Re: [Regression] depmod fails on kernel 6.17.1 rc1

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Borislav Petkov]

Lemme add a whole bunch of relevant ppl to Cc.

So I'm seeing this with an allmodconfig build too:

depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
depmod: ERROR: Found 2 modules in dependency cycles!
make[6]: *** [scripts/Makefile.modinst:132: depmod] Error 1
make[5]: *** [Makefile:1917: modules_install] Error 2
make[4]: *** [Makefile:2140: run-command] Error 2
make[3]: *** [debian/rules:66: binary-image] Error 2
dpkg-buildpackage: error: make -f debian/rules binary subprocess returned exit status 2
make[2]: *** [scripts/Makefile.package:126: bindeb-pkg] Error 2
make[1]: *** [/home/amd/kernel/linux/Makefile:1643: bindeb-pkg] Error 2
make: *** [Makefile:248: __sub-make] Error 2

That's 6.17.7

On Sun, Oct 05, 2025 at 09:40:28PM -0400, Alex Davis wrote:
> When running make modules_install on linux 6.17. with the attached
> .config, I get the following error:
> 
>   DEPMOD  /lib/modules/6.17.1-rc1
> depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> depmod: ERROR: Cycle detected: serial_mctrl_gpio
> depmod: ERROR: Found 2 modules in dependency cycles!
> make[2]: *** [scripts/Makefile.modinst:132: depmod] Error 1
> make[1]: *** [/spare/linux/linux-6.17/Makefile:1917: modules_install] Error 2
> make: *** [Makefile:248: __sub-make] Error 2
> 
> 
> This also happens with 6.17; it does not happen with 6.16.x.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Andy Shevchenko]

On Thu, Nov 06, 2025 at 05:02:35PM +0100, Borislav Petkov wrote:
> Lemme add a whole bunch of relevant ppl to Cc.

Thanks!

> So I'm seeing this with an allmodconfig build too:
> 
> depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> depmod: ERROR: Found 2 modules in dependency cycles!

I'm surprised it took so long to people to start complaining.

So, some of those are modules? Can you share the

	grep 8250 .config

part?

> make[6]: *** [scripts/Makefile.modinst:132: depmod] Error 1
> make[5]: *** [Makefile:1917: modules_install] Error 2
> make[4]: *** [Makefile:2140: run-command] Error 2
> make[3]: *** [debian/rules:66: binary-image] Error 2
> dpkg-buildpackage: error: make -f debian/rules binary subprocess returned exit status 2
> make[2]: *** [scripts/Makefile.package:126: bindeb-pkg] Error 2
> make[1]: *** [/home/amd/kernel/linux/Makefile:1643: bindeb-pkg] Error 2
> make: *** [Makefile:248: __sub-make] Error 2
> 
> That's 6.17.7
> 
> On Sun, Oct 05, 2025 at 09:40:28PM -0400, Alex Davis wrote:
> > When running make modules_install on linux 6.17. with the attached
> > .config, I get the following error:
> > 
> >   DEPMOD  /lib/modules/6.17.1-rc1
> > depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > depmod: ERROR: Cycle detected: serial_mctrl_gpio
> > depmod: ERROR: Found 2 modules in dependency cycles!
> > make[2]: *** [scripts/Makefile.modinst:132: depmod] Error 1
> > make[1]: *** [/spare/linux/linux-6.17/Makefile:1917: modules_install] Error 2
> > make: *** [Makefile:248: __sub-make] Error 2
> > 
> > This also happens with 6.17; it does not happen with 6.16.x.

-- 
With Best Regards,
Andy Shevchenko

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Borislav Petkov]

On Thu, Nov 06, 2025 at 06:15:57PM +0200, Andy Shevchenko wrote:
> > So I'm seeing this with an allmodconfig build too:
			       ^^^^^^^^^^^^

> > depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > depmod: ERROR: Found 2 modules in dependency cycles!
> 
> I'm surprised it took so long to people to start complaining.
> 
> So, some of those are modules? Can you share the
> 
> 	grep 8250 .config
> 
> part?

