serial console on Mikrotik RB532 non-working

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 is the first bad commit commit 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6

Author: Yi Yang <[email protected]>
Date:   Tue Jun 28 16:35:15 2022 +0800

serial: 8250: fix return error code in serial8250_request_std_resource()

If port->mapbase = NULL in serial8250_request_std_resource() , it need return a error code instead of 0. If uart_set_info() fail to request new regions by serial8250_request_std_resource() but the return value of serial8250_request_std_resource() is 0, The system incorrectly considers that the resource application is successful and does not attempt to restore the old setting. A null pointer reference is triggered when the port resource is later invoked.

Signed-off-by Yi Yang <[email protected]>

Cc stable <[email protected]>

Link https://lore.kernel.org/r/[email protected] open here

Signed-off-by Greg Kroah-Hartman <[email protected]>

 drivers/tty/serial/8250/8250_port.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

When I compile the commit before this commit everything is fine. When I just try to revert it on master I still get no output on my serial console.

Anyone here have an idea what is defect in arch/mips/rb532 nowadays?

Thanks in advance,
Waldemar

Quoted 5 lines · Waldemar Brodkorb

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Hi,

Quoted 56 lines · Waldemar Brodkorb

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Quoted 2 lines

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Have you tried the patch on top of master or
6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6?

If it works on top of 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6, then it fixes the serial issue, and you can do a new git bisect between 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 and master to find the next breakage, with the patch re-applied on top each step (so serial stays working).

Hi Jonas,
Jonas Gorski wrote,

Quoted 62 lines

Hi,

Quoted 56 lines · Waldemar Brodkorb

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Quoted 2 lines

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Quoted 8 lines

Have you tried the patch on top of master or
6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6?

If it works on top of 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6, then it fixes the serial issue, and you can do a new git bisect between 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 and master to find the next breakage, with the patch re-applied on top each step (so serial stays working).

Yes, thanks for the idea. I already had this idea in mind and I am 9 steps before I get the next breakage :)

Thanks for the heads up,

Quoted 75 lines

Hi Jonas,
Jonas Gorski wrote,

Quoted 62 lines

Hi,

Quoted 56 lines · Waldemar Brodkorb

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Quoted 2 lines

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Quoted 8 lines

Have you tried the patch on top of master or
6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6?

If it works on top of 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6, then it fixes the serial issue, and you can do a new git bisect between 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 and master to find the next breakage, with the patch re-applied on top each step (so serial stays working).

Quoted 4 lines

Yes, thanks for the idea. I already had this idea in mind and I am 9 steps before I get the next breakage :)

Thanks for the heads up,

Also a small warning when doing a bisect with git am'd patches on top:
always (git) reset to the original tested commit before doing git bisect <bood|bad>, else git bisect will account for these commits on top in calculating the next revision to test and may suggest the very same commit again you were originally testing.

It took a me a few steps until I noticed I was running in circles when I had to bisect with a few patches on top.

Hi Mips people,
Jonas Gorski wrote,

Quoted 81 lines · Waldemar Brodkorb

Quoted 75 lines

Hi Jonas,
Jonas Gorski wrote,

Quoted 62 lines

Hi,

Quoted 56 lines · Waldemar Brodkorb

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Quoted 2 lines

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Quoted 8 lines

Have you tried the patch on top of master or
6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6?

If it works on top of 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6, then it fixes the serial issue, and you can do a new git bisect between 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 and master to find the next breakage, with the patch re-applied on top each step (so serial stays working).

Quoted 4 lines

Yes, thanks for the idea. I already had this idea in mind and I am 9 steps before I get the next breakage :)

Thanks for the heads up,

Quoted 8 lines

Also a small warning when doing a bisect with git am'd patches on top:
always (git) reset to the original tested commit before doing git bisect <bood|bad>, else git bisect will account for these commits on top in calculating the next revision to test and may suggest the very same commit again you were originally testing.

It took a me a few steps until I noticed I was running in circles when I had to bisect with a few patches on top.

