[PATCH v3 00/10] MIPS: DEC: Fix serial device regressions + RTC cleanup

[PATCH v3 00/10] MIPS: DEC: Fix serial device regressions + RTC cleanup

[Maciej W. Rozycki]

Hi,

 This is v3 of the series, addressing a compilation warning discovered in 
the course of unrelated experimentation in 02/10 where a 64-bit->32-bit 
cast was directly made from a pointer to integer and missed owing to 
`-Wno-error' needed due to an unrelated issue, a fix for which has been 
posted already.  No change to code produced.

 In v2 fallout for 64-bit configurations has been fixed where the initial 
console output handler that uses calls into 32-bit firmware has to ensure 
no register input including the stack pointer is outside the 32-bit range, 
now that the handler is now also called from kernel threads other beyond 
the initial one.

 Two extra introductory changes, 01/10 and 02/10 have now been added to 
platform code.  No modification has been made to original changes.  The 
original description follows, updated for patch renumbering.

 Starting from commit 84a9582fd203 ("serial: core: Start managing serial 
controllers to enable runtime PM") drivers for serial devices used with 
the DEC platform have stopped working due to a null pointer dereference in 
the serial core, which means a kernel crash at bootstrap if the relevant 
driver has been enabled, as is usually the case for the system console.

 This patch series addresses the issue by switching the drivers away from 
legacy probing to using platform devices.  A notable consequence of this 
is the serial console is only switched from the bootconsole handler that 
uses firmware calls over to our serial driver at the time the driver is 
being properly brought up.  This causes messages to be produced to the 
console between the device reset and console setup, which in turn causes 
the firmware still being called via the bootconsole handler to loop 
forever owing to the transmitter having been disabled.

 Both drivers are affected and therefore introductory changes 04/10 and 
05/10 are included to fix the issue by doing a rudimentary device setup 
right after reset, using parameters compatible with those used by the 
firmware (9600n8).  There are auxiliary changes 03/10 and 06/10 included 
as well, that respectively prevent a message corruption regression from 
happening at reset due to the change in timing of messages produced to the 
console with the dz driver switch to the platform device, and simplify 
reset handling in the zs driver by issuing a channel rather hardware 
reset.  Then 07/10 and 08/10 actually switch the respective drivers to use 
platform devices.

 A tiny update follows, 09/10, that enables modular build for the dz 
driver, not to be backported as not a bug fix.

 Finally 10/10 is a small cleanup for the existing RTC device, included in 
the series and then last only, due to a mechanical dependency and neither 
for backporting, as it only addresses a code quality issue for a failure 
scenario that is not expected to trigger in reality.

 See individual change descriptions for details.  Verified with a 5000/200 
machine for the dz driver, and with 5000/150 and 5000/260 systems for the 
zs driver.

 Please apply.

 Previous iterations:

- v2 at: <https://lore.kernel.org/r/alpine.DEB.2.21.2604302336260.38805@angie.orcam.me.uk/>,

- v1 at: <https://lore.kernel.org/r/alpine.DEB.2.21.2604102250060.29980@angie.orcam.me.uk/>.

  Maciej

[PATCH v3 01/10] MIPS: DEC: Ensure 32-bit stack location for o32 prom_printf()

[Maciej W. Rozycki]

In 64-bit configurations calling any firmware entry points from a kernel 
thread other than the initial one will result in a situation where the 
stack has been placed in the XKPHYS 64-bit memory segment.

Consequently the stack pointer is no longer a 32-bit value and when the 
32-bit firmware code called uses 32-bit ALU operations to manipulate the 
stack pointer, the calculated result is incorrect (in fact in the 64-bit 
MIPS ISA almost all 32-bit ALU operations will produce an unpredictable 
result when executed on 64-bit data) and control goes astray.

This may happen when no final console driver has been enabled in the
configuration and consequently the initial console continues being used
late into bootstrap, or with an upcoming change that will switch the zs
driver to use a platform device, which in turn will make the console
handover happen only after other kernel threads have already been
started, and the kernel will hang at:

  pid_max: default: 32768 minimum: 301

or somewhat later, but always before:

  cblist_init_generic: Setting adjustable number of callback queues.

has been printed.

It seems that only the prom_printf() entry point is affected.  Of all 
the other entry points wired only rex_slot_address() and rex_gettcinfo() 
are called from a kernel thread other than the initial one, specifically 
kernel_init(), and they are leaf functions that do no business with the 
stack, having worked with no issue ever since 64-bit support was added 
for the platform back in 2002.

To address this issue then, arrange for the stack to be switched in the 
o32 wrapper as required for prom_printf() only, by supplying call_o32() 
with a pointer to a chunk of initdata space, which is placed in the 
CKSEG0 32-bit compatibility segment, observing that prom_printf() is 
only called from console output handler and therefore with the console 
lock held, implying no need for this code to be reentrant.

Other firmware entry points may be called with interrupts enabled and no 
lock held, and may therefore require that call_o32() be reentrant.  They 
trigger no issue at this point and "if it ain't broke, don't fix it," so 
just leave them alone.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # v2.6.12+
---
No change from v2, 
<https://lore.kernel.org/r/alpine.DEB.2.21.2605010132340.38805@angie.orcam.me.uk/>.

New change in v2.
---
 arch/mips/dec/prom/init.c        |    6 +++++-
 arch/mips/include/asm/dec/prom.h |   15 +++++++++++++--
 2 files changed, 18 insertions(+), 3 deletions(-)

linux-mips-dec-call-o32-stack.diff
Index: linux-macro/arch/mips/dec/prom/init.c
===================================================================
--- linux-macro.orig/arch/mips/dec/prom/init.c
+++ linux-macro/arch/mips/dec/prom/init.c
@@ -3,7 +3,7 @@
  * init.c: PROM library initialisation code.
  *
  * Copyright (C) 1998 Harald Koerfgen
- * Copyright (C) 2002, 2004  Maciej W. Rozycki
+ * Copyright (C) 2002, 2004, 2026  Maciej W. Rozycki
  */
 #include <linux/init.h>
 #include <linux/kernel.h>
@@ -20,6 +20,10 @@
 #include <asm/dec/prom.h>
 
 
+#ifdef CONFIG_64BIT
+unsigned long o32_stk[O32_STK_SIZE] __initdata = { 0 };
+#endif
+
 int (*__rex_bootinit)(void);
 int (*__rex_bootread)(void);
 int (*__rex_getbitmap)(memmap *);
Index: linux-macro/arch/mips/include/asm/dec/prom.h
===================================================================
--- linux-macro.orig/arch/mips/include/asm/dec/prom.h
+++ linux-macro/arch/mips/include/asm/dec/prom.h
@@ -4,7 +4,7 @@
  *
  *	DECstation PROM interface.
  *
- *	Copyright (C) 2002  Maciej W. Rozycki
+ *	Copyright (C) 2002, 2026  Maciej W. Rozycki
  *
  *	Based on arch/mips/dec/prom/prom.h by the Anonymous.
  */
@@ -97,6 +97,17 @@ extern int (*__pmax_close)(int);
 
 #ifdef CONFIG_64BIT
 
+#define O32_STK_SIZE 512
+extern unsigned long o32_stk[];
+
+/* Switch the stack if outside the 32-bit address space.  */
+static inline unsigned long *o32_get_stk(void)
+{
+	long fp = (long)__builtin_frame_address(0);
+
+	return fp != (int)fp ? o32_stk + O32_STK_SIZE : NULL;
+}
+
 /*
  * On MIPS64 we have to call PROM functions via a helper
  * dispatcher to accommodate ABI incompatibilities.
@@ -128,7 +139,7 @@ int __DEC_PROM_O32(_prom_printf, (int (*
 
 #define prom_getchar()		_prom_getchar(__prom_getchar, NULL)
 #define prom_getenv(x)		_prom_getenv(__prom_getenv, NULL, x)
-#define prom_printf(x...)	_prom_printf(__prom_printf, NULL, x)
+#define prom_printf(x...)	_prom_printf(__prom_printf, o32_get_stk(), x)
 
 #else /* !CONFIG_64BIT */
 

[PATCH v3 02/10] MIPS: DEC: Prevent initial console buffer from landing in XKPHYS

[Maciej W. Rozycki]

In 64-bit configurations calling the initial console output handler from 
a kernel thread other than the initial one will result in a situation 
where the stack has been placed in the XKPHYS 64-bit memory segment and 
consequently so has been the buffer allocated there that is used as the 
argument corresponding to the `%s' output conversion specifier for the 
firmware's printf() entry point.

This 64-bit address will then be truncated by 32-bit firmware, resulting 
in an attempt to access the wrong memory location, which in turn will 
cause all kinds of unpredictable behaviour, such as a kernel crash:

  Console: colour dummy device 160x64
  Calibrating delay loop... 49.36 BogoMIPS (lpj=192512)
  pid_max: default: 32768 minimum: 301
  CPU 0 Unable to handle kernel paging request at virtual address 000000000203bd00, epc == ffffffffbfc08364, ra == ffffffffbfc08800
  Oops[#1]:
  CPU: 0 PID: 0 Comm: swapper Not tainted 5.18.0-rc2-00254-gfb649bda6f56-dirty #121
  $ 0   : 0000000000000000 0000000000000001 0000000000000023 ffffffff80684ba0
  $ 4   : 000000000203bd00 ffffffffbfc0f3b4 ffffffffffffffff 0000000000000073
  $ 8   : 0a303d7469000000 0000000000000000 0000000000000073 ffffffffbfc0f473
  $12   : 0000000000000002 0000000000000000 ffffffff80684c1c 0000000000000000
  $16   : 0000000000000000 ffffffff80596dc9 0000000000000000 ffffffffbfc09240
  $20   : ffffffff80684c40 ffffffffbfc0f400 000000000000002d 000000000000002b
  $24   : ffffffffffffffbf 000000000203bd00                                  
  $28   : ffffffff805f0000 ffffffff80684b58 0000000000000030 ffffffffbfc08800
  Hi    : 0000000000000000
  Lo    : 0000000000000aa8
  epc   : ffffffffbfc08364 0xffffffffbfc08364
  ra    : ffffffffbfc08800 0xffffffffbfc08800
  Status: 140120e2        KX SX UX KERNEL EXL 
  Cause : 00000008 (ExcCode 02)
  BadVA : 000000000203bd00
  PrId  : 00000430 (R4000SC)
  Modules linked in:
  Process swapper (pid: 0, threadinfo=(____ptrval____), task=(____ptrval____), tls=0000000000000000)
  Stack : 0000000000000000 0000000000000000 0000000000000000 0000004d0000004d
          80684cc0806a2a40 80596dc80000004d 8061000000000000 bfc0850c80684c38
          0000000000000000 000000000203bd00 0000000000000000 0000000000000000
          0000000000000000 00000000bfc0f3b4 0000000000000000 0000000000000000
          0000000000000000 0000000000000000 0000000000000000 0000000000000000
          0000000000000000 0000000000000000 0000000000000000 0000000000000000
          0000002500000000 0000000000000000 0000000000000000 802c1a7400000000
          0203bd0080596dc8 0203bd4d69000000 6c61632000000018 5f746567646e6172
          6c616320625f6d6f 5f736e5f6d6f7266 206361323778302b 303d74696e726320
          806a0a38806b0000 806a0a38806b0000 00000000806b0000 80683c58806b0000
          ...
  Call Trace:
  
