCE

In reply to Dmitry Safonov

Many embedded boards have a disconnected TTL level serial which can generate some garbage that can lead to spurious false sysrq detects.

Currently, sysrq can be either completely disabled for serial console or always disabled (with CONFIG_MAGIC_SYSRQ_SERIAL), since
commit 732dbf3a6104 ("serial: do not accept sysrq characters via serial port")

At Arista, we have such boards that can generate BREAK and random garbage. While disabling sysrq for serial console would solve the problem with spurious false sysrq triggers, it's also desirable to have a way to enable sysrq back.

As a measure of balance between on and off options, add MAGIC_SYSRQ_SERIAL_SEQUENCE which is a string sequence that can enable sysrq if it follows BREAK on a serial line. The longer the string - the less likely it may be in the garbage.

Having the way to enable sysrq was beneficial to debug lockups with a manual investigation in field and on the other side preventing false sysrq detections.

---
 drivers/tty/serial/serial_core.c | 52 ++++++++++++++++++++++++++++----
 include/linux/serial_core.h      |  2 +-
 lib/Kconfig.debug                |  8 +++++
 3 files changed, 55 insertions(+), 7 deletions(-)
drivers/tty/serial/serial_core.c +46 -6
index 6ac9dfed3423..f70eba032d0b 100644
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
+
+static void uart_sysrq_on(struct work_struct *w)
+{
+	sysrq_toggle_support(1);
+	pr_info("SysRq is enabled by magic sequience on serial\n");
+}
+static DECLARE_WORK(sysrq_enable_work, uart_sysrq_on);
+
+static int uart_try_toggle_sysrq(struct uart_port *port, unsigned int ch)
+{
+	if (sysrq_toggle_seq[0] == '\0')
+		return 0;
+
+	BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >= sizeof(port->sysrq_seq)*U8_MAX);
+	if (sysrq_toggle_seq[port->sysrq_seq] != ch) {
+		port->sysrq_seq = 0;
+		return 0;
+	}
+
+	/* Without the last \0 */
+	if (++port->sysrq_seq < (ARRAY_SIZE(sysrq_toggle_seq) - 1)) {
+		port->sysrq = jiffies + HZ*5;
+		return 1;
+	}
+
+	schedule_work(&sysrq_enable_work);
+
+	port->sysrq = 0;
+	return 1;
+}
+
 int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch)
 {
 	if (!IS_ENABLED(CONFIG_MAGIC_SYSRQ_SERIAL))
@@ -3090,9 +3122,13 @@ int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch)
 		return 0;
 
 	if (ch && time_before(jiffies, port->sysrq)) {
-		handle_sysrq(ch);
-		port->sysrq = 0;
-		return 1;
+		if (sysrq_get_mask()) {
+			handle_sysrq(ch);
+			port->sysrq = 0;
+			return 1;
+		}
+		if (uart_try_toggle_sysrq(port, ch))
+			return 1;
 	}
 	port->sysrq = 0;
 
@@ -3109,9 +3145,13 @@ int uart_prepare_sysrq_char(struct uart_port *port, unsigned int ch)
 		return 0;
 
 	if (ch && time_before(jiffies, port->sysrq)) {
-		port->sysrq_ch = ch;
-		port->sysrq = 0;
-		return 1;
+		if (sysrq_get_mask()) {
+			port->sysrq_ch = ch;
+			port->sysrq = 0;
+			return 1;
+		}
+		if (uart_try_toggle_sysrq(port, ch))
+			return 1;
 	}
 	port->sysrq = 0;
 
include/linux/serial_core.h +1 -1
index 255e86a474e9..1f4443db5474 100644
@@ -243,10 +243,10 @@ struct uart_port {
 	unsigned long		sysrq;			/* sysrq timeout */
 	unsigned int		sysrq_ch;		/* char for sysrq */
 	unsigned char		has_sysrq;
+	unsigned char		sysrq_seq;		/* index in sysrq_toggle_seq */
 
 	unsigned char		hub6;			/* this should be in the 8250 driver */
 	unsigned char		suspended;
-	unsigned char		unused;
 	const char		*name;			/* port name */
 	struct attribute_group	*attr_group;		/* port specific attributes */
 	const struct attribute_group **tty_groups;	/* all attributes (serial core use only) */
lib/Kconfig.debug +8 -0
index 330909c600b1..f6f3e82fee25 100644
@@ -431,6 +431,14 @@ config MAGIC_SYSRQ_SERIAL
 	  This option allows you to decide whether you want to enable the
 	  magic SysRq key.
 