See above.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Ilpo Järvinen]

On Thu, 6 Nov 2025, Borislav Petkov wrote:

> On Thu, Nov 06, 2025 at 06:15:57PM +0200, Andy Shevchenko wrote:
> > > So I'm seeing this with an allmodconfig build too:
> 			       ^^^^^^^^^^^^
> 
> > > depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > > depmod: ERROR: Found 2 modules in dependency cycles!
> > 
> > I'm surprised it took so long to people to start complaining.
> > 
> > So, some of those are modules? Can you share the
> > 
> > 	grep 8250 .config
> > 
> > part?
> 
> See above.

https://lore.kernel.org/all/87frc3sd8d.fsf@posteo.net/

I wonder if 8250_rsa.o can be put into 8250_base.o where most of its 
callers are anyway?

-- 
 i.

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 6890 bytes --]

On Thu, 6 Nov 2025, Ilpo Järvinen wrote:

> On Thu, 6 Nov 2025, Borislav Petkov wrote:
> 
> > On Thu, Nov 06, 2025 at 06:15:57PM +0200, Andy Shevchenko wrote:
> > > > So I'm seeing this with an allmodconfig build too:
> > 			       ^^^^^^^^^^^^
> > > > depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > > > depmod: ERROR: Found 2 modules in dependency cycles!
> > > 
> > > I'm surprised it took so long to people to start complaining.
> > > 
> > > So, some of those are modules? Can you share the
> > > 
> > > 	grep 8250 .config
> > > 
> > > part?
> > 
> > See above.
> 
> https://lore.kernel.org/all/87frc3sd8d.fsf@posteo.net/
> 
> I wonder if 8250_rsa.o can be put into 8250_base.o where most of its 
> callers are anyway?

This seems to resolve the build issue for me:

--
From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>
Subject: [PATCH 1/1] serial: 8250: Fix 8250_rsa symbol loop

make allmodconfig build fails due to dependency loop:

  depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
  depmod: ERROR: Found 2 modules in dependency cycles!

Break dependency loop by moving 8250_rsa.o into 8250_base and by
passing univ8250_port_base_ops to univ8250_rsa_support() that can make
a local copy of it.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reported-by: Alex Davis <alex47794@gmail.com>
Fixes: b20d6576cdb3 ("serial: 8250: export RSA functions")
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
---
 drivers/tty/serial/8250/8250.h          |  2 +-
 drivers/tty/serial/8250/8250_platform.c |  2 +-
 drivers/tty/serial/8250/8250_rsa.c      | 18 +++++++++---------
 drivers/tty/serial/8250/Makefile        |  2 +-
 4 files changed, 12 insertions(+), 12 deletions(-)

diff --git a/drivers/tty/serial/8250/8250.h b/drivers/tty/serial/8250/8250.h
index 58e64c4e1e3a..bd294fbf0e65 100644
--- a/drivers/tty/serial/8250/8250.h
+++ b/drivers/tty/serial/8250/8250.h
@@ -322,7 +322,7 @@ static inline void serial8250_pnp_exit(void) { }
 #endif
 
 #ifdef CONFIG_SERIAL_8250_RSA
-void univ8250_rsa_support(struct uart_ops *ops);
+void univ8250_rsa_support(struct uart_ops *ops, const struct uart_ops *core_ops);
 void rsa_enable(struct uart_8250_port *up);
 void rsa_disable(struct uart_8250_port *up);
 void rsa_autoconfig(struct uart_8250_port *up);
diff --git a/drivers/tty/serial/8250/8250_platform.c b/drivers/tty/serial/8250/8250_platform.c
index b27981340e76..fe7ec440ffa5 100644
--- a/drivers/tty/serial/8250/8250_platform.c
+++ b/drivers/tty/serial/8250/8250_platform.c
@@ -75,7 +75,7 @@ static void __init __serial8250_isa_init_ports(void)
 