I just used patch and git stash for the job :) So here is the result, the failing commit is:
commit 9f048fa487409e364cf866c957cf0b0d782ca5a3 (HEAD)

Author: Maciej W. Rozycki <[email protected]>
Date:   Thu Nov 13 05:21:10 2025 +0000

MIPS: mm: Prevent a TLB shutdown on initial uniquification

Depending on the particular CPU implementation a TLB shutdown may occur if multiple matching entries are detected upon the execution of a TLBP or the TLBWI/TLBWR instructions. Given that we don't know what entries we have been handed we need to be very careful with the initial TLB setup and avoid all these instructions.

Therefore read all the TLB entries one by one with the TLBR instruction, bypassing the content addressing logic, and truncate any large pages in place so as to avoid a case in the second step where an incoming entry for a large page at a lower address overlaps with a replacement entry chosen at another index. Then preinitialize the TLB using addresses outside our usual unique range and avoiding clashes with any entries received, before making the usual call to local_flush_tlb_all().

This fixes (at least) R4x00 cores if TLBP hits multiple matching TLB entries (SGI IP22 PROM for examples sets up all TLBs to the same virtual address).

Signed-off-by Maciej W. Rozycki <[email protected]>

Fixes 35ad7e181541 ("MIPS: mm: tlb-r4k: Uniquify TLB entries on init")

Cc [email protected]

Reviewed-by Jiaxun Yang <[email protected]>

Tested-by Jiaxun Yang <[email protected]> # Boston I6400, M5150 sim

Signed-off-by Thomas Bogendoerfer <[email protected]>

Reverting this commit, make the RB532 boot. I need the UART patch, too.

Attached is the dmesg from the device.

Anyone knows how to fix this issue?

best regards
Waldemar

Hi Waldemar,

Quoted 126 lines · Waldemar Brodkorb

Hi Mips people,
Jonas Gorski wrote,

Quoted 81 lines · Waldemar Brodkorb

Quoted 75 lines

Hi Jonas,
Jonas Gorski wrote,

Quoted 62 lines

Hi,

Quoted 56 lines · Waldemar Brodkorb

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Quoted 2 lines

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Quoted 8 lines

Have you tried the patch on top of master or
6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6?

If it works on top of 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6, then it fixes the serial issue, and you can do a new git bisect between 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 and master to find the next breakage, with the patch re-applied on top each step (so serial stays working).

Quoted 4 lines

Yes, thanks for the idea. I already had this idea in mind and I am 9 steps before I get the next breakage :)

Thanks for the heads up,

Quoted 8 lines

Also a small warning when doing a bisect with git am'd patches on top:
always (git) reset to the original tested commit before doing git bisect <bood|bad>, else git bisect will account for these commits on top in calculating the next revision to test and may suggest the very same commit again you were originally testing.

It took a me a few steps until I noticed I was running in circles when I had to bisect with a few patches on top.

I just used patch and git stash for the job :) So here is the result, the failing commit is:
commit 9f048fa487409e364cf866c957cf0b0d782ca5a3 (HEAD)

Author: Maciej W. Rozycki <[email protected]>
Date:   Thu Nov 13 05:21:10 2025 +0000

MIPS: mm: Prevent a TLB shutdown on initial uniquification

Depending on the particular CPU implementation a TLB shutdown may occur if multiple matching entries are detected upon the execution of a TLBP or the TLBWI/TLBWR instructions. Given that we don't know what entries we have been handed we need to be very careful with the initial TLB setup and avoid all these instructions.

Therefore read all the TLB entries one by one with the TLBR instruction, bypassing the content addressing logic, and truncate any large pages in place so as to avoid a case in the second step where an incoming entry for a large page at a lower address overlaps with a replacement entry chosen at another index. Then preinitialize the TLB using addresses outside our usual unique range and avoiding clashes with any entries received, before making the usual call to local_flush_tlb_all().

This fixes (at least) R4x00 cores if TLBP hits multiple matching TLB entries (SGI IP22 PROM for examples sets up all TLBs to the same virtual address).