+config MAGIC_SYSRQ_SERIAL_SEQUENCE
+	string "Char sequence that enables magic SysRq over serial"
+	depends on MAGIC_SYSRQ_SERIAL
+	default ""
+	help
+	  Specifies a sequence of characters that can follow BREAK to enable
+	  SysRq on a serial console.
+
 config DEBUG_FS
 	bool "Debug Filesystem"
 	help
Signature · 1 line
2.24.1
#10 lore raw

Hi,

Quoted 14 lines · Dmitry Safonov
---
 drivers/tty/serial/serial_core.c | 52 ++++++++++++++++++++++++++++----
 include/linux/serial_core.h      |  2 +-
 lib/Kconfig.debug                |  8 +++++
 3 files changed, 55 insertions(+), 7 deletions(-)
drivers/tty/serial/serial_core.c +0 -0
index 6ac9dfed3423..f70eba032d0b 100644

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3089

Quoted 9 lines
drivers/tty/serial/serial_core.c +6 -0
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
+
+static void uart_sysrq_on(struct work_struct *w)
+{
+	sysrq_toggle_support(1);
+	pr_info("SysRq is enabled by magic sequience on serial\n");

typo: sequence

Quoted 2 lines
drivers/tty/serial/serial_core.c +2 -0
+}
+static DECLARE_WORK(sysrq_enable_work, uart_sysrq_on);
Signature · 1 line
~Randy
#11 lore raw

Hi Randy,

Quoted 20 lines · Randy Dunlap

Hi,

Quoted 17 lines · Dmitry Safonov
---
 drivers/tty/serial/serial_core.c | 52 ++++++++++++++++++++++++++++----
 include/linux/serial_core.h      |  2 +-
 lib/Kconfig.debug                |  8 +++++
 3 files changed, 55 insertions(+), 7 deletions(-)
drivers/tty/serial/serial_core.c +0 -0
index 6ac9dfed3423..f70eba032d0b 100644
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
Quoted 8 lines

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3089

Quoted 6 lines
drivers/tty/serial/serial_core.c +6 -0
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
+
+static void uart_sysrq_on(struct work_struct *w)
+{
+	sysrq_toggle_support(1);
+	pr_info("SysRq is enabled by magic sequience on serial\n");

typo: sequence

Thanks on catching this,
Dmitry

#12 lore raw
In reply to Dmitry Safonov
Quoted 34 lines · Dmitry Safonov

Many embedded boards have a disconnected TTL level serial which can generate some garbage that can lead to spurious false sysrq detects.

Currently, sysrq can be either completely disabled for serial console or always disabled (with CONFIG_MAGIC_SYSRQ_SERIAL), since
commit 732dbf3a6104 ("serial: do not accept sysrq characters via serial port")

At Arista, we have such boards that can generate BREAK and random garbage. While disabling sysrq for serial console would solve the problem with spurious false sysrq triggers, it's also desirable to have a way to enable sysrq back.

As a measure of balance between on and off options, add MAGIC_SYSRQ_SERIAL_SEQUENCE which is a string sequence that can enable sysrq if it follows BREAK on a serial line. The longer the string - the less likely it may be in the garbage.

Having the way to enable sysrq was beneficial to debug lockups with a manual investigation in field and on the other side preventing false sysrq detections.