  
  Code: a082ffff  03e00008  00601021 <80820000> 00001821  10400005  24840001  80820000  24630001 
  
  ---[ end trace 0000000000000000 ]---
  Kernel panic - not syncing: Fatal exception in interrupt
  
  KN04 V2.1k    (PC: 0xa0026768, SP: 0x806848e8)
  >>

In this case the pointer in $4 was truncated from 0x980000000203bd00 to 
0x000000000203bd00.

This may happen when no final console driver has been enabled in the 
configuration and consequently the initial console continues being used 
late into bootstrap or with an upcoming change that will switch the zs 
driver to use a platform device, which in turn will make the console 
handover happen only after other kernel threads have already been 
started.

Fix the issue by making the buffer static and initdata, and therefore 
placed in the CKSEG0 32-bit compatibility segment, observing that the 
console output handler is called with the console lock held, implying 
no need for this code to be reentrant.  Add an assertion to verify the 
buffer actually has been placed in a compatibility segment.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # v2.6.12+
---
Changes from v2, 
<https://lore.kernel.org/r/alpine.DEB.2.21.2605010059310.38805@angie.orcam.me.uk/>:

- Cast `buf' to `long' before casting to `int', fixing:

  In file included from ./include/linux/bug.h:5,
                   from arch/mips/dec/prom/console.c:7:
  arch/mips/dec/prom/console.c: In function 'prom_console_write':
  arch/mips/dec/prom/console.c:21:29: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
     21 |         BUG_ON((long)buf != (int)buf);
        |                             ^
  ./arch/mips/include/asm/bug.h:34:44: note: in definition of macro 'BUG_ON'
     34 | #define BUG_ON(C) __BUG_ON((unsigned long)(C))
        |                                            ^

New change in v2.
---
 arch/mips/dec/prom/console.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

linux-mips-dec-prom-console-xkphys.diff
Index: linux-macro/arch/mips/dec/prom/console.c
===================================================================
--- linux-macro.orig/arch/mips/dec/prom/console.c
+++ linux-macro/arch/mips/dec/prom/console.c
@@ -2,8 +2,9 @@
 /*
  *	DECstation PROM-based early console support.
  *
- *	Copyright (C) 2004, 2007  Maciej W. Rozycki
+ *	Copyright (C) 2004, 2007, 2026  Maciej W. Rozycki
  */
+#include <linux/bug.h>
 #include <linux/console.h>
 #include <linux/init.h>
 #include <linux/kernel.h>
@@ -14,9 +15,11 @@
 static void __init prom_console_write(struct console *con, const char *s,
 				      unsigned int c)
 {
-	char buf[81];
+	static char buf[81] __initdata = { 0 };
 	unsigned int chunk = sizeof(buf) - 1;
 
+	BUG_ON((long)buf != (int)(long)buf);
+
 	while (c > 0) {
 		if (chunk > c)
 			chunk = c;

[PATCH v3 03/10] serial: dz: Fix bootconsole message clobbering at chip reset

[Maciej W. Rozycki]

In the DZ interface as implemented by the DC7085 gate array the serial 
transmitters are double buffered, meaning that at the time a transmitter 
is ready to accept the next character there is one in the transmit shift 
register still being sent to the line.  Issuing a master clear at this 
time causes this character to be lost, so wait an extra amount of time 
sufficient for the transmit shift register to drain at 9600bps, which is 
the baud rate setting used by the firmware console.

Mind the specified 1.4us TRDY recovery time in the course and continue 
using iob() as the completion barrier, since the platforms involved use 
a write buffer that can delay and combine writes, and reorder them with 
respect to reads regardless of the MMIO locations accessed and we still 
lack a platform-independent handler for that.

When called from dz_serial_console_init() this is too early for fsleep() 
to work and even before lpj has been calculated and therefore the delay 
is actually not sufficient for the transmitter to drain and is merely a 
placeholder now.  This will be addressed in a follow-up change.

Fixes: e6ee512f5a77 ("dz.c: Resource management")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # v2.6.25+
---
No change from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605011617300.60398@angie.orcam.me.uk/>.

No change from v1 (1/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604121508060.29980@angie.orcam.me.uk/>.
---
 drivers/tty/serial/dz.c |   21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

linux-serial-dz-reset-drain.diff
Index: linux-macro/drivers/tty/serial/dz.c
===================================================================
--- linux-macro.orig/drivers/tty/serial/dz.c
+++ linux-macro/drivers/tty/serial/dz.c
@@ -542,10 +542,31 @@ static int dz_encode_baud_rate(unsigned
 static void dz_reset(struct dz_port *dport)
 {
 	struct dz_mux *mux = dport->mux;
+	unsigned short tcr;
+	int loops = 10000;
 
 	if (mux->initialised)
 		return;
 
+	tcr = dz_in(dport, DZ_TCR);
+
+	/* Do not disturb any ongoing transmissions.  */
+	if (dz_in(dport, DZ_CSR) & DZ_MSE) {
+		unsigned short csr, mask;
+
+		mask = tcr;
+		while ((mask & DZ_LNENB) && loops--) {
+			csr = dz_in(dport, DZ_CSR);
+			if (!(csr & DZ_TRDY))
+				continue;
+			mask &= ~(1 << ((csr & DZ_TLINE) >> 8));
+			dz_out(dport, DZ_TCR, mask);
+			iob();
+			udelay(2);		/* 1.4us TRDY recovery.  */
+		}
+		udelay(1200);			/* Transmitter drain.  */
+	}
+
 	dz_out(dport, DZ_CSR, DZ_CLR);
 	while (dz_in(dport, DZ_CSR) & DZ_CLR);
 	iob();

[PATCH v3 04/10] serial: dz: Fix bootconsole handover lockup

[Maciej W. Rozycki]

Calling dz_reset() in the course of setting up the serial device causes 
line parameters to be reset and the transmitter disabled.  We've been 
lucky in that no message is usually produced to the kernel log between 
this call and the later call to uart_set_options() in the course of 
console setup done by dz_serial_console_init(), or the system would hang 
as the console output handler in the firmware tried to access a port the 
transmitter of which has been disabled and line parameters messed up.

This will change with the next change to the driver, so fix dz_reset() 
such that line parameters are set for 9600n8 console operation as with 
the system firmware and the transmitter re-enabled after reset.  This 
also means dz_pm() serves no purpose anymore, so drop it.

Fixes: e6ee512f5a77 ("dz.c: Resource management")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # v2.6.25+
---
Changes from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605012013410.11074@angie.orcam.me.uk/>:

- Make a minor style improvement to the commit description.

No change from v1 (2/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604102338300.29980@angie.orcam.me.uk/>.
---
 drivers/tty/serial/dz.c |   36 ++++++++++++------------------------
 1 file changed, 12 insertions(+), 24 deletions(-)

linux-serial-dz-prom-console.diff
Index: linux-macro/drivers/tty/serial/dz.c
===================================================================
--- linux-macro.orig/drivers/tty/serial/dz.c
+++ linux-macro/drivers/tty/serial/dz.c
@@ -571,6 +571,18 @@ static void dz_reset(struct dz_port *dpo
 	while (dz_in(dport, DZ_CSR) & DZ_CLR);
 	iob();
 
+	/*
+	 * Set parameters across all lines such as not to interfere
+	 * with the initial PROM-based console.  Otherwise any output
+	 * produced before the console handover would cause the system
+	 * firmware to produce rubbish.
+	 */
+	for (int line = 0; line < DZ_NB_PORT; line++)
+		dz_out(dport, DZ_LPR, DZ_B9600 | DZ_CS8 | line);
+
+	/* Re-enable transmission for the initial PROM-based console.  */
+	dz_out(dport, DZ_TCR, tcr);
+
 	/* Enable scanning.  */
 	dz_out(dport, DZ_CSR, DZ_MSE);
 
@@ -654,26 +666,6 @@ static void dz_set_termios(struct uart_p
 	uart_port_unlock_irqrestore(&dport->port, flags);
 }
 
-/*
- * Hack alert!
- * Required solely so that the initial PROM-based console
- * works undisturbed in parallel with this one.
- */
-static void dz_pm(struct uart_port *uport, unsigned int state,
-		  unsigned int oldstate)
-{
-	struct dz_port *dport = to_dport(uport);
-	unsigned long flags;
-
-	uart_port_lock_irqsave(&dport->port, &flags);
-	if (state < 3)
-		dz_start_tx(&dport->port);
-	else
-		dz_stop_tx(&dport->port);
-	uart_port_unlock_irqrestore(&dport->port, flags);
-}
-
-
 static const char *dz_type(struct uart_port *uport)
 {
 	return "DZ";
@@ -769,7 +761,6 @@ static const struct uart_ops dz_ops = {
 	.startup	= dz_startup,
 	.shutdown	= dz_shutdown,
 	.set_termios	= dz_set_termios,
-	.pm		= dz_pm,
 	.type		= dz_type,
 	.release_port	= dz_release_port,
 	.request_port	= dz_request_port,
@@ -894,10 +885,7 @@ static int __init dz_console_setup(struc
 	if (ret)
 		return ret;
 
-	spin_lock_init(&dport->port.lock);	/* For dz_pm().  */
-
 	dz_reset(dport);
-	dz_pm(uport, 0, -1);
 
 	if (options)
 		uart_parse_options(options, &baud, &parity, &bits, &flow);

[PATCH v3 05/10] serial: zs: Fix bootconsole handover lockup

[Maciej W. Rozycki]

Calling zs_reset() in the course of setting up the serial device causes 
line parameters to be reset and the transmitter disabled.  We've been 
lucky in that no message is usually produced to the kernel log between 
this call and the later call to uart_set_options() in the course of 
console setup done by zs_serial_console_init(), or the system would hang 
as the console output handler in the firmware tried to access a port the
transmitter of which has been disabled and line parameters messed up.