Signed-off-by Maciej W. Rozycki <[email protected]>

Fixes 35ad7e181541 ("MIPS: mm: tlb-r4k: Uniquify TLB entries on init")

Cc [email protected]

Reviewed-by Jiaxun Yang <[email protected]>

Tested-by Jiaxun Yang <[email protected]> # Boston I6400, M5150 sim

Signed-off-by Thomas Bogendoerfer <[email protected]>

Quoted 1 line

Reverting this commit, make the RB532 boot. I need the UART patch, too.

FYI, this commit is buggy and has been fixed by commit 841ecc979b18d ("MIPS: mm: kmalloc tlb_vpn array to avoid stack overflow").

Hi Rong,
Rong Zhang wrote,

Quoted 129 lines

Hi Waldemar,

Quoted 126 lines · Waldemar Brodkorb

Hi Mips people,
Jonas Gorski wrote,

Quoted 81 lines · Waldemar Brodkorb

Quoted 75 lines

Hi Jonas,
Jonas Gorski wrote,

Quoted 62 lines

Hi,

Quoted 56 lines · Waldemar Brodkorb

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Quoted 2 lines

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Quoted 8 lines

Have you tried the patch on top of master or
6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6?

If it works on top of 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6, then it fixes the serial issue, and you can do a new git bisect between 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 and master to find the next breakage, with the patch re-applied on top each step (so serial stays working).

Quoted 4 lines

Yes, thanks for the idea. I already had this idea in mind and I am 9 steps before I get the next breakage :)

Thanks for the heads up,

Quoted 8 lines

Also a small warning when doing a bisect with git am'd patches on top:
always (git) reset to the original tested commit before doing git bisect <bood|bad>, else git bisect will account for these commits on top in calculating the next revision to test and may suggest the very same commit again you were originally testing.

It took a me a few steps until I noticed I was running in circles when I had to bisect with a few patches on top.

I just used patch and git stash for the job :) So here is the result, the failing commit is:
commit 9f048fa487409e364cf866c957cf0b0d782ca5a3 (HEAD)

Author: Maciej W. Rozycki <[email protected]>
Date:   Thu Nov 13 05:21:10 2025 +0000

MIPS: mm: Prevent a TLB shutdown on initial uniquification

Depending on the particular CPU implementation a TLB shutdown may occur if multiple matching entries are detected upon the execution of a TLBP or the TLBWI/TLBWR instructions. Given that we don't know what entries we have been handed we need to be very careful with the initial TLB setup and avoid all these instructions.

Therefore read all the TLB entries one by one with the TLBR instruction, bypassing the content addressing logic, and truncate any large pages in place so as to avoid a case in the second step where an incoming entry for a large page at a lower address overlaps with a replacement entry chosen at another index. Then preinitialize the TLB using addresses outside our usual unique range and avoiding clashes with any entries received, before making the usual call to local_flush_tlb_all().

This fixes (at least) R4x00 cores if TLBP hits multiple matching TLB entries (SGI IP22 PROM for examples sets up all TLBs to the same virtual address).

Signed-off-by Maciej W. Rozycki <[email protected]>

Fixes 35ad7e181541 ("MIPS: mm: tlb-r4k: Uniquify TLB entries on init")

Cc [email protected]

Reviewed-by Jiaxun Yang <[email protected]>

Tested-by Jiaxun Yang <[email protected]> # Boston I6400, M5150 sim

Signed-off-by Thomas Bogendoerfer <[email protected]>

Quoted 4 lines

Quoted 1 line

Reverting this commit, make the RB532 boot. I need the UART patch, too.

FYI, this commit is buggy and has been fixed by commit 841ecc979b18d ("MIPS: mm: kmalloc tlb_vpn array to avoid stack overflow").

But if you take a look at the dmesg, I am using latest Linux GIT master, which includes this commit. But still it is failing for me.

best regards
Waldemar

just for your information. This commit also breaks bootup of Microchip PIC32MZ Graphics (DA) Starter Kit.

Quoted 3 lines

just for your information. This commit also breaks bootup of Microchip PIC32MZ Graphics (DA) Starter Kit.

I'm working on a fix, but it's a tough issue, because at kernel entry we
don't know the state of the TLB and the existence/activation of the hidden TLB "entry disable" bit can't be determined (in which case we could just proceed to the final initialisation). I should have something in the next week or so, but in the meanwhile please just revert the commits that affect your particular configuration.