 	/* chain base port ops to support Remote Supervisor Adapter */
 	univ8250_port_ops = *univ8250_port_base_ops;
-	univ8250_rsa_support(&univ8250_port_ops);
+	univ8250_rsa_support(&univ8250_port_ops, univ8250_port_base_ops);
 
 	if (share_irqs)
 		irqflag = IRQF_SHARED;
diff --git a/drivers/tty/serial/8250/8250_rsa.c b/drivers/tty/serial/8250/8250_rsa.c
index 40a3dbd9e452..d76d1be86625 100644
--- a/drivers/tty/serial/8250/8250_rsa.c
+++ b/drivers/tty/serial/8250/8250_rsa.c
@@ -14,6 +14,8 @@
 static unsigned long probe_rsa[PORT_RSA_MAX];
 static unsigned int probe_rsa_count;
 
+static const struct uart_ops *core_port_base_ops;
+
 static int rsa8250_request_resource(struct uart_8250_port *up)
 {
 	struct uart_port *port = &up->port;
@@ -67,7 +69,7 @@ static void univ8250_config_port(struct uart_port *port, int flags)
 		}
 	}
 
-	univ8250_port_base_ops->config_port(port, flags);
+	core_port_base_ops->config_port(port, flags);
 
 	if (port->type != PORT_RSA && up->probe & UART_PROBE_RSA)
 		rsa8250_release_resource(up);
@@ -78,11 +80,11 @@ static int univ8250_request_port(struct uart_port *port)
 	struct uart_8250_port *up = up_to_u8250p(port);
 	int ret;
 
-	ret = univ8250_port_base_ops->request_port(port);
+	ret = core_port_base_ops->request_port(port);
 	if (ret == 0 && port->type == PORT_RSA) {
 		ret = rsa8250_request_resource(up);
 		if (ret < 0)
-			univ8250_port_base_ops->release_port(port);
+			core_port_base_ops->release_port(port);
 	}
 
 	return ret;
@@ -94,15 +96,17 @@ static void univ8250_release_port(struct uart_port *port)
 
 	if (port->type == PORT_RSA)
 		rsa8250_release_resource(up);
-	univ8250_port_base_ops->release_port(port);
+	core_port_base_ops->release_port(port);
 }
 
-void univ8250_rsa_support(struct uart_ops *ops)
+void univ8250_rsa_support(struct uart_ops *ops, const struct uart_ops *core_ops)
 {
+	core_port_base_ops = core_ops;
 	ops->config_port  = univ8250_config_port;
 	ops->request_port = univ8250_request_port;
 	ops->release_port = univ8250_release_port;
 }
+EXPORT_SYMBOL_FOR_MODULES(univ8250_rsa_support, "8250");
 
 module_param_hw_array(probe_rsa, ulong, ioport, &probe_rsa_count, 0444);
 MODULE_PARM_DESC(probe_rsa, "Probe I/O ports for RSA");
@@ -146,7 +150,6 @@ void rsa_enable(struct uart_8250_port *up)
 	if (up->port.uartclk == SERIAL_RSA_BAUD_BASE * 16)
 		serial_out(up, UART_RSA_FRR, 0);
 }
-EXPORT_SYMBOL_FOR_MODULES(rsa_enable, "8250_base");
 