---
 drivers/tty/serial/serial_core.c | 52 ++++++++++++++++++++++++++++----
 include/linux/serial_core.h      |  2 +-
 lib/Kconfig.debug                |  8 +++++
 3 files changed, 55 insertions(+), 7 deletions(-)
drivers/tty/serial/serial_core.c +0 -0
index 6ac9dfed3423..f70eba032d0b 100644

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3089

Quoted 9 lines
drivers/tty/serial/serial_core.c +6 -0
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
+
+static void uart_sysrq_on(struct work_struct *w)
+{
+	sysrq_toggle_support(1);
+	pr_info("SysRq is enabled by magic sequience on serial\n");

Do we want to say what serial port it is enabled on?

And why is this done in a workqueue?

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3096

Quoted 7 lines
drivers/tty/serial/serial_core.c +7 -0
+}
+static DECLARE_WORK(sysrq_enable_work, uart_sysrq_on);
+
+static int uart_try_toggle_sysrq(struct uart_port *port, unsigned int ch)
+{
+	if (sysrq_toggle_seq[0] == '\0')
+		return 0;

Is constantly checking the data stream like this going to slow things down overall? Ah, we are just checking this after BREAK, right? So that hopefully will not be that bad...

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3106

Quoted 10 lines
drivers/tty/serial/serial_core.c +10 -0
+
+	BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >= sizeof(port->sysrq_seq)*U8_MAX);
+	if (sysrq_toggle_seq[port->sysrq_seq] != ch) {
+		port->sysrq_seq = 0;
+		return 0;
+	}
+
+	/* Without the last \0 */
+	if (++port->sysrq_seq < (ARRAY_SIZE(sysrq_toggle_seq) - 1)) {
+		port->sysrq = jiffies + HZ*5;

5 second delay? You should document what this value is for somewhere here...

Quoted 14 lines
drivers/tty/serial/serial_core.c +9 -0
+		return 1;
+	}
+
+	schedule_work(&sysrq_enable_work);
+
+	port->sysrq = 0;
+	return 1;
+}
+
 int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch)
 {
 	if (!IS_ENABLED(CONFIG_MAGIC_SYSRQ_SERIAL))
@@ -3090,9 +3122,13 @@ int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch)
 		return 0;

drivers/tty/serial/serial_core.c int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch) · line 3129

Quoted 10 lines
drivers/tty/serial/serial_core.c +5 -3
 
 	if (ch && time_before(jiffies, port->sysrq)) {
-		handle_sysrq(ch);
-		port->sysrq = 0;
-		return 1;
+		if (sysrq_get_mask()) {
+			handle_sysrq(ch);
+			port->sysrq = 0;
+			return 1;
+		}

Isn't this change to test for sysrq_get_mask() a different change than checking for the "magic" data stream?

Quoted 27 lines
drivers/tty/serial/serial_core.c +9 -3
+		if (uart_try_toggle_sysrq(port, ch))
+			return 1;
 	}
 	port->sysrq = 0;
 
@@ -3109,9 +3145,13 @@ int uart_prepare_sysrq_char(struct uart_port *port, unsigned int ch)
 		return 0;
 
 	if (ch && time_before(jiffies, port->sysrq)) {
-		port->sysrq_ch = ch;
-		port->sysrq = 0;
-		return 1;
+		if (sysrq_get_mask()) {
+			port->sysrq_ch = ch;
+			port->sysrq = 0;
+			return 1;
+		}
+		if (uart_try_toggle_sysrq(port, ch))
+			return 1;
 	}
 	port->sysrq = 0;
 
include/linux/serial_core.h +0 -0
index 255e86a474e9..1f4443db5474 100644

include/linux/serial_core.h struct uart_port { · line 249

Quoted 8 lines
include/linux/serial_core.h +1 -1
@@ -243,10 +243,10 @@ struct uart_port {
 	unsigned long		sysrq;			/* sysrq timeout */
 	unsigned int		sysrq_ch;		/* char for sysrq */
 	unsigned char		has_sysrq;
+	unsigned char		sysrq_seq;		/* index in sysrq_toggle_seq */
 
 	unsigned char		hub6;			/* this should be in the 8250 driver */
 	unsigned char		suspended;
-	unsigned char		unused;

This is an unrelated change, let's leave it for a different patch that cleans up the layout of this structure, ok?

thanks,

greg k-h

#13 lore raw

On 1/10/20 4:46 PM, Greg Kroah-Hartman wrote:
[..]