This will change with the next change to the driver, so fix zs_reset() 
such that line parameters are set for 9600n8 console operation as with 
the system firmware and the transmitter re-enabled after reset.  This 
also means zs_pm() serves no purpose anymore, so drop it.

Fixes: 8b4a40809e53 ("zs: move to the serial subsystem")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # v2.6.23+
---
Changes from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605012016460.11074@angie.orcam.me.uk/>:

- Make a minor style improvement to the commit description.

- Drop a leftover paragraph in the commit description (doh!).

No change from v1 (3/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604102346290.29980@angie.orcam.me.uk/>.
---
 drivers/tty/serial/zs.c |   29 ++++++++---------------------
 1 file changed, 8 insertions(+), 21 deletions(-)

linux-serial-zs-prom-console.diff
Index: linux-macro/drivers/tty/serial/zs.c
===================================================================
--- linux-macro.orig/drivers/tty/serial/zs.c
+++ linux-macro/drivers/tty/serial/zs.c
@@ -105,18 +105,24 @@ struct zs_parms {
 
 static struct zs_scc zs_sccs[ZS_NUM_SCCS];
 
+/*
+ * Set parameters in WR5, WR12, WR13 such as not to interfere
+ * with the initial PROM-based console.  Otherwise any output
+ * produced before the console handover would cause the system
+ * firmware to hang (TxENAB) or produce rubbish (Tx8, B9600).
+ */
 static u8 zs_init_regs[ZS_NUM_REGS] __initdata = {
 	0,				/* write 0 */
 	PAR_SPEC,			/* write 1 */
 	0,				/* write 2 */
 	0,				/* write 3 */
 	X16CLK | SB1,			/* write 4 */
-	0,				/* write 5 */
+	Tx8 | TxENAB,			/* write 5 */
 	0, 0, 0,			/* write 6, 7, 8 */
 	MIE | DLC | NV,			/* write 9 */
 	NRZ,				/* write 10 */
 	TCBR | RCBR,			/* write 11 */
-	0, 0,				/* BRG time constant, write 12 + 13 */
+	0x16, 0x00,			/* BRG time constant, write 12 + 13 */
 	BRSRC | BRENABL,		/* write 14 */
 	0,				/* write 15 */
 };
@@ -956,23 +962,6 @@ static void zs_set_termios(struct uart_p
 	spin_unlock_irqrestore(&scc->zlock, flags);
 }
 
-/*
- * Hack alert!
- * Required solely so that the initial PROM-based console
- * works undisturbed in parallel with this one.
- */
-static void zs_pm(struct uart_port *uport, unsigned int state,
-		  unsigned int oldstate)
-{
-	struct zs_port *zport = to_zport(uport);
-
-	if (state < 3)
-		zport->regs[5] |= TxENAB;
-	else
-		zport->regs[5] &= ~TxENAB;
-	write_zsreg(zport, R5, zport->regs[5]);
-}
-
 
 static const char *zs_type(struct uart_port *uport)
 {
@@ -1055,7 +1044,6 @@ static const struct uart_ops zs_ops = {
 	.startup	= zs_startup,
 	.shutdown	= zs_shutdown,
 	.set_termios	= zs_set_termios,
-	.pm		= zs_pm,
 	.type		= zs_type,
 	.release_port	= zs_release_port,
 	.request_port	= zs_request_port,
@@ -1210,7 +1198,6 @@ static int __init zs_console_setup(struc
 		return ret;
 
 	zs_reset(zport);
-	zs_pm(uport, 0, -1);
 
 	if (options)
 		uart_parse_options(options, &baud, &parity, &bits, &flow);

[PATCH v3 06/10] serial: zs: Switch to using channel reset

[Maciej W. Rozycki]

Switch the driver to using the channel reset rather than hardware reset, 
simplifying handling by removing an interference between channels that 
causes the other channel to become uninitialised afterwards.

There is little difference between the two kinds of reset in terms of 
register settings that result, and we initialise the whole register set 
right away anyway.  However this prevents a hang from happening should 
the console output handler in the firmware try to access the other port 
whose transmitter has been disabled and line parameters messed up.

For example this will happen if the keyboard port (port A) is chosen for 
the system console, unusually but not insanely for a headless system, as 
the port is wired to a standard DA-15 connector and an adapter can be 
easily made.  Or with the next change in place this would happen for the 
regular console port (port B), since the keyboard port (port A) will be 
initialised first.

Just remove the unnecessary complication then, a channel reset is good 
enough.  We still need the initialisation marker, now per channel rather 
than per SCC, as for the console port zs_reset() will be called twice: 
once early on via zs_serial_console_init() for the console setup only, 
and then again via zs_config_port() as the port is associated with a TTY 
device.

Fixes: 8b4a40809e53 ("zs: move to the serial subsystem")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # v2.6.23+
---
No change from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605012041220.11074@angie.orcam.me.uk/>.

No change from v1 (4/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604122134330.29980@angie.orcam.me.uk/>.
---
 drivers/tty/serial/zs.c |    7 ++++---
 drivers/tty/serial/zs.h |    2 +-
 2 files changed, 5 insertions(+), 4 deletions(-)

linux-serial-zs-reset-channel.diff
Index: linux-macro/drivers/tty/serial/zs.c
===================================================================
--- linux-macro.orig/drivers/tty/serial/zs.c
+++ linux-macro/drivers/tty/serial/zs.c
@@ -832,21 +832,22 @@ static void zs_shutdown(struct uart_port
 
 static void zs_reset(struct zs_port *zport)
 {
+	struct zs_port *zport_a = &zport->scc->zport[ZS_CHAN_A];
 	struct zs_scc *scc = zport->scc;
 	int irq;
 	unsigned long flags;
 
 	spin_lock_irqsave(&scc->zlock, flags);
 	irq = !irqs_disabled_flags(flags);
-	if (!scc->initialised) {
+	if (!zport->initialised) {
 		/* Reset the pointer first, just in case...  */
 		read_zsreg(zport, R0);
 		/* And let the current transmission finish.  */
 		zs_line_drain(zport, irq);
-		write_zsreg(zport, R9, FHWRES);
+		write_zsreg(zport, R9, zport == zport_a ? CHRA : CHRB);
 		udelay(10);
 		write_zsreg(zport, R9, 0);
-		scc->initialised = 1;
+		zport->initialised = 1;
 	}
 	load_zsregs(zport, zport->regs, irq);
 	spin_unlock_irqrestore(&scc->zlock, flags);
Index: linux-macro/drivers/tty/serial/zs.h
===================================================================
--- linux-macro.orig/drivers/tty/serial/zs.h
+++ linux-macro/drivers/tty/serial/zs.h
@@ -22,6 +22,7 @@
 struct zs_port {
 	struct zs_scc	*scc;			/* Containing SCC.  */
 	struct uart_port port;			/* Underlying UART.  */
+	int		initialised;		/* For the console port.  */
 
 	int		clk_mode;		/* May be 1, 16, 32, or 64.  */
 
@@ -41,7 +42,6 @@ struct zs_scc {
 	struct zs_port	zport[2];
 	spinlock_t	zlock;
 	atomic_t	irq_guard;
-	int		initialised;
 };
 
 #endif /* __KERNEL__ */

[PATCH v3 07/10] serial: dz: Convert to use a platform device

[Maciej W. Rozycki]

Prevent a crash from happening as the first serial port is initialised:

  Console: switching to colour frame buffer device 160x64
  tgafb: SFB+ detected, rev=0x02
  fb0: Digital ZLX-E1 frame buffer device at 0x1e000000
  DECstation DZ serial driver version 1.04
  CPU 0 Unable to handle kernel paging request at virtual address 000000bc, epc == 8048b3a4, ra == 80470a78
  Oops[#1]:
  CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.19.0-dirty #35 NONE 
  $ 0   : 00000000 1000ac00 00000004 804707ac
  $ 4   : 00000000 80e20850 80e20858 81000030
  $ 8   : 00000000 8072c81c 00000008 fefefeff
  $12   : 6c616972 00000006 80c5917f 69726420
  $16   : 80e20800 00000000 808f8968 80e20800
  $20   : 00000000 807f5a90 808b0094 808d3bc8
  $24   : 00000018 80479030                  
  $28   : 80c2e000 80c2fd70 00000069 80470a78
  Hi    : 00000004
  Lo    : 00000000
  epc   : 8048b3a4 __dev_fwnode+0x0/0xc
  ra    : 80470a78 serial_base_ctrl_add+0xa0/0x168
  Status: 1000ac04	IEp 
  Cause : 30000008 (ExcCode 02)
  BadVA : 000000bc
  PrId  : 00000220 (R3000)
  Modules linked in:
  Process swapper/0 (pid: 1, threadinfo=(ptrval), task=(ptrval), tls=00000000)
  Stack : 00400044 00400040 8046f4cc 00000000 808a6148 808a0000 808f8968 8086983c
          808e0000 8046fc84 1000ac01 00000028 80e20700 802ba3f8 80e20700 80d34a94
          80c1b900 80e20700 80e20700 80e20700 80e20700 80444650 00000000 00000000
          00000000 807f5a90 808b0094 80447080 00400040 808e0000 80d34a94 808a6148
          80d34a94 00000004 80e20700 00000000 8076974c 80469810 80c2fe3c 1000ac01
          ...
  Call Trace:
  [<8048b3a4>] __dev_fwnode+0x0/0xc
  [<80470a78>] serial_base_ctrl_add+0xa0/0x168
  [<8046fc84>] serial_core_register_port+0x1c8/0x974
  [<808c6af0>] dz_init+0x74/0xc8
  [<800470e0>] do_one_initcall+0x44/0x2d4
  [<808b111c>] kernel_init_freeable+0x258/0x308
  [<8072e434>] kernel_init+0x20/0x114
  [<80049cd0>] ret_from_kernel_thread+0x14/0x1c
  
  Code: 27bd0018  03e00008  2402ffea <8c8200bc> 03e00008  00000000  27bdffc0  afbe0038  afb30024 
  
  ---[ end trace 0000000000000000 ]---

-- where a pointer is dereferenced that has been derived from a null 
pointer to the port's parent device.