 /*
  * Attempts to turn off the RSA FIFO and resets the RSA board back to 115kbps compat mode. It is
@@ -178,7 +181,6 @@ void rsa_disable(struct uart_8250_port *up)
 	if (result)
 		up->port.uartclk = SERIAL_RSA_BAUD_BASE_LO * 16;
 }
-EXPORT_SYMBOL_FOR_MODULES(rsa_disable, "8250_base");
 
 void rsa_autoconfig(struct uart_8250_port *up)
 {
@@ -191,7 +193,6 @@ void rsa_autoconfig(struct uart_8250_port *up)
 	if (__rsa_enable(up))
 		up->port.type = PORT_RSA;
 }
-EXPORT_SYMBOL_FOR_MODULES(rsa_autoconfig, "8250_base");
 
 void rsa_reset(struct uart_8250_port *up)
 {
@@ -200,7 +201,6 @@ void rsa_reset(struct uart_8250_port *up)
 
 	serial_out(up, UART_RSA_FRR, 0);
 }
-EXPORT_SYMBOL_FOR_MODULES(rsa_reset, "8250_base");
 
 #ifdef CONFIG_SERIAL_8250_DEPRECATED_OPTIONS
 #ifndef MODULE
diff --git a/drivers/tty/serial/8250/Makefile b/drivers/tty/serial/8250/Makefile
index 513a0941c284..9ec4d5fe64de 100644
--- a/drivers/tty/serial/8250/Makefile
+++ b/drivers/tty/serial/8250/Makefile
@@ -7,7 +7,6 @@ obj-$(CONFIG_SERIAL_8250)		+= 8250.o
 8250-y					:= 8250_core.o
 8250-y					+= 8250_platform.o
 8250-$(CONFIG_SERIAL_8250_PNP)		+= 8250_pnp.o
-8250-$(CONFIG_SERIAL_8250_RSA)		+= 8250_rsa.o
 
 obj-$(CONFIG_SERIAL_8250)		+= 8250_base.o
 8250_base-y				:= 8250_port.o
@@ -15,6 +14,7 @@ obj-$(CONFIG_SERIAL_8250)		+= 8250_base.o
 8250_base-$(CONFIG_SERIAL_8250_DWLIB)	+= 8250_dwlib.o
 8250_base-$(CONFIG_SERIAL_8250_FINTEK)	+= 8250_fintek.o
 8250_base-$(CONFIG_SERIAL_8250_PCILIB)	+= 8250_pcilib.o
+8250_base-$(CONFIG_SERIAL_8250_RSA)	+= 8250_rsa.o
 
 obj-$(CONFIG_SERIAL_8250_CONSOLE)	+= 8250_early.o
 

base-commit: 719f3df3e113e03d2c8cf324827da1fd17a9bd8f
-- 
2.39.5

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Andy Shevchenko]

On Thu, Nov 06, 2025 at 08:00:36PM +0200, Ilpo Järvinen wrote:
> On Thu, 6 Nov 2025, Ilpo Järvinen wrote:
> > On Thu, 6 Nov 2025, Borislav Petkov wrote:
> > > On Thu, Nov 06, 2025 at 06:15:57PM +0200, Andy Shevchenko wrote:
> > > > > So I'm seeing this with an allmodconfig build too:
> > > 			       ^^^^^^^^^^^^
> > > > > depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > > > > depmod: ERROR: Found 2 modules in dependency cycles!
> > > > 
> > > > I'm surprised it took so long to people to start complaining.
> > > > 
> > > > So, some of those are modules? Can you share the
> > > > 
> > > > 	grep 8250 .config
> > > > 
> > > > part?
> > > 
> > > See above.
> > 
> > https://lore.kernel.org/all/87frc3sd8d.fsf@posteo.net/
> > 
> > I wonder if 8250_rsa.o can be put into 8250_base.o where most of its 
> > callers are anyway?
> 
> This seems to resolve the build issue for me:

I prefer this solution to anything that suggests to merge the 8250_rsa in
another module.

