Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Message-ID: <2a3f8b4c-2c0d-28bc-8dcd-c56c7b8a2bb4@landley.net>
Date:   Mon, 2 May 2022 15:07:08 -0500
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.6.2
Subject: Re: [PATCH v3] sh: avoid using IRQ0 on SH3/4
Content-Language: en-US
To:     Geert Uytterhoeven <geert@linux-m68k.org>
Cc:     John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>,
        Sergey Shtylyov <s.shtylyov@omp.ru>,
        Rich Felker <dalias@libc.org>,
        Linux-sh list <linux-sh@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Yoshinori Sato <ysato@users.sourceforge.jp>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>
References: <2584ba18-9653-9310-efc1-8b3b3e221eea@omp.ru>
 <11021433-66c0-3c56-42bd-207a5ae8d267@physik.fu-berlin.de>
 <2ebef1ac-e5c5-980c-9413-22a6cccdfa1d@landley.net>
 <CAMuHMdWN0vRYhK7O0MgOSCtisw3RDvp4vxSS2VF-9uGDdOEb7g@mail.gmail.com>
From:   Rob Landley <rob@landley.net>
In-Reply-To: <CAMuHMdWN0vRYhK7O0MgOSCtisw3RDvp4vxSS2VF-9uGDdOEb7g@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 5/2/22 03:37, Geert Uytterhoeven wrote:
> Hi Rob,
...
> Until I tried "ifconfig eth0 up", which causes a lock-up.
> Interestingly, the 8139 irq was 112 with and without Sergey's patch,
> so there must be an irq remapping missing.

Yup, that's it.

> I also test regularly on landisk, where 8139 Ethernet works fine.
> Turns out landisk uses arch/sh/drivers/pci/fixups-landisk.c to fixup
> the irq...

I didn't think the patch was wrong per se, just that something broke when
jiggled. :(

> arch/sh/include/mach-common/mach/r2d.h has:
> #define R2D_FPGA_IRQ_BASE       100
> Subtracting 16 here does not help.
> 
> With this (gmail-whitespace-damaged) patch:
> 
> --- a/arch/sh/drivers/pci/fixups-rts7751r2d.c
> +++ b/arch/sh/drivers/pci/fixups-rts7751r2d.c
> @@ -31,9 +31,9 @@ static char lboxre2_irq_tab[] = {
>  int pcibios_map_platform_irq(const struct pci_dev *pdev, u8 slot, u8 pin)
>  {
>         if (mach_is_lboxre2())
> -               return lboxre2_irq_tab[slot];
> +               return lboxre2_irq_tab[slot] - 16;
>         else
> -               return rts7751r2d_irq_tab[slot];
> +               return rts7751r2d_irq_tab[slot] - 16;
>  }
> 
>  int pci_fixup_pcic(struct pci_channel *chan)
> 
> it no longer crashes, but ifconfig still fails:
> 
> / # ifconfig eth0 up
> ifconfig: ioctl 0x8914 failed: Invalid argument

Sounds like it's now outside of the IRQ range allocation, but I can't find where
that's requested when registering the controller? (What is a "swizzle" anyway?)

I'm looking at kernel/cpu/sh4/setup-sh7750.c but I don't understand why it might
work for landisk but not there. (Bit out of my depth in this plumbing.
Head-scratching at include/linux/sh_intc.h #defining DECLARE_INTC_DESC()... hard
to work backwards to find where this stuff STARTS...)

> Note that there are more implementations of pcibios_map_platform_irq()
> that do not use evt2irq(), and thus are probably broken by this patch.

Yup. Sounds like something could be consolidated. Unfortunately I only have 4
test systems for this platform, only 2 of which are easy to cycle...

> Gr{oetje,eeting}s,
> 
>                         Geert

I can try sticking printk() into this to track it down if you haven't got any
more time to look at it. I don't understand this plumbing very well but "error
return code comes from here, that tested this variable, which was set here..."
is generally a deterministic approach, if glacial.

Rob