Re: serial console on Mikrotik RB532 non-working

[Waldemar Brodkorb <wbx@openadk.org>]

Hi Again,

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

best regards
 Waldemar

Waldemar Brodkorb wrote,

> Hi Rong,
> Rong Zhang wrote,
> 
> > Hi Waldemar,
> > 
> > On Wed, 2026-02-04 at 17:38 +0100, Waldemar Brodkorb wrote:
> > > Hi Mips people,
> > > Jonas Gorski wrote,
> > > 
> > > > On Wed, Feb 4, 2026 at 12:19 PM Waldemar Brodkorb <wbx@openadk.org> wrote:
> > > > > 
> > > > > Hi Jonas,
> > > > > Jonas Gorski wrote,
> > > > > 
> > > > > > Hi,
> > > > > > 
> > > > > > On Tue, Feb 3, 2026 at 7:25 PM Waldemar Brodkorb <wbx@openadk.org> wrote:
> > > > > > > 
> > > > > > > Hi Jiaxun,
> > > > > > > Jiaxun Yang wrote,
> > > > > > > > 
> > > > > > > > On Tue, 3 Feb 2026, at 3:24 PM, Jiaxun Yang wrote:
> > > > > > > > > On Sun, 1 Feb 2026, at 6:39 PM, Waldemar Brodkorb wrote:
> > > > > > > > > > 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:
> > > > > > > > 
> > > > > > > > diff --git a/arch/mips/rb532/devices.c b/arch/mips/rb532/devices.c
> > > > > > > > index 8ecb56be81ac..239018540221 100644
> > > > > > > > --- a/arch/mips/rb532/devices.c
> > > > > > > > +++ b/arch/mips/rb532/devices.c
> > > > > > > > @@ -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),
> > > > > > > > +               .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.
> > > > > > 
> > > > > > 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,
> > > > 
> > > > 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 <macro@orcam.me.uk>
> > > 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 <macro@orcam.me.uk>
> > >     Fixes: 35ad7e181541 ("MIPS: mm: tlb-r4k: Uniquify TLB entries on init")
> > >     Cc: stable@vger.kernel.org
> > >     Reviewed-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
> > >     Tested-by: Jiaxun Yang <jiaxun.yang@flygoat.com> # Boston I6400, M5150 sim
> > >     Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
> > > 
> > > 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
>