Quoted 8 lines
Quoted 8 lines
drivers/tty/serial/serial_core.c +0 -0
index 6ac9dfed3423..f70eba032d0b 100644
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
Quoted 8 lines

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3089

Quoted 6 lines
drivers/tty/serial/serial_core.c +6 -0
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
+
+static void uart_sysrq_on(struct work_struct *w)
+{
+	sysrq_toggle_support(1);
+	pr_info("SysRq is enabled by magic sequience on serial\n");

Do we want to say what serial port it is enabled on?

Makes sense, will add.

Quoted 1 line

And why is this done in a workqueue?

uart_try_toggle_sysrq() sometimes is called under
spin_lock_irqsave(&port->lock, flags);

And sysrq_toggle_support() calls input_register_handler() internally which can sleep.

Quoted 11 lines

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3096

Quoted 7 lines
drivers/tty/serial/serial_core.c +7 -0
+}
+static DECLARE_WORK(sysrq_enable_work, uart_sysrq_on);
+
+static int uart_try_toggle_sysrq(struct uart_port *port, unsigned int ch)
+{
+	if (sysrq_toggle_seq[0] == '\0')
+		return 0;

Is constantly checking the data stream like this going to slow things down overall? Ah, we are just checking this after BREAK, right? So that hopefully will not be that bad...

Yes, it's after BREAK. In my POV it's fine as originally it would cause sysrq handler being called (if sysrq is enabled).

Quoted 14 lines

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3106

Quoted 11 lines
drivers/tty/serial/serial_core.c +10 -0
+
+	BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >= sizeof(port->sysrq_seq)*U8_MAX);
+	if (sysrq_toggle_seq[port->sysrq_seq] != ch) {
+		port->sysrq_seq = 0;
+		return 0;
+	}
+
+	/* Without the last \0 */
+	if (++port->sysrq_seq < (ARRAY_SIZE(sysrq_toggle_seq) - 1)) {
+		port->sysrq = jiffies + HZ*5;

5 second delay? You should document what this value is for somewhere here...

Fair enough, I'll add
#define SYSRQ_TIMEOUT (HZ*5)

And use it in uart_handle_break() too.

Quoted 15 lines

drivers/tty/serial/serial_core.c int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch) · line 3129

Quoted 12 lines
drivers/tty/serial/serial_core.c +5 -3
@@ -3090,9 +3122,13 @@ int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch)
 		return 0;
 
 	if (ch && time_before(jiffies, port->sysrq)) {
-		handle_sysrq(ch);
-		port->sysrq = 0;
-		return 1;
+		if (sysrq_get_mask()) {
+			handle_sysrq(ch);
+			port->sysrq = 0;
+			return 1;
+		}

Isn't this change to test for sysrq_get_mask() a different change than checking for the "magic" data stream?

It's for the case when sysrq is already enabled.
Than sysrq_get_mask() will return something and it makes uart call handle_sysrq() instead of checking the toggle sequence.

Quoted 10 lines
Quoted 10 lines
include/linux/serial_core.h +1 -0
index 255e86a474e9..1f4443db5474 100644
@@ -243,10 +243,10 @@ struct uart_port {
 	unsigned long		sysrq;			/* sysrq timeout */
 	unsigned int		sysrq_ch;		/* char for sysrq */
 	unsigned char		has_sysrq;
+	unsigned char		sysrq_seq;		/* index in sysrq_toggle_seq */
 
Quoted 6 lines

include/linux/serial_core.h struct uart_port { · line 249

Quoted 3 lines
include/linux/serial_core.h +0 -1
 	unsigned char		hub6;			/* this should be in the 8250 driver */
 	unsigned char		suspended;
-	unsigned char		unused;

This is an unrelated change, let's leave it for a different patch that cleans up the layout of this structure, ok?