Since no device is available with legacy probing and it's not anymore a 
preferable way to discover devices anyway, switch the driver to using a 
platform device and use it as the port's parent device.  Update resource 
handling accordingly and only request the actual span of addresses used 
within the slot, which will have had its resource already requested by 
generic platform device code.

Use platform_driver_probe() not just because the DZ device is fixed with 
solder on board and not straightforward to remove, but foremost because 
the associated TTY's major device number is the same as used by the zs 
driver and the first driver to claim it will prevent the other one from 
using it.  Either one DZ device or some SCC devices will be present in a 
given system but never both at a time, and therefore we want the major 
device number to be claimed by the first driver to actually successfully 
bind to its device and platform_driver_probe() is a way to fulfil that.

An unfortunate consequence of the switch to a platform device is we now 
hand the console over from the bootconsole much later in the bootstrap.  
The firmware console handler appears good enough though to work so late 
and in particular with interrupts enabled.

Conversely only starting the console port so late lets the reset code 
fully utilise our delay handlers, so switch from udelay() to fsleep() 
for transmitter draining so as to avoid busy-waiting for an excessive 
amount of time.

Fixes: 84a9582fd203 ("serial: core: Start managing serial controllers to enable runtime PM")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # needs to use .remove_new for <= 6.10
---
No change from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605012043570.11074@angie.orcam.me.uk/>.

No change from v1 (5/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604102351280.29980@angie.orcam.me.uk/>.
---
 arch/mips/dec/platform.c |   55 +++++++++++++++++++++-
 drivers/tty/serial/dz.c  |  116 ++++++++++++++++++++++-------------------------
 2 files changed, 110 insertions(+), 61 deletions(-)

linux-serial-dz-platform.diff
Index: linux-macro/arch/mips/dec/platform.c
===================================================================
--- linux-macro.orig/arch/mips/dec/platform.c
+++ linux-macro/arch/mips/dec/platform.c
@@ -10,6 +10,13 @@
 #include <linux/mc146818rtc.h>
 #include <linux/platform_device.h>
 
+#include <asm/bootinfo.h>
+
+#include <asm/dec/interrupts.h>
+#include <asm/dec/kn01.h>
+#include <asm/dec/kn02.h>
+#include <asm/dec/system.h>
+
 static struct resource dec_rtc_resources[] = {
 	{
 		.name = "rtc",
@@ -30,11 +37,57 @@ static struct platform_device dec_rtc_de
 	.num_resources = ARRAY_SIZE(dec_rtc_resources),
 };
 
+static struct resource dec_dz_resources[] = {
+	{ .name = "dz", .flags = IORESOURCE_MEM, },
+	{ .name = "dz", .flags = IORESOURCE_IRQ, },
+};
+
+static struct platform_device dec_dz_device = {
+	.name = "dz",
+	.id = PLATFORM_DEVID_NONE,
+	.resource = dec_dz_resources,
+	.num_resources = ARRAY_SIZE(dec_dz_resources),
+};
+
+static struct platform_device *dec_dz_devices[] __initdata = {
+	&dec_dz_device,
+};
+
 static int __init dec_add_devices(void)
 {
+	int ret1, ret2;
+	int num_dz;
+	int irq, i;
+
 	dec_rtc_resources[0].start = RTC_PORT(0);
 	dec_rtc_resources[0].end = RTC_PORT(0) + dec_kn_slot_size - 1;
-	return platform_device_register(&dec_rtc_device);
+
+	i = 0;
+	irq = dec_interrupt[DEC_IRQ_DZ11];
+	if (IS_ENABLED(CONFIG_32BIT) && irq >= 0) {
+		resource_size_t base;
+
+		switch (mips_machtype) {
+		case MACH_DS23100:
+		case MACH_DS5100:
+			base = dec_kn_slot_base + KN01_DZ11;
+			break;
+		default:
+			base = dec_kn_slot_base + KN02_DZ11;
+			break;
+		}
+		dec_dz_device.resource[0].start = base;
+		dec_dz_device.resource[0].end = base + dec_kn_slot_size - 1;
+		dec_dz_device.resource[1].start = irq;
+		dec_dz_device.resource[1].end = irq;
+		i++;
+	}
+	num_dz = i;
+
+	ret1 = platform_device_register(&dec_rtc_device);
+	ret2 = IS_ENABLED(CONFIG_32BIT) ?
+	       platform_add_devices(dec_dz_devices, num_dz) : 0;
+	return ret1 ? ret1 : ret2;
 }
 
 device_initcall(dec_add_devices);
Index: linux-macro/drivers/tty/serial/dz.c
===================================================================
--- linux-macro.orig/drivers/tty/serial/dz.c
+++ linux-macro/drivers/tty/serial/dz.c
@@ -40,6 +40,7 @@
 #include <linux/kernel.h>
 #include <linux/major.h>
 #include <linux/module.h>
+#include <linux/platform_device.h>
 #include <linux/serial.h>
 #include <linux/serial_core.h>
 #include <linux/sysrq.h>
@@ -48,14 +49,6 @@
 
 #include <linux/atomic.h>
 #include <linux/io.h>
-#include <asm/bootinfo.h>
-
-#include <asm/dec/interrupts.h>
-#include <asm/dec/kn01.h>
-#include <asm/dec/kn02.h>
-#include <asm/dec/machtype.h>
-#include <asm/dec/prom.h>
-#include <asm/dec/system.h>
 
 #include "dz.h"
 
@@ -65,7 +58,9 @@ MODULE_LICENSE("GPL");
 
 
 static char dz_name[] __initdata = "DECstation DZ serial driver version ";
-static char dz_version[] __initdata = "1.04";
+static char dz_version[] __initdata = "1.05";
+
+#define DZ_IO_SIZE 0x20			/* IOMEM space size.  */
 
 struct dz_port {
 	struct dz_mux		*mux;
@@ -81,6 +76,7 @@ struct dz_mux {
 };
 
 static struct dz_mux dz_mux;
+static struct uart_driver dz_reg;
 
 static inline struct dz_port *to_dport(struct uart_port *uport)
 {
@@ -564,7 +560,7 @@ static void dz_reset(struct dz_port *dpo
 			iob();
 			udelay(2);		/* 1.4us TRDY recovery.  */
 		}
-		udelay(1200);			/* Transmitter drain.  */
+		fsleep(1200);			/* Transmitter drain.  */
 	}
 
 	dz_out(dport, DZ_CSR, DZ_CLR);
@@ -681,14 +677,13 @@ static void dz_release_port(struct uart_
 
 	map_guard = atomic_add_return(-1, &mux->map_guard);
 	if (!map_guard)
-		release_mem_region(uport->mapbase, dec_kn_slot_size);
+		release_mem_region(uport->mapbase, DZ_IO_SIZE);
 }
 
 static int dz_map_port(struct uart_port *uport)
 {
 	if (!uport->membase)
-		uport->membase = ioremap(uport->mapbase,
-						 dec_kn_slot_size);
+		uport->membase = ioremap(uport->mapbase, DZ_IO_SIZE);
 	if (!uport->membase) {
 		printk(KERN_ERR "dz: Cannot map MMIO\n");
 		return -ENOMEM;
@@ -704,8 +699,7 @@ static int dz_request_port(struct uart_p
 
 	map_guard = atomic_add_return(1, &mux->map_guard);
 	if (map_guard == 1) {
-		if (!request_mem_region(uport->mapbase, dec_kn_slot_size,
-					"dz")) {
+		if (!request_mem_region(uport->mapbase, DZ_IO_SIZE, "dz")) {
 			atomic_add(-1, &mux->map_guard);
 			printk(KERN_ERR
 			       "dz: Unable to reserve MMIO resource\n");
@@ -716,7 +710,7 @@ static int dz_request_port(struct uart_p
 	if (ret) {
 		map_guard = atomic_add_return(-1, &mux->map_guard);
 		if (!map_guard)
-			release_mem_region(uport->mapbase, dec_kn_slot_size);
+			release_mem_region(uport->mapbase, DZ_IO_SIZE);
 		return ret;
 	}
 	return 0;
@@ -768,20 +762,15 @@ static const struct uart_ops dz_ops = {
 	.verify_port	= dz_verify_port,
 };
 
-static void __init dz_init_ports(void)
+static int __init dz_probe(struct platform_device *pdev)
 {
-	static int first = 1;
-	unsigned long base;
+	struct resource *mem_resource, *irq_resource;
 	int line;
 
-	if (!first)
-		return;
-	first = 0;
-
-	if (mips_machtype == MACH_DS23100 || mips_machtype == MACH_DS5100)
-		base = dec_kn_slot_base + KN01_DZ11;
-	else
-		base = dec_kn_slot_base + KN02_DZ11;
+	mem_resource = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq_resource = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
+	if (!mem_resource || !irq_resource)
+		return -ENODEV;
 
 	for (line = 0; line < DZ_NB_PORT; line++) {
 		struct dz_port *dport = &dz_mux.dport[line];
@@ -789,14 +778,33 @@ static void __init dz_init_ports(void)
 
 		dport->mux	= &dz_mux;
 