Hi Mips people,
Jonas Gorski wrote,

Quoted 81 lines · Waldemar Brodkorb

Quoted 75 lines

Hi Jonas,
Jonas Gorski wrote,

Quoted 62 lines

Hi,

Quoted 56 lines · Waldemar Brodkorb

Hi Jiaxun,
Jiaxun Yang wrote,

Quoted 25 lines · Jiaxun Yang

Quoted 10 lines · Waldemar Brodkorb

Quoted 5 lines

Hi MIPS hackers,

I want to use the latest Linux kernel on my Mikrotik RB532, but the serial console is non working. I bisected the problem and the first breaking change is this commit:

Hi,

Can you try this?

Oops sorry I missed a flag, it should be:

arch/mips/rb532/devices.c +0 -1

index 8ecb56be81ac..239018540221 100644
@@ -213,11 +213,12 @@ static struct platform_device rb532_wdt = {
 static struct plat_serial8250_port rb532_uart_res[] = {
        {
                .type           = PORT_16550A,
-               .membase        = (char *)KSEG1ADDR(REGBASE + UART0BASE),

arch/mips/rb532/devices.c static struct platform_device rb532_wdt = { · line 224

Quoted 10 lines

arch/mips/rb532/devices.c +3 -1

+               .mapbase        = REGBASE + UART0BASE,
+               .mapsize        = SZ_4K,
                .irq            = UART0_IRQ,
                .regshift       = 2,
                .iotype         = UPIO_MEM,
-               .flags          = UPF_BOOT_AUTOCONF,
+               .flags          = UPF_BOOT_AUTOCONF | UPF_IOREMAP,
        },
        {
                .flags          = 0,

I tried the patch, but it still hangs here:

RouterBOOT booter 2.18

RouterBoard 532

CPU frequency: 399 MHz
Memory size: 32 MB

Press any key within 3 seconds to enter setup...
trying dhcp protocol... OK
resolved mac address 9C:BF:0D:00:D6:4E
Gateway: 192.168.1.254
transfer started ........................................ transfer ok, time=525.70s setting up elf image... OK
jumping to kernel code

Quoted 2 lines

Nothing in tcpdump tries to mount nfs, so I believe serial console breakage is not the only problem here.

Quoted 8 lines

Have you tried the patch on top of master or
6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6?

If it works on top of 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6, then it fixes the serial issue, and you can do a new git bisect between 6e690d54cfa802f939cefbd2fa2c91bd0b8bd1b6 and master to find the next breakage, with the patch re-applied on top each step (so serial stays working).

Quoted 4 lines

Yes, thanks for the idea. I already had this idea in mind and I am 9 steps before I get the next breakage :)

Thanks for the heads up,

Quoted 8 lines

Also a small warning when doing a bisect with git am'd patches on top:
always (git) reset to the original tested commit before doing git bisect <bood|bad>, else git bisect will account for these commits on top in calculating the next revision to test and may suggest the very same commit again you were originally testing.

It took a me a few steps until I noticed I was running in circles when I had to bisect with a few patches on top.

I just used patch and git stash for the job :) So here is the result, the failing commit is:
commit 9f048fa487409e364cf866c957cf0b0d782ca5a3 (HEAD)

Author: Maciej W. Rozycki <[email protected]>
Date:   Thu Nov 13 05:21:10 2025 +0000

MIPS: mm: Prevent a TLB shutdown on initial uniquification

Depending on the particular CPU implementation a TLB shutdown may occur if multiple matching entries are detected upon the execution of a TLBP or the TLBWI/TLBWR instructions. Given that we don't know what entries we have been handed we need to be very careful with the initial TLB setup and avoid all these instructions.

Therefore read all the TLB entries one by one with the TLBR instruction, bypassing the content addressing logic, and truncate any large pages in place so as to avoid a case in the second step where an incoming entry for a large page at a lower address overlaps with a replacement entry chosen at another index. Then preinitialize the TLB using addresses outside our usual unique range and avoiding clashes with any entries received, before making the usual call to local_flush_tlb_all().