Yes, sure.

Thanks,
Dmitry

#14 lore raw
Quoted 19 lines · Dmitry Safonov

On 1/10/20 4:46 PM, Greg Kroah-Hartman wrote:
[..]

Quoted 8 lines
Quoted 8 lines
drivers/tty/serial/serial_core.c +0 -0
index 6ac9dfed3423..f70eba032d0b 100644
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
Quoted 8 lines

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3089

Quoted 6 lines
drivers/tty/serial/serial_core.c +6 -0
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
+
+static void uart_sysrq_on(struct work_struct *w)
+{
+	sysrq_toggle_support(1);
+	pr_info("SysRq is enabled by magic sequience on serial\n");

Do we want to say what serial port it is enabled on?

Quoted 1 line

Makes sense, will add.

Ah, I've managed to forget to mention that I didn't add the port name into the message in v2. I experimented a bit - it's getting a bit complicated how-to protect (char *name) for just this message. Like, SysRq can be theoretically enabled on two serials at the same moment - so some locking is needed to make the printed name sane.

As sysrq_toggle_support() is a global-enable knob for sysrq (also can be switched in /proc/sys/kernel/sysrq) I'm not sure if it's worth to complicate code to print through which serial console SysRq has been enabled.

I can still do it in v3 if you insist.

And sorry about forgetting to mention this - thought I'll write reply after I send v2 and somehow it slipped my mind.

Thanks,
Dmitry

#15 lore raw
Quoted 28 lines · Dmitry Safonov
Quoted 19 lines · Dmitry Safonov

On 1/10/20 4:46 PM, Greg Kroah-Hartman wrote:
[..]

Quoted 8 lines
Quoted 8 lines
drivers/tty/serial/serial_core.c +0 -0
index 6ac9dfed3423..f70eba032d0b 100644
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,
 }
 EXPORT_SYMBOL_GPL(uart_insert_char);
 
Quoted 8 lines

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3089

Quoted 6 lines
drivers/tty/serial/serial_core.c +6 -0
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
+
+static void uart_sysrq_on(struct work_struct *w)
+{
+	sysrq_toggle_support(1);
+	pr_info("SysRq is enabled by magic sequience on serial\n");

Do we want to say what serial port it is enabled on?

Quoted 1 line

Makes sense, will add.

Ah, I've managed to forget to mention that I didn't add the port name into the message in v2. I experimented a bit - it's getting a bit complicated how-to protect (char *name) for just this message. Like, SysRq can be theoretically enabled on two serials at the same moment - so some locking is needed to make the printed name sane.

Quoted 6 lines

As sysrq_toggle_support() is a global-enable knob for sysrq (also can be switched in /proc/sys/kernel/sysrq) I'm not sure if it's worth to complicate code to print through which serial console SysRq has been enabled.

I can still do it in v3 if you insist.

No worries, if it's too hard to get to, that's fine. If it was available, it would have been nice to have.

thanks,

greg k-h

#16 lore raw
In reply to Dmitry Safonov
Quoted 2 lines · Dmitry Safonov

Many embedded boards have a disconnected TTL level serial which can generate some garbage that can lead to spurious false sysrq detects.

trivia:

drivers/tty/serial/serial_core.c

Quoted 1 line
drivers/tty/serial/serial_core.c +0 -0

[]

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status,

Quoted 1 line
drivers/tty/serial/serial_core.c +0 -0
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,

[]

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3084

Quoted 1 line
drivers/tty/serial/serial_core.c +1 -0
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;

static const?