If it fixes for the reporters, feel free to add my
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

-- 
With Best Regards,
Andy Shevchenko

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Andy Shevchenko]

On Thu, Nov 06, 2025 at 08:14:35PM +0200, Andy Shevchenko wrote:
> On Thu, Nov 06, 2025 at 08:00:36PM +0200, Ilpo Järvinen wrote:
> > On Thu, 6 Nov 2025, Ilpo Järvinen wrote:
> > > On Thu, 6 Nov 2025, Borislav Petkov wrote:
> > > > On Thu, Nov 06, 2025 at 06:15:57PM +0200, Andy Shevchenko wrote:
> > > > > > So I'm seeing this with an allmodconfig build too:
> > > > 			       ^^^^^^^^^^^^
> > > > > > depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > > > > > depmod: ERROR: Found 2 modules in dependency cycles!
> > > > > 
> > > > > I'm surprised it took so long to people to start complaining.
> > > > > 
> > > > > So, some of those are modules? Can you share the
> > > > > 
> > > > > 	grep 8250 .config
> > > > > 
> > > > > part?
> > > > 
> > > > See above.
> > > 
> > > https://lore.kernel.org/all/87frc3sd8d.fsf@posteo.net/
> > > 
> > > I wonder if 8250_rsa.o can be put into 8250_base.o where most of its 
> > > callers are anyway?
> > 
> > This seems to resolve the build issue for me:
> 
> I prefer this solution to anything that suggests to merge the 8250_rsa in
> another module.

Oh, I meant "merge the source of the 8250_rsa", and here we are talking about
merging the object file, which is totally fine! Sorry for the confusion it
might made.

> If it fixes for the reporters, feel free to add my
> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

-- 
With Best Regards,
Andy Shevchenko

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 1893 bytes --]

On Fri, 7 Nov 2025, Andy Shevchenko wrote:

> On Thu, Nov 06, 2025 at 08:14:35PM +0200, Andy Shevchenko wrote:
> > On Thu, Nov 06, 2025 at 08:00:36PM +0200, Ilpo Järvinen wrote:
> > > On Thu, 6 Nov 2025, Ilpo Järvinen wrote:
> > > > On Thu, 6 Nov 2025, Borislav Petkov wrote:
> > > > > On Thu, Nov 06, 2025 at 06:15:57PM +0200, Andy Shevchenko wrote:
> > > > > > > So I'm seeing this with an allmodconfig build too:
> > > > > 			       ^^^^^^^^^^^^
> > > > > > > depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > > > > > > depmod: ERROR: Found 2 modules in dependency cycles!
> > > > > > 
> > > > > > I'm surprised it took so long to people to start complaining.
> > > > > > 
> > > > > > So, some of those are modules? Can you share the
> > > > > > 
> > > > > > 	grep 8250 .config
> > > > > > 
> > > > > > part?
> > > > > 
> > > > > See above.
> > > > 
> > > > https://lore.kernel.org/all/87frc3sd8d.fsf@posteo.net/
> > > > 
> > > > I wonder if 8250_rsa.o can be put into 8250_base.o where most of its 
> > > > callers are anyway?
> > > 
> > > This seems to resolve the build issue for me:
> > 
> > I prefer this solution to anything that suggests to merge the 8250_rsa in
> > another module.
> 
> Oh, I meant "merge the source of the 8250_rsa", and here we are talking about
> merging the object file, which is totally fine! Sorry for the confusion it
> might made.

The suggestion in that link was to merge the modules (8250.ko with 
8250_base.ko) and that certainly isn't a great idea.

My patch only moves 8250_rsa.o from 8250.ko to 8250_base.ko and breaks 
the symbol loop by passing what is needed from 8250.ko in function 
arguments instead of directly referencing them in 8250_rsa.o.

> > If it fixes for the reporters, feel free to add my
> > Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> 
> 

-- 
 i.

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Jiri Slaby]

On 06. 11. 25, 19:00, Ilpo Järvinen wrote:
> This seems to resolve the build issue for me:
> 
> --
> From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>
> Subject: [PATCH 1/1] serial: 8250: Fix 8250_rsa symbol loop
> 
> make allmodconfig build fails due to dependency loop:
> 
>    depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
>    depmod: ERROR: Found 2 modules in dependency cycles!
> 
> Break dependency loop by moving 8250_rsa.o into 8250_base and by
> passing univ8250_port_base_ops to univ8250_rsa_support() that can make
> a local copy of it.
> 
> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> Reported-by: Alex Davis <alex47794@gmail.com>
> Fixes: b20d6576cdb3 ("serial: 8250: export RSA functions")
> Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>

LGTM, thanks for the fix.