This fixes (at least) R4x00 cores if TLBP hits multiple matching TLB entries (SGI IP22 PROM for examples sets up all TLBs to the same virtual address).

Signed-off-by Maciej W. Rozycki <[email protected]>

Fixes 35ad7e181541 ("MIPS: mm: tlb-r4k: Uniquify TLB entries on init")

Cc [email protected]

Reviewed-by Jiaxun Yang <[email protected]>

Tested-by Jiaxun Yang <[email protected]> # Boston I6400, M5150 sim

Signed-off-by Thomas Bogendoerfer <[email protected]>

Reverting this commit, make the RB532 boot. I need the UART patch, too.

Attached is the dmesg from the device.

Anyone knows how to fix this issue?

best regards
Waldemar

[    0.000000] Linux version 6.19.0-rc8+ (wbx@fluor) (mipsel-openadk-linux-uclibc-gcc (GCC) 15.2.0, GNU ld (GNU Binutils) 2.45.1) #50 Wed Feb  4 16:41:19 CET 2026
[    0.000000] CPU0 revision is: 0001800a (MIPS 4Kc)
[    0.000000] User-defined physical RAM map overwrite
[    0.000000] Primary instruction cache 8kB, VIPT, 4-way, linesize 16 bytes.
[    0.000000] Primary data cache 8kB, 4-way, VIPT, no aliases, linesize 16 bytes
[    0.000000] Zone ranges:
[    0.000000]   Normal   [mem 0x0000000000000000-0x0000000001ffffff]
[    0.000000] Movable zone start for each node
[    0.000000] Early memory node ranges
[    0.000000]   node   0: [mem 0x0000000000000000-0x0000000001ffffff]
[    0.000000] Initmem setup node 0 [mem 0x0000000000000000-0x0000000001ffffff]
[    0.000000] pcpu-alloc: s0 r0 d32768 u32768 alloc=1*32768
[    0.000000] pcpu-alloc: [0] 0 
[    0.000000] Kernel command line: console=ttyS0,115200 gpio=16383 mem=32M kmac=00:0C:42:0A:7D:D3 board=500 boot=1   console=ttyS0,115200 rw root=/dev/nfs ip=dhcp nfsroot=192.168.1.254:/srv/nfs/rb532
[    0.000000] korina mac = 00:0C:42:0A:7D:D3
[    0.000000] Unknown kernel command line parameters "gpio=16383 board=500 boot=1", will be passed to user space.
[    0.000000] printk: log buffer data + meta data: 131072 + 409600 = 540672 bytes
[    0.000000] Dentry cache hash table entries: 4096 (order: 2, 16384 bytes, linear)
[    0.000000] Inode-cache hash table entries: 2048 (order: 1, 8192 bytes, linear)
[    0.000000] Built 1 zonelists, mobility grouping on.  Total pages: 8192
[    0.000000] mem auto-init: stack:all(zero), heap alloc:off, heap free:off
[    0.000000] SLUB: HWalign=16, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
[    0.000000] NR_IRQS: 256
[    0.000000] Initializing IRQ's: 168 out of 256
[    0.000000] calculating r4koff... 
[    0.000000] 000c34f8(799992)
[    0.000000] CPU frequency 400.00 MHz
[    0.000000] clocksource: MIPS: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 9556397797 ns
[    0.000010] sched_clock: 32 bits at 200MHz, resolution 5ns, wraps every 10737525757ns
[    0.000236] Calibrating delay loop... 397.82 BogoMIPS (lpj=795648)
[    0.032139] pid_max: default: 32768 minimum: 301
[    0.033029] Mount-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
[    0.033136] Mountpoint-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
[    0.042812] Memory: 26612K/32768K available (3993K kernel code, 542K rwdata, 548K rodata, 152K init, 188K bss, 5760K reserved, 0K cma-reserved)
[    0.044819] devtmpfs: initialized
