[PATCH v3 0/4] Handle NBCON consoles on KDB

[PATCH v3 0/4] Handle NBCON consoles on KDB

[Marcos Paulo de Souza]

These changes reached v3 (since the first version was tagged as v2 by me...),
thanks to all reviewers for the suggestions and questions about the code.

A new patch was introduced this time (3/4), adding an exception to nbcon
code when trying to acquire the context of a console when KDB is running,
suggested by John and Petr. Thanks a lot!

Testing
-------

I did the tests using qemu and reapplying commit f79b163c4231
('Revert "serial: 8250: Switch to nbcon console"') created originally by
John, just to exercise the common 8250 serial from qemu. The commit can
be checked on [1]. I had to solve some conflicts since the code has been
reworked after the commit was reverted.

Without the patches, I can't see any mirrored messages on the NBCON
console when KDB is triggered. With the patches I can see the messages.

[1]: https://github.com/marcosps/linux/commit/618bd49f8533db85d9c322f9ad1cb0da22aca9ee

Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
---
Changes in v3:
- Only call nbcon_context_release if nbcon_context_exit_unsafe returns true (John Ogness)
- Dropped the prototype of console_is_usable from kernel/printk/internal. (Petr Mladek)
- Add comments to the new functions introduced (Petr Mladek)
- Flush KDB console on nbcon_kdb_release (Petr Mladek)
- Add an exception for KDB on nbcon_context_try_acquire_direct (John Ogness and Petr Mladek)
- Link to v2: https://lore.kernel.org/r/20250811-nbcon-kgdboc-v2-0-c7c72bcdeaf6@suse.com

Changes in v2:
- Set by mistake ..
- Link to v1: https://lore.kernel.org/r/20250713-nbcon-kgdboc-v1-0-51eccd9247a8@suse.com

---
Marcos Paulo de Souza (4):
      printk: nbcon: Export console_is_usable
      printk: nbcon: Introduce KDB helpers
      printk: nbcon: Allow KDB to acquire the NBCON context
      kdb: Adapt kdb_msg_write to work with NBCON consoles

 include/linux/console.h   | 50 +++++++++++++++++++++++++++++++++
 kernel/debug/kdb/kdb_io.c | 46 +++++++++++++++++++++----------
 kernel/printk/internal.h  | 44 -----------------------------
 kernel/printk/nbcon.c     | 70 +++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 151 insertions(+), 59 deletions(-)
---
base-commit: 618bd49f8533db85d9c322f9ad1cb0da22aca9ee
change-id: 20250713-nbcon-kgdboc-efcfc37fde46

Best regards,
-- 
Marcos Paulo de Souza <mpdesouza@suse.com>

[PATCH v3 1/4] printk: nbcon: Export console_is_usable

[Marcos Paulo de Souza]

The helper will be used on KDB code in the next commits.

Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
---
 include/linux/console.h  | 44 ++++++++++++++++++++++++++++++++++++++++++++
 kernel/printk/internal.h | 44 --------------------------------------------
 2 files changed, 44 insertions(+), 44 deletions(-)

diff --git a/include/linux/console.h b/include/linux/console.h
index 8f10d0a85bb4536e4b0dda4e8ccbdf87978bbb4a..67af483574727c00eea1d5a1eacc994755c92607 100644
--- a/include/linux/console.h
+++ b/include/linux/console.h
@@ -605,6 +605,48 @@ extern bool nbcon_can_proceed(struct nbcon_write_context *wctxt);
 extern bool nbcon_enter_unsafe(struct nbcon_write_context *wctxt);
 extern bool nbcon_exit_unsafe(struct nbcon_write_context *wctxt);
 extern void nbcon_reacquire_nobuf(struct nbcon_write_context *wctxt);
+
+/*
+ * Check if the given console is currently capable and allowed to print
+ * records. Note that this function does not consider the current context,
+ * which can also play a role in deciding if @con can be used to print
+ * records.
+ */
+static inline bool console_is_usable(struct console *con, short flags, bool use_atomic)
+{
+	if (!(flags & CON_ENABLED))
+		return false;
+
+	if ((flags & CON_SUSPENDED))
+		return false;
+
+	if (flags & CON_NBCON) {
+		/* The write_atomic() callback is optional. */
+		if (use_atomic && !con->write_atomic)
+			return false;
+
+		/*
+		 * For the !use_atomic case, @printk_kthreads_running is not
+		 * checked because the write_thread() callback is also used
+		 * via the legacy loop when the printer threads are not
+		 * available.
+		 */
+	} else {
+		if (!con->write)
+			return false;
+	}
+
+	/*
+	 * Console drivers may assume that per-cpu resources have been
+	 * allocated. So unless they're explicitly marked as being able to
+	 * cope (CON_ANYTIME) don't call them until this CPU is officially up.
+	 */
+	if (!cpu_online(raw_smp_processor_id()) && !(flags & CON_ANYTIME))
+		return false;
+
+	return true;
+}
+
 #else
 static inline void nbcon_cpu_emergency_enter(void) { }
 static inline void nbcon_cpu_emergency_exit(void) { }
@@ -612,6 +654,8 @@ static inline bool nbcon_can_proceed(struct nbcon_write_context *wctxt) { return
 static inline bool nbcon_enter_unsafe(struct nbcon_write_context *wctxt) { return false; }
 static inline bool nbcon_exit_unsafe(struct nbcon_write_context *wctxt) { return false; }
 static inline void nbcon_reacquire_nobuf(struct nbcon_write_context *wctxt) { }
+static inline bool console_is_usable(struct console *con, short flags,
+				     bool use_atomic) { return false; }
 #endif
 