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3093

Quoted 1 line
drivers/tty/serial/serial_core.c +1 -0
+static int uart_try_toggle_sysrq(struct uart_port *port, unsigned int ch)

This function return might read better as bool not int

Quoted 11 lines
drivers/tty/serial/serial_core.c +11 -0
+{
+	if (sysrq_toggle_seq[0] == '\0')
+		return 0;
+
+	BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >= sizeof(port->sysrq_seq)*U8_MAX);
+	if (sysrq_toggle_seq[port->sysrq_seq] != ch) {
+		port->sysrq_seq = 0;
+		return 0;
+	}
+
+	/* Without the last \0 */

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3114

Quoted 10 lines
drivers/tty/serial/serial_core.c +10 -0
+	if (++port->sysrq_seq < (ARRAY_SIZE(sysrq_toggle_seq) - 1)) {
+		port->sysrq = jiffies + HZ*5;
+		return 1;
+	}
+
+	schedule_work(&sysrq_enable_work);
+
+	port->sysrq = 0;
+	return 1;
+}

[]

include/linux/serial_core.h struct uart_port { · line 246

Quoted 9 lines
include/linux/serial_core.h +1 -0
index 255e86a474e9..1f4443db5474 100644
@@ -243,10 +243,10 @@ struct uart_port {
 	unsigned long		sysrq;			/* sysrq timeout */
 	unsigned int		sysrq_ch;		/* char for sysrq */
 	unsigned char		has_sysrq;
+	unsigned char		sysrq_seq;		/* index in sysrq_toggle_seq */

unsigned int?

Or maybe set a maximum length of MAGIC_SYSRQ_SERIAL_SEQUENCE.

Quoted 26 lines
include/linux/serial_core.h +0 -1
 	unsigned char		hub6;			/* this should be in the 8250 driver */
 	unsigned char		suspended;
-	unsigned char		unused;
 	const char		*name;			/* port name */
 	struct attribute_group	*attr_group;		/* port specific attributes */
 	const struct attribute_group **tty_groups;	/* all attributes (serial core use only) */
lib/Kconfig.debug +8 -0
index 330909c600b1..f6f3e82fee25 100644
@@ -431,6 +431,14 @@ config MAGIC_SYSRQ_SERIAL
 	  This option allows you to decide whether you want to enable the
 	  magic SysRq key.
 
+config MAGIC_SYSRQ_SERIAL_SEQUENCE
+	string "Char sequence that enables magic SysRq over serial"
+	depends on MAGIC_SYSRQ_SERIAL
+	default ""
+	help
+	  Specifies a sequence of characters that can follow BREAK to enable
+	  SysRq on a serial console.
+
 config DEBUG_FS
 	bool "Debug Filesystem"
 	help
#17 lore raw

Hi Joe,

Quoted 6 lines · Joe Perches
Quoted 2 lines · Dmitry Safonov

Many embedded boards have a disconnected TTL level serial which can generate some garbage that can lead to spurious false sysrq detects.

trivia:

Quoted 7 lines

drivers/tty/serial/serial_core.c

Quoted 1 line
drivers/tty/serial/serial_core.c +0 -0

[]

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status,

Quoted 1 line
drivers/tty/serial/serial_core.c +0 -0
@@ -3081,6 +3081,38 @@ void uart_insert_char(struct uart_port *port, unsigned int status,

[]

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3084

Quoted 1 line
drivers/tty/serial/serial_core.c +1 -0
+const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;

static const?

Will do, thanks!

Quoted 4 lines

drivers/tty/serial/serial_core.c void uart_insert_char(struct uart_port *port, unsigned int status, · line 3093

Quoted 2 lines
drivers/tty/serial/serial_core.c +1 -0
+static int uart_try_toggle_sysrq(struct uart_port *port, unsigned int ch)

This function return might read better as bool not int

Yeah, no hard feelings, will convert.

[..]

Quoted 9 lines

drivers/tty/serial/serial_core.c struct uart_port { · line 246

Quoted 5 lines
drivers/tty/serial/serial_core.c +1 -0
@@ -243,10 +243,10 @@ struct uart_port {
 	unsigned long		sysrq;			/* sysrq timeout */
 	unsigned int		sysrq_ch;		/* char for sysrq */
 	unsigned char		has_sysrq;
+	unsigned char		sysrq_seq;		/* index in sysrq_toggle_seq */

unsigned int?

Or maybe set a maximum length of MAGIC_SYSRQ_SERIAL_SEQUENCE.