[    0.053484] clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 7645041785100000 ns
[    0.053636] posixtimers hash table entries: 512 (order: 0, 2048 bytes, linear)
[    0.053740] futex hash table entries: 256 (4096 bytes on 1 NUMA nodes, total 4 KiB, linear).
[    0.060069] NET: Registered PF_NETLINK/PF_ROUTE protocol family
[    0.062045] gpio gpiochip0: Static allocation of GPIO base is deprecated, use dynamic allocation.
[    0.064668] PCI: Initializing PCI
[    0.070830] SCSI subsystem initialized
[    0.071977] libata version 3.00 loaded.
[    0.074816] PCI host bridge to bus 0000:00
[    0.074882] pci_bus 0000:00: root bus resource [mem 0x50000000-0x5fffffff]
[    0.074978] pci_bus 0000:00: root bus resource [io  0x18800000-0x188fffff]
[    0.075066] pci_bus 0000:00: No busn resource found for root bus, will use [bus 00-ff]
[    0.075217] pci 0000:00:00.0: [111d:0000] type 00 class 0x000000 conventional PCI endpoint
[    0.075353] pci 0000:00:00.0: BAR 0 [mem 0x00000000-0x07ffffff pref]
[    0.075449] pci 0000:00:00.0: [Firmware Bug]: BAR 1: invalid; can't size
[    0.075512] pci 0000:00:00.0: [Firmware Bug]: BAR 2: invalid; can't size
[    0.101583] pci 0000:00:02.0: [1106:3106] type 00 class 0x020000 conventional PCI endpoint
[    0.101736] pci 0000:00:02.0: BAR 0 [io  0x0000-0x00ff]
[    0.101831] pci 0000:00:02.0: BAR 1 [mem 0x00000000-0x000000ff]
[    0.102022] pci 0000:00:02.0: supports D1 D2
[    0.102087] pci 0000:00:02.0: PME# supported from D1 D2 D3hot D3cold
[    0.102883] pci 0000:00:03.0: [1106:3106] type 00 class 0x020000 conventional PCI endpoint
[    0.103036] pci 0000:00:03.0: BAR 0 [io  0x0000-0x00ff]
[    0.103131] pci 0000:00:03.0: BAR 1 [mem 0x00000000-0x000000ff]
[    0.103322] pci 0000:00:03.0: supports D1 D2
[    0.103386] pci 0000:00:03.0: PME# supported from D1 D2 D3hot D3cold
[    0.104215] pci 0000:00:04.0: [13a3:0020] type 00 class 0x0b4000 conventional PCI endpoint
[    0.104368] pci 0000:00:04.0: BAR 0 [mem 0x00000000-0x00000fff]
[    0.104463] pci 0000:00:04.0: BAR 1 [mem 0x00000000-0x00001fff]
[    0.104553] pci 0000:00:04.0: BAR 2 [mem 0x00000000-0x00007fff]
[    0.584389] pci_bus 0000:00: busn_res: [bus 00-ff] end is updated to 00
[    0.584536] pci 0000:00:04.0: BAR 2 [mem 0x50000000-0x50007fff]: assigned
[    0.584656] pci 0000:00:04.0: BAR 1 [mem 0x50008000-0x50009fff]: assigned
[    0.584771] pci 0000:00:04.0: BAR 0 [mem 0x5000a000-0x5000afff]: assigned
[    0.584887] pci 0000:00:02.0: BAR 0 [io  0x18800000-0x188000ff]: assigned
[    0.585003] pci 0000:00:02.0: BAR 1 [mem 0x5000b000-0x5000b0ff]: assigned
[    0.585119] pci 0000:00:03.0: BAR 0 [io  0x18800400-0x188004ff]: assigned
[    0.585235] pci 0000:00:03.0: BAR 1 [mem 0x5000b100-0x5000b1ff]: assigned
[    0.586040] clocksource: Switched to clocksource MIPS
[    0.595022] NET: Registered PF_INET protocol family
[    0.596277] IP idents hash table entries: 2048 (order: 2, 16384 bytes, linear)
[    0.599075] tcp_listen_portaddr_hash hash table entries: 1024 (order: 0, 4096 bytes, linear)
[    0.599207] Table-perturb hash table entries: 65536 (order: 6, 262144 bytes, linear)
[    0.599357] TCP established hash table entries: 1024 (order: 0, 4096 bytes, linear)