 extern int console_set_on_cmdline;
diff --git a/kernel/printk/internal.h b/kernel/printk/internal.h
index ef282001f200fdbbacae3171932bf9f049037a85..bef97f2d11793191280bfb46f7f8b13bf2560351 100644
--- a/kernel/printk/internal.h
+++ b/kernel/printk/internal.h
@@ -112,47 +112,6 @@ bool nbcon_kthread_create(struct console *con);
 void nbcon_kthread_stop(struct console *con);
 void nbcon_kthreads_wake(void);
 
-/*
- * Check if the given console is currently capable and allowed to print
- * records. Note that this function does not consider the current context,
- * which can also play a role in deciding if @con can be used to print
- * records.
- */
-static inline bool console_is_usable(struct console *con, short flags, bool use_atomic)
-{
-	if (!(flags & CON_ENABLED))
-		return false;
-
-	if ((flags & CON_SUSPENDED))
-		return false;
-
-	if (flags & CON_NBCON) {
-		/* The write_atomic() callback is optional. */
-		if (use_atomic && !con->write_atomic)
-			return false;
-
-		/*
-		 * For the !use_atomic case, @printk_kthreads_running is not
-		 * checked because the write_thread() callback is also used
-		 * via the legacy loop when the printer threads are not
-		 * available.
-		 */
-	} else {
-		if (!con->write)
-			return false;
-	}
-
-	/*
-	 * Console drivers may assume that per-cpu resources have been
-	 * allocated. So unless they're explicitly marked as being able to
-	 * cope (CON_ANYTIME) don't call them until this CPU is officially up.
-	 */
-	if (!cpu_online(raw_smp_processor_id()) && !(flags & CON_ANYTIME))
-		return false;
-
-	return true;
-}
-
 /**
  * nbcon_kthread_wake - Wake up a console printing thread
  * @con:	Console to operate on
@@ -204,9 +163,6 @@ static inline bool nbcon_legacy_emit_next_record(struct console *con, bool *hand
 static inline void nbcon_kthread_wake(struct console *con) { }
 static inline void nbcon_kthreads_wake(void) { }
 
-static inline bool console_is_usable(struct console *con, short flags,
-				     bool use_atomic) { return false; }
-
 #endif /* CONFIG_PRINTK */
 
 extern bool have_boot_console;

-- 
2.50.0

[PATCH v3 2/4] printk: nbcon: Introduce KDB helpers

[Marcos Paulo de Souza]

These helpers will be used when calling console->write_atomic on
KDB code in the next patch. It's basically the same implementaion
as nbcon_device_try_acquire, but using NBCON_PORIO_EMERGENCY when
acquiring the context.

For release we need to flush the console, since some messages could be
added before the context was acquired, as KDB emits the messages using
con->{write,write_atomic} instead of storing them on the ring buffer.

Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
---
 include/linux/console.h |  6 +++++
 kernel/printk/nbcon.c   | 66 +++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 72 insertions(+)

diff --git a/include/linux/console.h b/include/linux/console.h
index 67af483574727c00eea1d5a1eacc994755c92607..b34c5a0b86303e2fb4583fa467d8be43761cf756 100644
--- a/include/linux/console.h
+++ b/include/linux/console.h
@@ -605,6 +605,9 @@ extern bool nbcon_can_proceed(struct nbcon_write_context *wctxt);
 extern bool nbcon_enter_unsafe(struct nbcon_write_context *wctxt);
 extern bool nbcon_exit_unsafe(struct nbcon_write_context *wctxt);
 extern void nbcon_reacquire_nobuf(struct nbcon_write_context *wctxt);
+extern bool nbcon_kdb_try_acquire(struct console *con,
+				  struct nbcon_write_context *wctxt);
+extern void nbcon_kdb_release(struct nbcon_write_context *wctxt);
 