Reviewed-by: Jiri Slaby <jirislaby@kernel.org>

If the reporters could give it a shot and mark this by Tested-by, it 
would be great...

thanks,
-- 
js
suse labs

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Borislav Petkov]

On Fri, Nov 07, 2025 at 07:20:26AM +0100, Jiri Slaby wrote:
> If the reporters could give it a shot and mark this by Tested-by, it would
> be great...

I'd love to but I can't reproduce it again, for the life of me!

I've rebuilt a bunch of kernels on the same machine but nada. It simply won't
trigger again.

Damn.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1257 bytes --]

Hi Jiri,

On Fri, 7 Nov 2025 07:20:26 +0100 Jiri Slaby <jirislaby@kernel.org> wrote:
>
> On 06. 11. 25, 19:00, Ilpo Järvinen wrote:
> > This seems to resolve the build issue for me:
> > 
> > --
> > From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>
> > Subject: [PATCH 1/1] serial: 8250: Fix 8250_rsa symbol loop
> > 
> > make allmodconfig build fails due to dependency loop:
> > 
> >    depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> >    depmod: ERROR: Found 2 modules in dependency cycles!
> > 
> > Break dependency loop by moving 8250_rsa.o into 8250_base and by
> > passing univ8250_port_base_ops to univ8250_rsa_support() that can make
> > a local copy of it.
> > 
> > Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> > Reported-by: Alex Davis <alex47794@gmail.com>
> > Fixes: b20d6576cdb3 ("serial: 8250: export RSA functions")
> > Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>  
> 
> LGTM, thanks for the fix.
> 
> Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
> 
> If the reporters could give it a shot and mark this by Tested-by, it would be great...

I have not seen this for quite some time ... I assumed it had been
fixed.
-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Greg Kroah-Hartman]

On Sun, Nov 09, 2025 at 10:52:03AM +1100, Stephen Rothwell wrote:
> Hi Jiri,
> 
> On Fri, 7 Nov 2025 07:20:26 +0100 Jiri Slaby <jirislaby@kernel.org> wrote:
> >
> > On 06. 11. 25, 19:00, Ilpo Järvinen wrote:
> > > This seems to resolve the build issue for me:
> > > 
> > > --
> > > From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>
> > > Subject: [PATCH 1/1] serial: 8250: Fix 8250_rsa symbol loop
> > > 
> > > make allmodconfig build fails due to dependency loop:
> > > 
> > >    depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > >    depmod: ERROR: Found 2 modules in dependency cycles!
> > > 
> > > Break dependency loop by moving 8250_rsa.o into 8250_base and by
> > > passing univ8250_port_base_ops to univ8250_rsa_support() that can make
> > > a local copy of it.
> > > 
> > > Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> > > Reported-by: Alex Davis <alex47794@gmail.com>
> > > Fixes: b20d6576cdb3 ("serial: 8250: export RSA functions")
> > > Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>  
> > 
> > LGTM, thanks for the fix.
> > 
> > Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
> > 
> > If the reporters could give it a shot and mark this by Tested-by, it would be great...
> 
> I have not seen this for quite some time ... I assumed it had been
> fixed.