-		uport->irq	= dec_interrupt[DEC_IRQ_DZ11];
+		uport->dev	= &pdev->dev;
+		uport->irq	= irq_resource->start;
 		uport->fifosize	= 1;
 		uport->iotype	= UPIO_MEM;
 		uport->flags	= UPF_BOOT_AUTOCONF;
 		uport->ops	= &dz_ops;
 		uport->line	= line;
-		uport->mapbase	= base;
+		uport->mapbase	= mem_resource->start;
 		uport->has_sysrq = IS_ENABLED(CONFIG_SERIAL_DZ_CONSOLE);
+
+		if (uart_add_one_port(&dz_reg, uport))
+			uport->dev = NULL;
+	}
+
+	return 0;
+}
+
+static void __exit dz_remove(struct platform_device *pdev)
+{
+	int line;
+
+	for (line = DZ_NB_PORT - 1; line >= 0; line--) {
+		struct dz_port *dport = &dz_mux.dport[line];
+		struct uart_port *uport = &dport->port;
+
+		if (uport->dev)
+			uart_remove_one_port(&dz_reg, uport);
 	}
 }
 
@@ -879,21 +887,14 @@ static int __init dz_console_setup(struc
 	int bits = 8;
 	int parity = 'n';
 	int flow = 'n';
-	int ret;
-
-	ret = dz_map_port(uport);
-	if (ret)
-		return ret;
-
-	dz_reset(dport);
 
+	if (!dport->mux)
+		return -ENODEV;
 	if (options)
 		uart_parse_options(options, &baud, &parity, &bits, &flow);
-
-	return uart_set_options(&dport->port, co, baud, parity, bits, flow);
+	return uart_set_options(uport, co, baud, parity, bits, flow);
 }
 
-static struct uart_driver dz_reg;
 static struct console dz_console = {
 	.name	= "ttyS",
 	.write	= dz_console_print,
@@ -904,18 +905,6 @@ static struct console dz_console = {
 	.data	= &dz_reg,
 };
 
-static int __init dz_serial_console_init(void)
-{
-	if (!IOASIC) {
-		dz_init_ports();
-		register_console(&dz_console);
-		return 0;
-	} else
-		return -ENXIO;
-}
-
-console_initcall(dz_serial_console_init);
-
 #define SERIAL_DZ_CONSOLE	&dz_console
 #else
 #define SERIAL_DZ_CONSOLE	NULL
@@ -931,25 +920,32 @@ static struct uart_driver dz_reg = {
 	.cons			= SERIAL_DZ_CONSOLE,
 };
 
+static struct platform_driver dz_driver = {
+	.remove = __exit_p(dz_remove),
+	.driver = { .name = "dz" },
+};
+
 static int __init dz_init(void)
 {
-	int ret, i;
-
-	if (IOASIC)
-		return -ENXIO;
+	int ret;
 
 	printk("%s%s\n", dz_name, dz_version);
 
-	dz_init_ports();
-
 	ret = uart_register_driver(&dz_reg);
 	if (ret)
 		return ret;
+	ret = platform_driver_probe(&dz_driver, dz_probe);
+	if (ret)
+		uart_unregister_driver(&dz_reg);
 
-	for (i = 0; i < DZ_NB_PORT; i++)
-		uart_add_one_port(&dz_reg, &dz_mux.dport[i].port);
+	return ret;
+}
 
-	return 0;
+static void __exit dz_exit(void)
+{
+	platform_driver_unregister(&dz_driver);
+	uart_unregister_driver(&dz_reg);
 }
 
 module_init(dz_init);
+module_exit(dz_exit);

[PATCH v3 08/10] serial: zs: Convert to use a platform device

[Maciej W. Rozycki]

Prevent a crash from happening as the first serial port is initialised:

  Console: switching to mono frame buffer device 160x64
  fb0: PMAG-AA frame buffer device at tc0
  DECstation Z85C30 serial driver version 0.10
  CPU 0 Unable to handle kernel paging request at virtual address 0000002c, epc == 803ab00c, ra == 803aafe0
  Oops[#1]:
  CPU: 0 PID: 1 Comm: swapper Not tainted 6.4.0-rc3-00031-g84a9582fd203-dirty #57
  $ 0   : 00000000 10012c00 803aaeb0 00000000
  $ 4   : 80e12f60 80e12f50 80e12f58 81000030
  $ 8   : 00000000 805ff37c 00000000 33433538
  $12   : 65732030 00000006 80c2915d 6c616972
  $16   : 80e12f00 807b7630 00000000 00000000
  $20   : 00000004 00000348 000001a0 807623b8
  $24   : 00000018 00000000                  
  $28   : 80c24000 80c25d60 8078b148 803aafe0
  Hi    : 00000000
  Lo    : 00000000
  epc   : 803ab00c serial_base_ctrl_add+0x78/0xf4
  ra    : 803aafe0 serial_base_ctrl_add+0x4c/0xf4
  Status: 10012c03	KERNEL EXL IE 
  Cause : 00000008 (ExcCode 02)
  BadVA : 0000002c
  PrId  : 00000440 (R4400SC)
  Modules linked in:
  Process swapper (pid: 1, threadinfo=(ptrval), task=(ptrval), tls=00000000)
  Stack : 80760000 00000cc0 00400044 00400040 803aa02c 80d61ab8 00000000 807b7630
          80760000 807623b8 807b7628 803aa644 80386998 00000000 80e17780 80220f68
          80e17780 80d61ab8 80c17d80 80e17780 80e17780 8063c798 80e17780 80383fa0
          00000010 80e17780 00000000 80386998 807a0000 00000000 00400040 8038f848
          807623b8 80d61ab8 00000004 80e17780 00000000 803a68e4 80c25e2c 803bb884
          ...
  Call Trace:
  [<803ab00c>] serial_base_ctrl_add+0x78/0xf4
  [<803aa644>] serial_core_register_port+0x174/0x69c
  [<8077e9ac>] zs_init+0xc8/0xfc
  [<800404d4>] do_one_initcall+0x40/0x2ac
  [<8076cecc>] kernel_init_freeable+0x1e4/0x270
  [<80605bec>] kernel_init+0x20/0x108
  [<800431e8>] ret_from_kernel_thread+0x14/0x1c
  
  Code: 2442aeb0  ae120024  ae0200d0 <8c67002c> 50e00001  8c670000  3c06806e  3c05806e  afb30010 
  
  ---[ end trace 0000000000000000 ]---

(report at the offending commit) -- where a pointer is dereferenced that 
has been derived from a null pointer to the port's parent device.

Since no device is available with legacy probing and it's not anymore a
preferable way to discover devices anyway, switch the driver to using a
platform device and use it as the port's parent device.  Update resource
handling accordingly and only request the actual span of addresses used
within the slot, which will have had its resource already requested by
generic platform device code.

Use platform_driver_probe() not just because SCC devices are fixed with 
solder on board and not straightforward to remove, but foremost because 
the associated TTY's major device number is the same as used by the dz 
driver and the first driver to claim it will prevent the other one from 
using it.  Either one DZ device or some SCC devices will be present in a 
given system but never both at a time, and therefore we want the major 
device number to be claimed by the first driver to actually successfully 
bind to its device and platform_driver_probe() is a way to fulfil that.

An unfortunate consequence of the switch to a platform device is we now
hand the console over from the bootconsole much later in the bootstrap.
The firmware console handler appears good enough though to work so late
and in particular with interrupts enabled.

Since there is one way only remaining to reach zs_reset() now, remove 
the port initialisation marker as no longer needed and go through the 
channel reset unconditionally.

Fixes: 84a9582fd203 ("serial: core: Start managing serial controllers to enable runtime PM")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: stable@vger.kernel.org # needs to use .remove_new for <= 6.10
---
No change from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605012100520.11074@angie.orcam.me.uk/>.

No change from v1 (6/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604110034160.29980@angie.orcam.me.uk/>.
---
 arch/mips/dec/platform.c |   60 ++++++++++++++-
 drivers/tty/serial/zs.c  |  184 +++++++++++++++++------------------------------
 drivers/tty/serial/zs.h  |    1 
 3 files changed, 125 insertions(+), 120 deletions(-)

linux-serial-zs-platform.diff
Index: linux-macro/arch/mips/dec/platform.c
===================================================================
--- linux-macro.orig/arch/mips/dec/platform.c
+++ linux-macro/arch/mips/dec/platform.c
@@ -13,6 +13,7 @@
 #include <asm/bootinfo.h>
 
 #include <asm/dec/interrupts.h>
+#include <asm/dec/ioasic_addrs.h>
 #include <asm/dec/kn01.h>
 #include <asm/dec/kn02.h>
 #include <asm/dec/system.h>
@@ -53,10 +54,37 @@ static struct platform_device *dec_dz_de
 	&dec_dz_device,
 };
 
+static struct resource dec_zs_resources[][2] = {
+	{
+		{ .name = "scc0", .flags = IORESOURCE_MEM, },
+		{ .name = "scc0", .flags = IORESOURCE_IRQ, },
+	},
+	{
+		{ .name = "scc1", .flags = IORESOURCE_MEM, },
+		{ .name = "scc1", .flags = IORESOURCE_IRQ, },
+	},
+};
+
+static struct platform_device dec_zs_device[] = {
+	{
+		.name = "zs",
+		.id = 0,
+		.resource = dec_zs_resources[0],
+		.num_resources = ARRAY_SIZE(dec_zs_resources[0]),
+	},
+	{
+		.name = "zs",
+		.id = 1,
+		.resource = dec_zs_resources[1],
+		.num_resources = ARRAY_SIZE(dec_zs_resources[1]),
+	},
+};
+
 static int __init dec_add_devices(void)
 {
-	int ret1, ret2;
-	int num_dz;
+	struct platform_device *dec_zs_devices[ARRAY_SIZE(dec_zs_device)];
+	int ret1, ret2, ret3;
+	int num_dz, num_zs;
 	int irq, i;
 
 	dec_rtc_resources[0].start = RTC_PORT(0);
@@ -84,10 +112,36 @@ static int __init dec_add_devices(void)
 	}
 	num_dz = i;
 