I think, 256 chars should be enough to send on serial (c)

I'm not aware of a way to put the max string length in Kconfig, so I did in the patch:
BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >= sizeof(port->sysrq_seq)*U8_MAX);

Do you have something more elegant in your mind?

Thanks,
Dmitry

#18 lore raw
Quoted 1 line · Dmitry Safonov

Hi Joe,

Hi Dmitry.

Quoted 1 line

I'm not aware of a way to put the max string length in Kconfig,

Nor am I.

Quoted 6 lines

so I did
in the patch:
BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >=
sizeof(port->sysrq_seq)*U8_MAX);

Do you have something more elegant in your mind?

No, I didn't notice this and think this is fine
so thanks for that.

It's be nice to have something like BUILD_BUG_ON_MSG.

#19 lore raw
In reply to Dmitry Safonov
Thank you for the patch! Yet something to improve:

Hi Dmitry,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on tty/tty-testing]
[also build test ERROR on next-20200109]
[cannot apply to linux/master usb/usb-testing linus/master v5.5-rc5] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system. BTW, we also suggest to use '--base' option to specify the base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/Dmitry-Safonov/serial-sysrq-Add-MAGIC_SYSRQ_SERIAL_SEQUENCE/20200110-191606
base:   https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git tty-testing
config: microblaze-nommu_defconfig (attached as .config)
compiler: microblaze-linux-gcc (GCC) 7.5.0

reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
GCC_VERSION=7.5.0 make.cross ARCH=microblaze

If you fix the issue, kindly add following tag

All errors (new ones prefixed by >>):

>> drivers/tty/serial/serial_core.c:3084:33: error: 'CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE' undeclared here (not in a function)
    const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
                                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   drivers/tty/serial/serial_core.c: In function 'uart_sysrq_on':
>> drivers/tty/serial/serial_core.c:3088:2: error: implicit declaration of function 'sysrq_toggle_support' [-Werror=implicit-function-declaration]
     sysrq_toggle_support(1);
     ^~~~~~~~~~~~~~~~~~~~
   drivers/tty/serial/serial_core.c: In function 'uart_handle_sysrq_char':
>> drivers/tty/serial/serial_core.c:3125:7: error: implicit declaration of function 'sysrq_get_mask'; did you mean 'xas_get_mark'? [-Werror=implicit-function-declaration]
      if (sysrq_get_mask()) {
          ^~~~~~~~~~~~~~
          xas_get_mark

cc1: some warnings being treated as errors

vim +/CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE +3084 drivers/tty/serial/serial_core.c

3083

> 3084	const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
  3085	
  3086	static void uart_sysrq_on(struct work_struct *w)
  3087	{
> 3088		sysrq_toggle_support(1);
  3089		pr_info("SysRq is enabled by magic sequience on serial\n");
  3090	}
  3091	static DECLARE_WORK(sysrq_enable_work, uart_sysrq_on);
  3092	
  3093	static int uart_try_toggle_sysrq(struct uart_port *port, unsigned int ch)
  3094	{
  3095		if (sysrq_toggle_seq[0] == '\0')
  3096			return 0;
  3097	
  3098		BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >= sizeof(port->sysrq_seq)*U8_MAX);
  3099		if (sysrq_toggle_seq[port->sysrq_seq] != ch) {
  3100			port->sysrq_seq = 0;
  3101			return 0;
  3102		}
  3103	
  3104		/* Without the last \0 */
  3105		if (++port->sysrq_seq < (ARRAY_SIZE(sysrq_toggle_seq) - 1)) {
  3106			port->sysrq = jiffies + HZ*5;
  3107			return 1;
  3108		}
  3109	
  3110		schedule_work(&sysrq_enable_work);
  3111	
  3112		port->sysrq = 0;
  3113		return 1;
  3114	}
  3115	
  3116	int uart_handle_sysrq_char(struct uart_port *port, unsigned int ch)
  3117	{
  3118		if (!IS_ENABLED(CONFIG_MAGIC_SYSRQ_SERIAL))
  3119			return 0;
  3120	
  3121		if (!port->has_sysrq || !port->sysrq)
  3122			return 0;
  3123	
  3124		if (ch && time_before(jiffies, port->sysrq)) {
> 3125			if (sysrq_get_mask()) {