I too thought this was fixed a while ago, what changed to cause this to
happen now?  Anyone have a .config to reproduce this, 'allmodconfig' on
x86 works for me.

greg k-h

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2141 bytes --]

Hi Greg,

On Sun, 9 Nov 2025 09:49:29 +0900 Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
>
> On Sun, Nov 09, 2025 at 10:52:03AM +1100, Stephen Rothwell wrote:
> > 
> > On Fri, 7 Nov 2025 07:20:26 +0100 Jiri Slaby <jirislaby@kernel.org> wrote:  
> > >
> > > On 06. 11. 25, 19:00, Ilpo Järvinen wrote:  
> > > > This seems to resolve the build issue for me:
> > > > 
> > > > --
> > > > From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>
> > > > Subject: [PATCH 1/1] serial: 8250: Fix 8250_rsa symbol loop
> > > > 
> > > > make allmodconfig build fails due to dependency loop:
> > > > 
> > > >    depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > > >    depmod: ERROR: Found 2 modules in dependency cycles!
> > > > 
> > > > Break dependency loop by moving 8250_rsa.o into 8250_base and by
> > > > passing univ8250_port_base_ops to univ8250_rsa_support() that can make
> > > > a local copy of it.
> > > > 
> > > > Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> > > > Reported-by: Alex Davis <alex47794@gmail.com>
> > > > Fixes: b20d6576cdb3 ("serial: 8250: export RSA functions")
> > > > Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>    
> > > 
> > > LGTM, thanks for the fix.
> > > 
> > > Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
> > > 
> > > If the reporters could give it a shot and mark this by Tested-by, it would be great...  
> > 
> > I have not seen this for quite some time ... I assumed it had been
> > fixed.  
> 
> I too thought this was fixed a while ago, what changed to cause this to
> happen now?  Anyone have a .config to reproduce this, 'allmodconfig' on
> x86 works for me.

Actually, I think the current report is for v6.17.1-rc1, so maybe
something got missed in a stable backport?  My original report was for
next-20250728 (July 28).  The only response I got was on October 1,
but I am pretty sure it was fixed well before then.

Also, you probably only get the error when you do a "make
modules_install" after the allmodconfig build (at least that is where I
got it).
-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [Regression] depmod fails on kernel 6.17.1 rc1

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 5044 bytes --]

On Sun, 9 Nov 2025, Stephen Rothwell wrote:
> On Sun, 9 Nov 2025 09:49:29 +0900 Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
> > On Sun, Nov 09, 2025 at 10:52:03AM +1100, Stephen Rothwell wrote:
> > > On Fri, 7 Nov 2025 07:20:26 +0100 Jiri Slaby <jirislaby@kernel.org> wrote:  
> > > >
> > > > On 06. 11. 25, 19:00, Ilpo Järvinen wrote:  
> > > > > This seems to resolve the build issue for me:
> > > > > 
> > > > > --
> > > > > From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>
> > > > > Subject: [PATCH 1/1] serial: 8250: Fix 8250_rsa symbol loop
> > > > > 
> > > > > make allmodconfig build fails due to dependency loop:
> > > > > 
> > > > >    depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
> > > > >    depmod: ERROR: Found 2 modules in dependency cycles!
> > > > > 
> > > > > Break dependency loop by moving 8250_rsa.o into 8250_base and by
> > > > > passing univ8250_port_base_ops to univ8250_rsa_support() that can make
> > > > > a local copy of it.
> > > > > 
> > > > > Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> > > > > Reported-by: Alex Davis <alex47794@gmail.com>
> > > > > Fixes: b20d6576cdb3 ("serial: 8250: export RSA functions")
> > > > > Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>    
> > > > 
> > > > LGTM, thanks for the fix.
> > > > 
> > > > Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
> > > > 
> > > > If the reporters could give it a shot and mark this by Tested-by, it would be great...  
> > > 
> > > I have not seen this for quite some time ... I assumed it had been
> > > fixed.  
> > 
> > I too thought this was fixed a while ago, what changed to cause this to
> > happen now?  Anyone have a .config to reproduce this, 'allmodconfig' on
> > x86 works for me.
> 
> Actually, I think the current report is for v6.17.1-rc1, so maybe
> something got missed in a stable backport? My original report was for
> next-20250728 (July 28).  The only response I got was on October 1,
> but I am pretty sure it was fixed well before then.
> 
> Also, you probably only get the error when you do a "make
> modules_install" after the allmodconfig build (at least that is where I
> got it).