+	i = 0;
+	irq = dec_interrupt[DEC_IRQ_SCC0];
+	if (irq >= 0) {
+		resource_size_t base = dec_kn_slot_base + IOASIC_SCC0;
+
+		dec_zs_device[i].resource[0].start = base;
+		dec_zs_device[i].resource[0].end = base + dec_kn_slot_size - 1;
+		dec_zs_device[i].resource[1].start = irq;
+		dec_zs_device[i].resource[1].end = irq;
+		dec_zs_devices[i] = &dec_zs_device[i];
+		i++;
+	}
+	irq = dec_interrupt[DEC_IRQ_SCC1];
+	if (irq >= 0) {
+		resource_size_t base = dec_kn_slot_base + IOASIC_SCC1;
+
+		dec_zs_device[i].resource[0].start = base;
+		dec_zs_device[i].resource[0].end = base + dec_kn_slot_size - 1;
+		dec_zs_device[i].resource[1].start = irq;
+		dec_zs_device[i].resource[1].end = irq;
+		dec_zs_devices[i] = &dec_zs_device[i];
+		i++;
+	}
+	num_zs = i;
+
 	ret1 = platform_device_register(&dec_rtc_device);
 	ret2 = IS_ENABLED(CONFIG_32BIT) ?
 	       platform_add_devices(dec_dz_devices, num_dz) : 0;
-	return ret1 ? ret1 : ret2;
+	ret3 = platform_add_devices(dec_zs_devices, num_zs);
+	return ret1 ? ret1 : ret2 ? ret2 : ret3;
 }
 
 device_initcall(dec_add_devices);
Index: linux-macro/drivers/tty/serial/zs.c
===================================================================
--- linux-macro.orig/drivers/tty/serial/zs.c
+++ linux-macro/drivers/tty/serial/zs.c
@@ -56,6 +56,7 @@
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/major.h>
+#include <linux/platform_device.h>
 #include <linux/serial.h>
 #include <linux/serial_core.h>
 #include <linux/spinlock.h>
@@ -66,10 +67,6 @@
 
 #include <linux/atomic.h>
 
-#include <asm/dec/interrupts.h>
-#include <asm/dec/ioasic_addrs.h>
-#include <asm/dec/system.h>
-
 #include "zs.h"
 
 
@@ -79,7 +76,7 @@ MODULE_LICENSE("GPL");
 
 
 static char zs_name[] __initdata = "DECstation Z85C30 serial driver version ";
-static char zs_version[] __initdata = "0.10";
+static char zs_version[] __initdata = "0.11";
 
 /*
  * It would be nice to dynamically allocate everything that
@@ -98,12 +95,8 @@ static char zs_version[] __initdata = "0
 
 #define to_zport(uport) container_of(uport, struct zs_port, port)
 
-struct zs_parms {
-	resource_size_t scc[ZS_NUM_SCCS];
-	int irq[ZS_NUM_SCCS];
-};
-
 static struct zs_scc zs_sccs[ZS_NUM_SCCS];
+static struct uart_driver zs_reg;
 
 /*
  * Set parameters in WR5, WR12, WR13 such as not to interfere
@@ -839,16 +832,15 @@ static void zs_reset(struct zs_port *zpo
 
 	spin_lock_irqsave(&scc->zlock, flags);
 	irq = !irqs_disabled_flags(flags);
-	if (!zport->initialised) {
-		/* Reset the pointer first, just in case...  */
-		read_zsreg(zport, R0);
-		/* And let the current transmission finish.  */
-		zs_line_drain(zport, irq);
-		write_zsreg(zport, R9, zport == zport_a ? CHRA : CHRB);
-		udelay(10);
-		write_zsreg(zport, R9, 0);
-		zport->initialised = 1;
-	}
+
+	/* Reset the pointer first, just in case...  */
+	read_zsreg(zport, R0);
+	/* And let the current transmission finish.  */
+	zs_line_drain(zport, irq);
+	write_zsreg(zport, R9, zport == zport_a ? CHRA : CHRB);
+	udelay(10);
+	write_zsreg(zport, R9, 0);
+
 	load_zsregs(zport, zport->regs, irq);
 	spin_unlock_irqrestore(&scc->zlock, flags);
 }
@@ -1055,63 +1047,62 @@ static const struct uart_ops zs_ops = {
 /*
  * Initialize Z85C30 port structures.
  */
-static int __init zs_probe_sccs(void)
+static int __init zs_probe(struct platform_device *pdev)
 {
-	static int probed;
-	struct zs_parms zs_parms;
-	int chip, side, irq;
-	int n_chips = 0;
+	struct resource *mem_resource, *irq_resource;
+	int chip, side;
 	int i;
 
-	if (probed)
-		return 0;
-
-	irq = dec_interrupt[DEC_IRQ_SCC0];
-	if (irq >= 0) {
-		zs_parms.scc[n_chips] = IOASIC_SCC0;
-		zs_parms.irq[n_chips] = dec_interrupt[DEC_IRQ_SCC0];
-		n_chips++;
-	}
-	irq = dec_interrupt[DEC_IRQ_SCC1];
-	if (irq >= 0) {
-		zs_parms.scc[n_chips] = IOASIC_SCC1;
-		zs_parms.irq[n_chips] = dec_interrupt[DEC_IRQ_SCC1];
-		n_chips++;
-	}
-	if (!n_chips)
-		return -ENXIO;
+	mem_resource = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq_resource = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
+	if (!mem_resource || !irq_resource)
+		return -ENODEV;
 
-	probed = 1;
+	chip = pdev->id;
+	spin_lock_init(&zs_sccs[chip].zlock);
+	for (side = 0; side < ZS_NUM_CHAN; side++) {
+		struct zs_port *zport = &zs_sccs[chip].zport[side];
+		struct uart_port *uport = &zport->port;
 
-	for (chip = 0; chip < n_chips; chip++) {
-		spin_lock_init(&zs_sccs[chip].zlock);
-		for (side = 0; side < ZS_NUM_CHAN; side++) {
-			struct zs_port *zport = &zs_sccs[chip].zport[side];
-			struct uart_port *uport = &zport->port;
+		zport->scc	= &zs_sccs[chip];
+		zport->clk_mode	= 16;
 
-			zport->scc	= &zs_sccs[chip];
-			zport->clk_mode	= 16;
+		uport->dev	= &pdev->dev;
+		uport->has_sysrq = IS_ENABLED(CONFIG_SERIAL_ZS_CONSOLE);
+		uport->irq	= irq_resource->start;
+		uport->uartclk	= ZS_CLOCK;
+		uport->fifosize	= 1;
+		uport->iotype	= UPIO_MEM;
+		uport->flags	= UPF_BOOT_AUTOCONF;
+		uport->ops	= &zs_ops;
+		uport->line	= chip * ZS_NUM_CHAN + side;
+		uport->mapbase	= mem_resource->start +
+				  (side ^ ZS_CHAN_B) * ZS_CHAN_IO_SIZE;
 
-			uport->has_sysrq = IS_ENABLED(CONFIG_SERIAL_ZS_CONSOLE);
-			uport->irq	= zs_parms.irq[chip];
-			uport->uartclk	= ZS_CLOCK;
-			uport->fifosize	= 1;
-			uport->iotype	= UPIO_MEM;
-			uport->flags	= UPF_BOOT_AUTOCONF;
-			uport->ops	= &zs_ops;
-			uport->line	= chip * ZS_NUM_CHAN + side;
-			uport->mapbase	= dec_kn_slot_base +
-					  zs_parms.scc[chip] +
-					  (side ^ ZS_CHAN_B) * ZS_CHAN_IO_SIZE;
+		for (i = 0; i < ZS_NUM_REGS; i++)
+			zport->regs[i] = zs_init_regs[i];
 
-			for (i = 0; i < ZS_NUM_REGS; i++)
-				zport->regs[i] = zs_init_regs[i];
-		}
+		if (uart_add_one_port(&zs_reg, uport))
+			uport->dev = NULL;
 	}
 
 	return 0;
 }
 
+static void __exit zs_remove(struct platform_device *pdev)
+{
+	int chip, side;
+
+	chip = pdev->id;
+	for (side = ZS_NUM_CHAN - 1; side >= 0; side--) {
+		struct zs_port *zport = &zs_sccs[chip].zport[side];
+		struct uart_port *uport = &zport->port;
+
+		if (uport->dev)
+			uart_remove_one_port(&zs_reg, uport);
+	}
+}
+
 
 #ifdef CONFIG_SERIAL_ZS_CONSOLE
 static void zs_console_putchar(struct uart_port *uport, unsigned char ch)
@@ -1192,20 +1183,14 @@ static int __init zs_console_setup(struc
 	int bits = 8;
 	int parity = 'n';
 	int flow = 'n';
-	int ret;
-
-	ret = zs_map_port(uport);
-	if (ret)
-		return ret;
-
-	zs_reset(zport);
 
+	if (!zport->scc)
+		return -ENODEV;
 	if (options)
 		uart_parse_options(options, &baud, &parity, &bits, &flow);
 	return uart_set_options(uport, co, baud, parity, bits, flow);
 }
 
-static struct uart_driver zs_reg;
 static struct console zs_console = {
 	.name	= "ttyS",
 	.write	= zs_console_write,
@@ -1216,23 +1201,6 @@ static struct console zs_console = {
 	.data	= &zs_reg,
 };
 
-/*
- *	Register console.
- */
-static int __init zs_serial_console_init(void)
-{
-	int ret;
-
-	ret = zs_probe_sccs();
-	if (ret)
-		return ret;
-	register_console(&zs_console);
-
-	return 0;
-}
-
-console_initcall(zs_serial_console_init);
-
 #define SERIAL_ZS_CONSOLE	&zs_console
 #else
 #define SERIAL_ZS_CONSOLE	NULL
@@ -1248,47 +1216,31 @@ static struct uart_driver zs_reg = {
 	.cons			= SERIAL_ZS_CONSOLE,
 };
 
+static struct platform_driver zs_driver = {
+	.remove = __exit_p(zs_remove),
+	.driver = { .name = "zs" },
+};
+
 /* zs_init inits the driver. */
 static int __init zs_init(void)
 {
-	int i, ret;
+	int ret;
 
 	pr_info("%s%s\n", zs_name, zs_version);
 