3126 handle_sysrq(ch);
3127 port->sysrq = 0;
3128 return 1;
3129 }
3130 if (uart_try_toggle_sysrq(port, ch))
3131 return 1;
3132 }
3133 port->sysrq = 0;
3134
3135 return 0;
3136 }
3137 EXPORT_SYMBOL_GPL(uart_handle_sysrq_char);
3138

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org Intel Corporation

#20 lore raw
In reply to Dmitry Safonov
Thank you for the patch! Yet something to improve:

Hi Dmitry,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on tty/tty-testing]
[also build test ERROR on next-20200110]
[cannot apply to linux/master usb/usb-testing linus/master v5.5-rc5] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system. BTW, we also suggest to use '--base' option to specify the base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/Dmitry-Safonov/serial-sysrq-Add-MAGIC_SYSRQ_SERIAL_SEQUENCE/20200110-191606
base:   https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git tty-testing
config: mips-randconfig-a001-20200109 (attached as .config)
compiler: mips-linux-gcc (GCC) 7.5.0

reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
GCC_VERSION=7.5.0 make.cross ARCH=mips

If you fix the issue, kindly add following tag

All errors (new ones prefixed by >>):

>> drivers/tty/serial/serial_core.c:3084:33: error: 'CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE' undeclared here (not in a function); did you mean 'CONFIG_MAGIC_SYSRQ_DEFAULT_ENABLE'?
    const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;
                                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                                    CONFIG_MAGIC_SYSRQ_DEFAULT_ENABLE

vim +3084 drivers/tty/serial/serial_core.c

3083

> 3084	const char sysrq_toggle_seq[] = CONFIG_MAGIC_SYSRQ_SERIAL_SEQUENCE;

3085

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org Intel Corporation

#21 lore raw
In reply to Dmitry Safonov
Thank you for the patch! Perhaps something to improve:

Hi Dmitry,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on tty/tty-testing]
[also build test WARNING on next-20200110]
[cannot apply to linux/master usb/usb-testing linus/master v5.5-rc5] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system. BTW, we also suggest to use '--base' option to specify the base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/Dmitry-Safonov/serial-sysrq-Add-MAGIC_SYSRQ_SERIAL_SEQUENCE/20200110-191606
base:   https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git tty-testing

If you fix the issue, kindly add following tag

New smatch warnings:
drivers/tty/serial/serial_core.c:3105 uart_try_toggle_sysrq() warn: unsigned '++port->sysrq_seq' is never less than zero.

Old smatch warnings:
drivers/tty/serial/serial_core.c:295 uart_shutdown() error: we previously assumed 'uport' could be null (see line 291) drivers/tty/serial/serial_core.c:2729 uart_get_attr_iomem_base() warn: argument 4 to %lX specifier is cast from pointer

vim +3105 drivers/tty/serial/serial_core.c

3092
3093 static int uart_try_toggle_sysrq(struct uart_port *port, unsigned int ch)
3094 {
3095 if (sysrq_toggle_seq[0] == '\0')
3096 return 0;
3097
3098 BUILD_BUG_ON(ARRAY_SIZE(sysrq_toggle_seq) >= sizeof(port->sysrq_seq)*U8_MAX);
3099 if (sysrq_toggle_seq[port->sysrq_seq] != ch) {
3100 port->sysrq_seq = 0;
3101 return 0;
3102 }
3103
3104 /* Without the last \0 */

> 3105		if (++port->sysrq_seq < (ARRAY_SIZE(sysrq_toggle_seq) - 1)) {

3106 port->sysrq = jiffies + HZ*5;
3107 return 1;
3108 }
3109
3110 schedule_work(&sysrq_enable_work);
3111
3112 port->sysrq = 0;
3113 return 1;
3114 }
3115

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org Intel Corporation