Hi all,

I don't want to waste my time building allmodconfig but this is with 
allnoconfig + a few key CONFIGs set and without my patch:

$ git log -n1
commit e9a6fb0bcdd7609be6969112f3fbfcce3b1d4a7c (grafted, HEAD -> master, 
tag: v6.18-rc5, origin/master, origin/HEAD)
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Sun Nov 9 15:10:19 2025 -0800

    Linux 6.18-rc5
$ make allnoconfig
  LEX     scripts/kconfig/lexer.lex.c
  YACC    scripts/kconfig/parser.tab.[ch]
  HOSTCC  scripts/kconfig/lexer.lex.o
  HOSTCC  scripts/kconfig/menu.o
  HOSTCC  scripts/kconfig/parser.tab.o
  HOSTCC  scripts/kconfig/preprocess.o
  HOSTCC  scripts/kconfig/symbol.o
  HOSTCC  scripts/kconfig/util.o
  HOSTLD  scripts/kconfig/conf
#
# configuration written to .config
#
$ echo -e "CONFIG_MODULES=y\nCONFIG_SERIAL_8250=m\nCONFIG_SERIAL_8250_EXTENDED=y\nCONFIG_SERIAL_8250_RSA=y" >> .config 
$ make olddefconfig
.config:1528:warning: override: reassigning to symbol MODULES
.config:1529:warning: override: reassigning to symbol SERIAL_8250
.config:1530:warning: override: reassigning to symbol SERIAL_8250_EXTENDED
#
# configuration written to .config
#
$ make -j4 > /dev/null && sudo make modules_install
  INSTALL /lib/modules/6.18.0-rc5/modules.order
  INSTALL /lib/modules/6.18.0-rc5/modules.builtin
  INSTALL /lib/modules/6.18.0-rc5/modules.builtin.modinfo
  SYMLINK /lib/modules/6.18.0-rc5/build
  INSTALL /lib/modules/6.18.0-rc5/kernel/drivers/tty/serial/8250/8250.ko
  INSTALL /lib/modules/6.18.0-rc5/kernel/drivers/tty/serial/8250/8250_base.ko
  INSTALL /lib/modules/6.18.0-rc5/kernel/drivers/tty/serial/serial_base.ko
  DEPMOD  /lib/modules/6.18.0-rc5
depmod: ERROR: Cycle detected: 8250 -> 8250_base -> 8250
depmod: ERROR: Found 2 modules in dependency cycles!
make[2]: *** [scripts/Makefile.modinst:132: depmod] Error 1
make[1]: *** [/home/user/linux/Makefile:1916: modules_install] Error 2
make: *** [Makefile:248: __sub-make] Error 2

...so no, it does not look fixed.

With the patch:

$ git am 0001-serial-8250-Fix-8250_rsa-symbol-loop.patch 
Applying: serial: 8250: Fix 8250_rsa symbol loop
user@disp1097:~/linux$ make -j4 > /dev/null && sudo make modules_install
  SYMLINK /lib/modules/6.18.0-rc5+/build
  INSTALL /lib/modules/6.18.0-rc5+/modules.order
  INSTALL /lib/modules/6.18.0-rc5+/modules.builtin
  INSTALL /lib/modules/6.18.0-rc5+/modules.builtin.modinfo
  INSTALL /lib/modules/6.18.0-rc5+/kernel/drivers/tty/serial/8250/8250.ko
  INSTALL /lib/modules/6.18.0-rc5+/kernel/drivers/tty/serial/8250/8250_base.ko
  INSTALL /lib/modules/6.18.0-rc5+/kernel/drivers/tty/serial/serial_base.ko
  DEPMOD  /lib/modules/6.18.0-rc5+
$ 

I'll just make the official submission with the #else block's prototype 
fixed (thanks to lkp).

-- 
 i.