-	/* Find out how many Z85C30 SCCs we have.  */
-	ret = zs_probe_sccs();
-	if (ret)
-		return ret;
-
 	ret = uart_register_driver(&zs_reg);
 	if (ret)
 		return ret;
+	ret = platform_driver_probe(&zs_driver, zs_probe);
+	if (ret)
+		uart_unregister_driver(&zs_reg);
 
-	for (i = 0; i < ZS_NUM_SCCS * ZS_NUM_CHAN; i++) {
-		struct zs_scc *scc = &zs_sccs[i / ZS_NUM_CHAN];
-		struct zs_port *zport = &scc->zport[i % ZS_NUM_CHAN];
-		struct uart_port *uport = &zport->port;
-
-		if (zport->scc)
-			uart_add_one_port(&zs_reg, uport);
-	}
-
-	return 0;
+	return ret;
 }
 
 static void __exit zs_exit(void)
 {
-	int i;
-
-	for (i = ZS_NUM_SCCS * ZS_NUM_CHAN - 1; i >= 0; i--) {
-		struct zs_scc *scc = &zs_sccs[i / ZS_NUM_CHAN];
-		struct zs_port *zport = &scc->zport[i % ZS_NUM_CHAN];
-		struct uart_port *uport = &zport->port;
-
-		if (zport->scc)
-			uart_remove_one_port(&zs_reg, uport);
-	}
-
+	platform_driver_unregister(&zs_driver);
 	uart_unregister_driver(&zs_reg);
 }
 
Index: linux-macro/drivers/tty/serial/zs.h
===================================================================
--- linux-macro.orig/drivers/tty/serial/zs.h
+++ linux-macro/drivers/tty/serial/zs.h
@@ -22,7 +22,6 @@
 struct zs_port {
 	struct zs_scc	*scc;			/* Containing SCC.  */
 	struct uart_port port;			/* Underlying UART.  */
-	int		initialised;		/* For the console port.  */
 
 	int		clk_mode;		/* May be 1, 16, 32, or 64.  */
 

[PATCH v3 09/10] serial: dz: Enable modular build

[Maciej W. Rozycki]

Enable modular build since the driver now has a proper module_exit() 
handler.  There's nothing specific to DZ hardware to prevent driver 
unloading and reloading from working.

Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
---
No change from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605012103360.11074@angie.orcam.me.uk/>.

No change from v1 (7/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604130133470.29980@angie.orcam.me.uk/>.
---
 drivers/tty/serial/Kconfig |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

linux-serial-dz-module.diff
Index: linux-macro/drivers/tty/serial/Kconfig
===================================================================
--- linux-macro.orig/drivers/tty/serial/Kconfig
+++ linux-macro/drivers/tty/serial/Kconfig
@@ -335,7 +335,7 @@ config SERIAL_MAX310X
 	  Say Y here if you want to support this ICs.
 
 config SERIAL_DZ
-	bool "DECstation DZ serial driver"
+	tristate "DECstation DZ serial driver"
 	depends on MACH_DECSTATION && 32BIT
 	select SERIAL_CORE
 	default y

[PATCH v3 10/10] MIPS: DEC: Ensure RTC platform device deregistration upon failure

[Maciej W. Rozycki]

Switch RTC platform device registration from platform_device_register() 
to platform_add_devices() so as to make sure any failure will result in 
automatic device unregistration.

Fixes: fae67ad43114 ("arch/mips/dec: switch DECstation systems to rtc-cmos")
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
---
Change from v2,
<https://lore.kernel.org/r/alpine.DEB.2.21.2605012105320.11074@angie.orcam.me.uk/>:

- Fix a minor style issue in the commit description.

No change from v1 (8/8),
<https://lore.kernel.org/r/alpine.DEB.2.21.2604110042130.29980@angie.orcam.me.uk/>.
---
 arch/mips/dec/platform.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

linux-mips-dec-platform-rtc-unregister.diff
Index: linux-macro/arch/mips/dec/platform.c
===================================================================
--- linux-macro.orig/arch/mips/dec/platform.c
+++ linux-macro/arch/mips/dec/platform.c
@@ -38,6 +38,10 @@ static struct platform_device dec_rtc_de
 	.num_resources = ARRAY_SIZE(dec_rtc_resources),
 };
 
+static struct platform_device *dec_rtc_devices[] __initdata = {
+	&dec_rtc_device,
+};
+
 static struct resource dec_dz_resources[] = {
 	{ .name = "dz", .flags = IORESOURCE_MEM, },
 	{ .name = "dz", .flags = IORESOURCE_IRQ, },
@@ -137,7 +141,7 @@ static int __init dec_add_devices(void)
 	}
 	num_zs = i;
 
-	ret1 = platform_device_register(&dec_rtc_device);
+	ret1 = platform_add_devices(dec_rtc_devices, 1);
 	ret2 = IS_ENABLED(CONFIG_32BIT) ?
 	       platform_add_devices(dec_dz_devices, num_dz) : 0;
 	ret3 = platform_add_devices(dec_zs_devices, num_zs);

Re: [PATCH v3 00/10] MIPS: DEC: Fix serial device regressions + RTC cleanup

[Maciej W. Rozycki]

Hi Thomas,

On Wed, 6 May 2026, Maciej W. Rozycki wrote:

>  Two extra introductory changes, 01/10 and 02/10 have now been added to 
> platform code.  No modification has been made to original changes.  The 
> original description follows, updated for patch renumbering.

 Will you be able to get these two changes merged sometime soon now that 
Greg has queued changes 03/10 through 09/10, so that these prerequisites 
land ahead and there's no 64-bit regression in mainline?

 Let me know if you need anything from me here before I disappear for a 
couple weeks in a fortnight's time.

  Maciej

Re: [PATCH v3 01/10] MIPS: DEC: Ensure 32-bit stack location for o32 prom_printf()

[Thomas Bogendoerfer]

On Wed, May 06, 2026 at 11:42:23PM +0100, Maciej W. Rozycki wrote:
> In 64-bit configurations calling any firmware entry points from a kernel 
> thread other than the initial one will result in a situation where the 
> stack has been placed in the XKPHYS 64-bit memory segment.
> 
> Consequently the stack pointer is no longer a 32-bit value and when the 
> 32-bit firmware code called uses 32-bit ALU operations to manipulate the 
> stack pointer, the calculated result is incorrect (in fact in the 64-bit 
> MIPS ISA almost all 32-bit ALU operations will produce an unpredictable 
> result when executed on 64-bit data) and control goes astray.
> 
> This may happen when no final console driver has been enabled in the
> configuration and consequently the initial console continues being used
> late into bootstrap, or with an upcoming change that will switch the zs
> driver to use a platform device, which in turn will make the console
> handover happen only after other kernel threads have already been
> started, and the kernel will hang at:
> 
>   pid_max: default: 32768 minimum: 301
> 
> or somewhat later, but always before:
> 
>   cblist_init_generic: Setting adjustable number of callback queues.
> 
> has been printed.
> 
> It seems that only the prom_printf() entry point is affected.  Of all 
> the other entry points wired only rex_slot_address() and rex_gettcinfo() 
> are called from a kernel thread other than the initial one, specifically 
> kernel_init(), and they are leaf functions that do no business with the 
> stack, having worked with no issue ever since 64-bit support was added 
> for the platform back in 2002.
> 
> To address this issue then, arrange for the stack to be switched in the 
> o32 wrapper as required for prom_printf() only, by supplying call_o32() 
> with a pointer to a chunk of initdata space, which is placed in the 
> CKSEG0 32-bit compatibility segment, observing that prom_printf() is 
> only called from console output handler and therefore with the console 
> lock held, implying no need for this code to be reentrant.
> 
> Other firmware entry points may be called with interrupts enabled and no 
> lock held, and may therefore require that call_o32() be reentrant.  They 
> trigger no issue at this point and "if it ain't broke, don't fix it," so 
> just leave them alone.
> 
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
> Cc: stable@vger.kernel.org # v2.6.12+
> ---
> No change from v2, 
> <https://lore.kernel.org/r/alpine.DEB.2.21.2605010132340.38805@angie.orcam.me.uk/>.
> 
> New change in v2.
> ---
>  arch/mips/dec/prom/init.c        |    6 +++++-
>  arch/mips/include/asm/dec/prom.h |   15 +++++++++++++--
>  2 files changed, 18 insertions(+), 3 deletions(-)

applied to mips-next

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]

Re: [PATCH v3 02/10] MIPS: DEC: Prevent initial console buffer from landing in XKPHYS

[Thomas Bogendoerfer]