 /*
  * Check if the given console is currently capable and allowed to print
@@ -654,6 +657,9 @@ static inline bool nbcon_can_proceed(struct nbcon_write_context *wctxt) { return
 static inline bool nbcon_enter_unsafe(struct nbcon_write_context *wctxt) { return false; }
 static inline bool nbcon_exit_unsafe(struct nbcon_write_context *wctxt) { return false; }
 static inline void nbcon_reacquire_nobuf(struct nbcon_write_context *wctxt) { }
+static inline bool nbcon_kdb_try_acquire(struct console *con,
+					 struct nbcon_write_context *wctxt) { return false; }
+static inline void nbcon_kdb_release(struct console *con) { }
 static inline bool console_is_usable(struct console *con, short flags,
 				     bool use_atomic) { return false; }
 #endif
diff --git a/kernel/printk/nbcon.c b/kernel/printk/nbcon.c
index 646801813415f0abe40cabf2f28ca9e30664f028..ff218e95a505fd10521c2c4dfb00ad5ec5773953 100644
--- a/kernel/printk/nbcon.c
+++ b/kernel/printk/nbcon.c
@@ -1855,3 +1855,69 @@ void nbcon_device_release(struct console *con)
 	console_srcu_read_unlock(cookie);
 }
 EXPORT_SYMBOL_GPL(nbcon_device_release);
+
+/**
+ * nbcon_kdb_try_acquire - Try to acquire nbcon console, enter unsafe
+ *				section, and initialized nbcon write context
+ * @con:	The nbcon console to acquire
+ * @wctxt:	The nbcon write context to be used on success
+ *
+ * Context:	Under console_srcu_read_lock() for emiting a single kdb message
+ *		using the given con->write_atomic() callback. Can be called
+ *		only when the console is usable at the moment.
+ *
+ * Return:	True if the console was acquired. False otherwise.
+ *
+ * kdb emits messages on consoles registered for printk() without
+ * storing them into the ring buffer. It has to acquire the console
+ * ownerhip so that it could call con->write_atomic() callback a safe way.
+ *
+ * This function acquires the nbcon console using priority NBCON_PRIO_EMERGENCY
+ * and marks it unsafe for handover/takeover.
+ */
+bool nbcon_kdb_try_acquire(struct console *con,
+			   struct nbcon_write_context *wctxt)
+{
+	struct nbcon_context *ctxt = &ACCESS_PRIVATE(wctxt, ctxt);
+
+	memset(ctxt, 0, sizeof(*ctxt));
+	ctxt->console = con;
+	ctxt->prio    = NBCON_PRIO_EMERGENCY;
+
+	if (!nbcon_context_try_acquire(ctxt, false))
+		return false;
+
+	if (!nbcon_context_enter_unsafe(ctxt))
+		return false;
+
+	return true;
+}
+
+/**
+ * nbcon_kdb_release - Exit unsafe section and release the nbcon console
+ *
+ * @wctxt:	The nbcon write context initialized by a successful
+ *		nbcon_kdb_try_acquire()
+ *
+ * Context:	Under console_srcu_read_lock() for emiting a single kdb message
+ *		using the given con->write_atomic() callback. Can be called
+ *		only when the console is usable at the moment.
+ */
+void nbcon_kdb_release(struct nbcon_write_context *wctxt)
+{
+	struct nbcon_context *ctxt = &ACCESS_PRIVATE(wctxt, ctxt);
+
+	if (!nbcon_context_exit_unsafe(ctxt))
+		return;
+
+	nbcon_context_release(ctxt);
+
+	/*
+	 * Flush any new printk() messages added when the console was blocked.
+	 * Only the console used by the given write context was	blocked.
+	 * The console was locked only when the write_atomic() callback
+	 * was usable.
+	 */
+	__nbcon_atomic_flush_pending_con(ctxt->console,
+					 prb_next_reserve_seq(prb), false);
+}

-- 
2.50.0

[PATCH v3 3/4] printk: nbcon: Allow KDB to acquire the NBCON context

[Marcos Paulo de Souza]

KDB can interrupt any console to execute the "mirrored printing" at any
time, so add an exception to nbcon_context_try_acquire_direct to allow
to get the context if the current CPU is the same as kdb_printf_cpu.

This change will be necessary for the next patch, which fixes
kdb_msg_write to work with NBCON consoles by calling ->write_atomic on
such consoles. But to print it first needs to acquire the ownership of
the console, so nbcon_context_try_acquire_direct is fixed here.

Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
---
 kernel/printk/nbcon.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/kernel/printk/nbcon.c b/kernel/printk/nbcon.c
index ff218e95a505fd10521c2c4dfb00ad5ec5773953..352235a0eb4a484caccf86d3a57d1a149218ecec 100644
--- a/kernel/printk/nbcon.c
+++ b/kernel/printk/nbcon.c
@@ -10,6 +10,7 @@
 #include <linux/export.h>
 #include <linux/init.h>
 #include <linux/irqflags.h>
+#include <linux/kdb.h>
 #include <linux/kthread.h>
 #include <linux/minmax.h>
 #include <linux/percpu.h>
@@ -247,6 +248,8 @@ static int nbcon_context_try_acquire_direct(struct nbcon_context *ctxt,
 		 * Panic does not imply that the console is owned. However,
 		 * since all non-panic CPUs are stopped during panic(), it
 		 * is safer to have them avoid gaining console ownership.
+		 * The only exception is if kdb is active, which may print
+		 * from multiple CPUs during a panic.
 		 *
 		 * If this acquire is a reacquire (and an unsafe takeover
 		 * has not previously occurred) then it is allowed to attempt
@@ -255,6 +258,7 @@ static int nbcon_context_try_acquire_direct(struct nbcon_context *ctxt,
 		 * interrupted by the panic CPU while printing.
 		 */
 		if (other_cpu_in_panic() &&
+		    READ_ONCE(kdb_printf_cpu) != raw_smp_processor_id() &&
 		    (!is_reacquire || cur->unsafe_takeover)) {
 			return -EPERM;
 		}

-- 
2.50.0

[PATCH v3 4/4] kdb: Adapt kdb_msg_write to work with NBCON consoles

[Marcos Paulo de Souza]

Function kdb_msg_write was calling con->write for any found console,
but it won't work on NBCON ones. In this case we should acquire the
ownership of the console using NBCON_PRIO_EMERGENCY, since printing
kdb messages should only be interrupted by a panic in the case it was
triggered by sysrq debug option. This is done by the
nbcon_kdb_{acquire,release} functions.

At this point, the console is required to use the atomic callback. The
console is skipped if the write_atomic callback is not set or if the
context could not be acquired. The validation of NBCON is done by the
console_is_usable helper. The context is released right after
write_atomic finishes.

The oops_in_progress handling is only needed in the legacy consoles,
so it was moved around the con->write callback.

Suggested-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
---
 kernel/debug/kdb/kdb_io.c | 46 +++++++++++++++++++++++++++++++---------------
 1 file changed, 31 insertions(+), 15 deletions(-)

diff --git a/kernel/debug/kdb/kdb_io.c b/kernel/debug/kdb/kdb_io.c
index 9b11b10b120cf07e451a7a4d92ce50f9a6c066b2..47bc31cc71bc84750db5d9304ed75a113cd382bf 100644
--- a/kernel/debug/kdb/kdb_io.c
+++ b/kernel/debug/kdb/kdb_io.c
@@ -589,24 +589,40 @@ static void kdb_msg_write(const char *msg, int msg_len)
 	 */
 	cookie = console_srcu_read_lock();
 	for_each_console_srcu(c) {
-		if (!(console_srcu_read_flags(c) & CON_ENABLED))
+		struct nbcon_write_context wctxt = { };
+		short flags = console_srcu_read_flags(c);
+
+		if (!console_is_usable(c, flags, true))
 			continue;
 		if (c == dbg_io_ops->cons)
 			continue;
-		if (!c->write)
-			continue;
-		/*
-		 * Set oops_in_progress to encourage the console drivers to
-		 * disregard their internal spin locks: in the current calling
-		 * context the risk of deadlock is a bigger problem than risks
-		 * due to re-entering the console driver. We operate directly on
-		 * oops_in_progress rather than using bust_spinlocks() because
-		 * the calls bust_spinlocks() makes on exit are not appropriate
-		 * for this calling context.
-		 */
-		++oops_in_progress;
-		c->write(c, msg, msg_len);
-		--oops_in_progress;
+
+		if (flags & CON_NBCON) {
+			/*
+			 * Do not continue if the console is NBCON and the context
+			 * can't be acquired.
+			 */
+			if (!nbcon_kdb_try_acquire(c, &wctxt))
+				continue;
+
+			wctxt.outbuf = (char *)msg;
+			wctxt.len = msg_len;
+			c->write_atomic(c, &wctxt);
+			nbcon_kdb_release(&wctxt);
+		} else {
+			/*
+			 * Set oops_in_progress to encourage the console drivers to
+			 * disregard their internal spin locks: in the current calling
+			 * context the risk of deadlock is a bigger problem than risks
+			 * due to re-entering the console driver. We operate directly on
+			 * oops_in_progress rather than using bust_spinlocks() because
+			 * the calls bust_spinlocks() makes on exit are not appropriate
+			 * for this calling context.
+			 */
+			++oops_in_progress;
+			c->write(c, msg, msg_len);
+			--oops_in_progress;
+		}
 		touch_nmi_watchdog();
 	}
 	console_srcu_read_unlock(cookie);

-- 
2.50.0