[    0.599483] TCP bind hash table entries: 1024 (order: 1, 8192 bytes, linear)
[    0.599665] TCP: Hash tables configured (established 1024 bind 1024)
[    0.600026] UDP hash table entries: 256 (order: 1, 8192 bytes, linear)
[    0.600186] UDP-Lite hash table entries: 256 (order: 1, 8192 bytes, linear)
[    0.601161] NET: Registered PF_UNIX/PF_LOCAL protocol family
[    0.603686] RPC: Registered named UNIX socket transport module.
[    0.603736] RPC: Registered udp transport module.
[    0.603769] RPC: Registered tcp transport module.
[    0.603802] RPC: Registered tcp-with-tls transport module.
[    0.603835] RPC: Registered tcp NFSv4.1 backchannel transport module.
[    0.603971] PCI: CLS 16 bytes, default 16
[    0.607134] workingset: timestamp_bits=30 max_order=13 bucket_order=0
[    0.613358] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
[    0.621159] printk: legacy console [ttyS0] disabled
[    0.622542] serial8250.0: ttyS0 at MMIO 0x18058000 (irq = 104, base_baud = 12499875) is a 16550A
[    0.622738] printk: legacy console [ttyS0] enabled
[    1.299755] scsi host0: pata-rb532-cf
[    1.305007] ata1: PATA max PIO4 irq 149 lpm-pol 0
[    1.313888] eth0: korina-0.20 15Sep2017
[    1.318861] via-rhine 0000:00:02.0: enabling device (0080 -> 0083)
[    1.334840] via-rhine 0000:00:02.0 eth1: VIA Rhine III at (ptrval), 00:0c:42:0a:7d:d4, IRQ 142
[    1.345448] via-rhine 0000:00:02.0 eth1: MII PHY found at address 1, status 0x7849 advertising 05e1 Link 0000
[    1.356989] via-rhine 0000:00:03.0: enabling device (0080 -> 0083)
[    1.372559] via-rhine 0000:00:03.0 eth2: VIA Rhine III at (ptrval), 00:0c:42:0a:7d:d5, IRQ 143
[    1.383162] via-rhine 0000:00:03.0 eth2: MII PHY found at address 1, status 0x7849 advertising 05e1 Link 0000
[    1.395174] rc32434_wdt: Stopped watchdog timer
[    1.401166] rc32434_wdt: Watchdog Timer version 1.0, timer margin: 20 sec
[    1.410007] NET: Registered PF_PACKET protocol family
[    1.488756] korina korina eth0: link up, 100Mbps, full-duplex, lpa 0xFFFFFF6F
[    1.514213] Sending DHCP requests ., OK
[    1.542622] IP-Config: Got DHCP answer from 192.168.1.254, my address is 192.168.1.9
[    1.551409] IP-Config: Complete:
[    1.555148]      device=eth0, hwaddr=00:0c:42:0a:7d:d3, ipaddr=192.168.1.9, mask=255.255.255.0, gw=192.168.1.254
[    1.566657]      host=192.168.1.9, domain=, nis-domain=(none)
[    1.573201]      bootserver=192.168.1.254, rootserver=192.168.1.254, rootpath=
[    1.573275]      nameserver0=192.168.1.254
[    1.588787] check access for rdinit=/init failed: -2, ignoring
[    1.658499] VFS: Mounted root (nfs filesystem) on device 0:13.
[    1.667866] devtmpfs: mounted
[    1.671780] Freeing unused kernel image (initmem) memory: 152K
[    1.677854] This architecture does not have kernel memory protection.
[    1.684548] Run /sbin/init as init process
[    1.688861]   with arguments:
[    1.688920]     /sbin/init
[    1.688977]   with environment:
[    1.689027]     HOME=/
[    1.689082]     TERM=linux
[    1.689137]     gpio=16383
[    1.689193]     board=500
[    1.689248]     boot=1
[    1.723116] process '/bin/busybox' started with executable stack
[  208.334338] random: crng init done