On Wed, May 06, 2026 at 11:42:27PM +0100, Maciej W. Rozycki wrote:
> In 64-bit configurations calling the initial console output handler from 
> a kernel thread other than the initial one will result in a situation 
> where the stack has been placed in the XKPHYS 64-bit memory segment and 
> consequently so has been the buffer allocated there that is used as the 
> argument corresponding to the `%s' output conversion specifier for the 
> firmware's printf() entry point.
> 
> This 64-bit address will then be truncated by 32-bit firmware, resulting 
> in an attempt to access the wrong memory location, which in turn will 
> cause all kinds of unpredictable behaviour, such as a kernel crash:
> 
>   Console: colour dummy device 160x64
>   Calibrating delay loop... 49.36 BogoMIPS (lpj=192512)
>   pid_max: default: 32768 minimum: 301
>   CPU 0 Unable to handle kernel paging request at virtual address 000000000203bd00, epc == ffffffffbfc08364, ra == ffffffffbfc08800
>   Oops[#1]:
>   CPU: 0 PID: 0 Comm: swapper Not tainted 5.18.0-rc2-00254-gfb649bda6f56-dirty #121
>   $ 0   : 0000000000000000 0000000000000001 0000000000000023 ffffffff80684ba0
>   $ 4   : 000000000203bd00 ffffffffbfc0f3b4 ffffffffffffffff 0000000000000073
>   $ 8   : 0a303d7469000000 0000000000000000 0000000000000073 ffffffffbfc0f473
>   $12   : 0000000000000002 0000000000000000 ffffffff80684c1c 0000000000000000
>   $16   : 0000000000000000 ffffffff80596dc9 0000000000000000 ffffffffbfc09240
>   $20   : ffffffff80684c40 ffffffffbfc0f400 000000000000002d 000000000000002b
>   $24   : ffffffffffffffbf 000000000203bd00                                  
>   $28   : ffffffff805f0000 ffffffff80684b58 0000000000000030 ffffffffbfc08800
>   Hi    : 0000000000000000
>   Lo    : 0000000000000aa8
>   epc   : ffffffffbfc08364 0xffffffffbfc08364
>   ra    : ffffffffbfc08800 0xffffffffbfc08800
>   Status: 140120e2        KX SX UX KERNEL EXL 
>   Cause : 00000008 (ExcCode 02)
>   BadVA : 000000000203bd00
>   PrId  : 00000430 (R4000SC)
>   Modules linked in:
>   Process swapper (pid: 0, threadinfo=(____ptrval____), task=(____ptrval____), tls=0000000000000000)
>   Stack : 0000000000000000 0000000000000000 0000000000000000 0000004d0000004d
>           80684cc0806a2a40 80596dc80000004d 8061000000000000 bfc0850c80684c38
>           0000000000000000 000000000203bd00 0000000000000000 0000000000000000
>           0000000000000000 00000000bfc0f3b4 0000000000000000 0000000000000000
>           0000000000000000 0000000000000000 0000000000000000 0000000000000000
>           0000000000000000 0000000000000000 0000000000000000 0000000000000000
>           0000002500000000 0000000000000000 0000000000000000 802c1a7400000000
>           0203bd0080596dc8 0203bd4d69000000 6c61632000000018 5f746567646e6172
>           6c616320625f6d6f 5f736e5f6d6f7266 206361323778302b 303d74696e726320
>           806a0a38806b0000 806a0a38806b0000 00000000806b0000 80683c58806b0000
>           ...
>   Call Trace:
>   
>   
>   Code: a082ffff  03e00008  00601021 <80820000> 00001821  10400005  24840001  80820000  24630001 
>   
>   ---[ end trace 0000000000000000 ]---
>   Kernel panic - not syncing: Fatal exception in interrupt
>   
>   KN04 V2.1k    (PC: 0xa0026768, SP: 0x806848e8)
>   >>
> 
> In this case the pointer in $4 was truncated from 0x980000000203bd00 to 
> 0x000000000203bd00.
> 
> This may happen when no final console driver has been enabled in the 
> configuration and consequently the initial console continues being used 
> late into bootstrap or with an upcoming change that will switch the zs 
> driver to use a platform device, which in turn will make the console 
> handover happen only after other kernel threads have already been 
> started.
> 
> Fix the issue by making the buffer static and initdata, and therefore 
> placed in the CKSEG0 32-bit compatibility segment, observing that the 
> console output handler is called with the console lock held, implying 
> no need for this code to be reentrant.  Add an assertion to verify the 
> buffer actually has been placed in a compatibility segment.
> 
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
> Cc: stable@vger.kernel.org # v2.6.12+
> ---
> Changes from v2, 
> <https://lore.kernel.org/r/alpine.DEB.2.21.2605010059310.38805@angie.orcam.me.uk/>:
> 
> - Cast `buf' to `long' before casting to `int', fixing:
> 
>   In file included from ./include/linux/bug.h:5,
>                    from arch/mips/dec/prom/console.c:7:
>   arch/mips/dec/prom/console.c: In function 'prom_console_write':
>   arch/mips/dec/prom/console.c:21:29: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
>      21 |         BUG_ON((long)buf != (int)buf);
>         |                             ^
>   ./arch/mips/include/asm/bug.h:34:44: note: in definition of macro 'BUG_ON'
>      34 | #define BUG_ON(C) __BUG_ON((unsigned long)(C))
>         |                                            ^
> 
> New change in v2.
> ---
>  arch/mips/dec/prom/console.c |    7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)

applied to mips-next

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]

Re: [PATCH v3 00/10] MIPS: DEC: Fix serial device regressions + RTC cleanup

[Thomas Bogendoerfer]

On Fri, May 22, 2026 at 04:15:06PM +0100, Maciej W. Rozycki wrote:
> Hi Thomas,
> 
> On Wed, 6 May 2026, Maciej W. Rozycki wrote:
> 
> >  Two extra introductory changes, 01/10 and 02/10 have now been added to 
> > platform code.  No modification has been made to original changes.  The 
> > original description follows, updated for patch renumbering.
> 
>  Will you be able to get these two changes merged sometime soon now that 
> Greg has queued changes 03/10 through 09/10, so that these prerequisites 
> land ahead and there's no 64-bit regression in mainline?

I've applied 01/10 and 02/10. 10/10 doesn't apply to my tree, because it
depends on the platform.c changes from dz/zs driver change not yet present
in my tree. So either Greg takes it via his tree or I'll pick it up in
a second pull request during the merge window

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]

Re: [PATCH v3 10/10] MIPS: DEC: Ensure RTC platform device deregistration upon failure

[Thomas Bogendoerfer]

On Wed, May 06, 2026 at 11:43:00PM +0100, Maciej W. Rozycki wrote:
> Switch RTC platform device registration from platform_device_register() 
> to platform_add_devices() so as to make sure any failure will result in 
> automatic device unregistration.
> 
> Fixes: fae67ad43114 ("arch/mips/dec: switch DECstation systems to rtc-cmos")
> Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
> ---
> Change from v2,
> <https://lore.kernel.org/r/alpine.DEB.2.21.2605012105320.11074@angie.orcam.me.uk/>:
> 
> - Fix a minor style issue in the commit description.
> 
> No change from v1 (8/8),
> <https://lore.kernel.org/r/alpine.DEB.2.21.2604110042130.29980@angie.orcam.me.uk/>.
> ---
>  arch/mips/dec/platform.c |    6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> linux-mips-dec-platform-rtc-unregister.diff
> Index: linux-macro/arch/mips/dec/platform.c
> ===================================================================
> --- linux-macro.orig/arch/mips/dec/platform.c
> +++ linux-macro/arch/mips/dec/platform.c
> @@ -38,6 +38,10 @@ static struct platform_device dec_rtc_de
>  	.num_resources = ARRAY_SIZE(dec_rtc_resources),
>  };
>  
> +static struct platform_device *dec_rtc_devices[] __initdata = {
> +	&dec_rtc_device,
> +};
> +
>  static struct resource dec_dz_resources[] = {
>  	{ .name = "dz", .flags = IORESOURCE_MEM, },
>  	{ .name = "dz", .flags = IORESOURCE_IRQ, },
> @@ -137,7 +141,7 @@ static int __init dec_add_devices(void)
>  	}
>  	num_zs = i;
>  
> -	ret1 = platform_device_register(&dec_rtc_device);
> +	ret1 = platform_add_devices(dec_rtc_devices, 1);
>  	ret2 = IS_ENABLED(CONFIG_32BIT) ?
>  	       platform_add_devices(dec_dz_devices, num_dz) : 0;
>  	ret3 = platform_add_devices(dec_zs_devices, num_zs);

Acked-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>

Greg,

since this only applies with the DZ/ZS changes, could you take this patch
via the serial tree ?

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]

Re: [PATCH v3 00/10] MIPS: DEC: Fix serial device regressions + RTC cleanup

[Maciej W. Rozycki]

On Tue, 26 May 2026, Thomas Bogendoerfer wrote:

> >  Will you be able to get these two changes merged sometime soon now that 
> > Greg has queued changes 03/10 through 09/10, so that these prerequisites 
> > land ahead and there's no 64-bit regression in mainline?
> 
> I've applied 01/10 and 02/10. 10/10 doesn't apply to my tree, because it
> depends on the platform.c changes from dz/zs driver change not yet present
> in my tree. So either Greg takes it via his tree or I'll pick it up in
> a second pull request during the merge window

 Fine with me either way, thanks for taking care of this stuff!  I'm glad 
I was able to find time and momentum to get these issues finally sorted, 
as I was aware of them for quite a while now, but there's so much to pick 
from.  More to come.

  Maciej

Re: [PATCH v3 10/10] MIPS: DEC: Ensure RTC platform device deregistration upon failure

[Thomas Bogendoerfer]

On Wed, May 06, 2026 at 11:43:00PM +0100, Maciej W. Rozycki wrote:
> Switch RTC platform device registration from platform_device_register() 
> to platform_add_devices() so as to make sure any failure will result in 
> automatic device unregistration.
> 
> Fixes: fae67ad43114 ("arch/mips/dec: switch DECstation systems to rtc-cmos")
> Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
> ---
> Change from v2,
> <https://lore.kernel.org/r/alpine.DEB.2.21.2605012105320.11074@angie.orcam.me.uk/>:
> 
> - Fix a minor style issue in the commit description.
> 
> No change from v1 (8/8),
> <https://lore.kernel.org/r/alpine.DEB.2.21.2604110042130.29980@angie.orcam.me.uk/>.
> ---
>  arch/mips/dec/platform.c |    6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> linux-mips-dec-platform-rtc-unregister.diff
> Index: linux-macro/arch/mips/dec/platform.c
> ===================================================================
> --- linux-macro.orig/arch/mips/dec/platform.c
> +++ linux-macro/arch/mips/dec/platform.c
> @@ -38,6 +38,10 @@ static struct platform_device dec_rtc_de
>  	.num_resources = ARRAY_SIZE(dec_rtc_resources),
>  };
>  
> +static struct platform_device *dec_rtc_devices[] __initdata = {
> +	&dec_rtc_device,
> +};
> +
>  static struct resource dec_dz_resources[] = {
>  	{ .name = "dz", .flags = IORESOURCE_MEM, },
>  	{ .name = "dz", .flags = IORESOURCE_IRQ, },
> @@ -137,7 +141,7 @@ static int __init dec_add_devices(void)
>  	}
>  	num_zs = i;
>  
> -	ret1 = platform_device_register(&dec_rtc_device);
> +	ret1 = platform_add_devices(dec_rtc_devices, 1);
>  	ret2 = IS_ENABLED(CONFIG_32BIT) ?
>  	       platform_add_devices(dec_dz_devices, num_dz) : 0;
>  	ret3 = platform_add_devices(dec_zs_devices, num_zs);

applied to mips-fixes

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]