Re: [Qemu-devel] Data bus error with redeonfb on mips_fulong2e

[BALATON Zoltan]

Hello,

On Fri, 15 Mar 2019, Philippe Mathieu-Daudé wrote:
> Hi Zoltan,
>
> On 3/14/19 11:08 PM, BALATON Zoltan wrote:
> > On Thu, 14 Mar 2019, Aleksandar Markovic wrote:
> > > > From: BALATON Zoltan <address@hidden>
> > > > Subject: Data bus error with redeonfb on mips_fulong2e
> > > >
> > > > Hello,
> > > >
> > > > Trying to debug the Linux kernel oops with radeonfb I've added some more
> > > > debug logs and got this:
> > > >
> > > > radeonfb_pci_register BEGIN
> > > > pci_host_data: pci_data_read: ati-vga: addr=04 val=00000000 len=4
> > > > PCI: Enabling device 0000:00:06.0 (0000 -> 0003)
> > > > pci_host_data: pci_data_read: ati-vga: addr=04 val=00000000 len=4
> > > > pci_host_data: pci_data_write: ati-vga: addr=04 val=00000003 len=4
> > > > pci_update_mappings: adding bar 0 to pci.mem @ 0x14000000
> > > > pci_update_mappings: adding bar 1 to io @ 0x4000
> > > > pci_update_mappings: adding bar 2 to pci.mem @ 0x15050000
> > > > pci_host_data: pci_data_read: ati-vga: addr=3c val=00000000 len=4
> > > > radeonfb_pci_register fb_base_phys=14000000
> > > > radeonfb_pci_register mmio_base_phys=15050000
> > > > radeonfb_pci_register request mem regions
> > > > radeonfb_pci_register map regions
> > > > radeonfb_pci_register mmio_base=9000000015050000
> > > > Data bus error, epc == ffffffff80418e4c, ra == ffffffff80418e4c
> > > > (full oops dump in previous message but I couldn't find anything
> > > > useful in
> > > > there so omitted here)
>
> (qemu) info mtree
> memory-region: system
>  0000000000000000-ffffffffffffffff (prio 0, i/o): system
>    0000000000000000-000000000fffffff (prio 0, ram): fulong2e.ram
>    0000000010000000-000000001bffffff (prio 0, i/o): pci.mem
>      00000000100a0000-00000000100bffff (prio 1, i/o): vga-lowmem
>      0000000024000000-0000000024ffffff (prio 1, ram): vga.vram
>      0000000025050000-0000000025053fff (prio 1, i/o): ati.mmregs
>    000000001fc00000-000000001fc7ffff (prio 0, rom): fulong2e.bios
>
> The CPU bus is right, there is nothing mapped at 0x0.1505.0000 (IO is
> 36bit wide on mips64).
>
> I look at the kernel and it doesn't configure the PCI but expect the
> display to be there already, so I checked the BIOS you cited in another
> thread and ran the same command con the "PMON>" prompt was reached:
>
> (qemu) info mtree
> memory-region: system
>  0000000000000000-ffffffffffffffff (prio 0, i/o): system
>    0000000000000000-000000000fffffff (prio 0, ram): fulong2e.ram
>    0000000010000000-000000001bffffff (prio 0, i/o): pci.mem
>      00000000100a0000-00000000100affff (prio 2, i/o): alias vga.chain4
> @vga.vram 0000000000000000-000000000000ffff
>      00000000100a0000-00000000100bffff (prio 1, i/o): vga-lowmem
>      0000000014000000-0000000014ffffff (prio 1, ram): vga.vram
>      0000000015000000-000000001503ffff (prio 1, rom): rtl8139.rom
>      0000000015040000-000000001504ffff (prio 1, rom): ati-vga.rom
>      0000000015050000-0000000015053fff (prio 1, i/o): ati.mmregs
>
>              ^ remapped

Thanks a lot for finding this. Indeed this makes sense that we may need to 
do further init of hardware normally done by firmware if we load kernel 
with -kernel option but I haven't thought about that. This explains the 
error we get.

> $ fgrep 'pci_cfg_write ati-vga' trace.log
> ...
> address@hidden:pci_cfg_write ati-vga 06:0 @0x10 <- 0x4000008
> address@hidden:pci_cfg_write ati-vga 06:0 @0x18 <- 0x5050000
> address@hidden:pci_cfg_write ati-vga 06:0 @0x30 <- 0x5040001
> address@hidden:pci_cfg_write ati-vga 06:0 @0x14 <- 0xa101
> address@hidden:pci_cfg_write ati-vga 06:0 @0x4 <- 0x147
>
> PCI_BASE_ADDRESS_0 = 0x4000008
> PCI_BASE_ADDRESS_2 = 0x5050000
> PCI_ROM_ADDRESS    = 0x5040001
> PCI_BASE_ADDRESS_1 = 0xa101
> PCI_COMMAND        = IO|MEMORY|MASTER|PARITY|SERR
>
> I dunno where we have to plug that.

Maybe in  mips_fulong2e.c:main_cpu_reset() where it has a comment:
/* TODO: 2E reset stuff */ within the if (loaderparams.kernel_filename) so 
we only do it when kernel is loaded via -kernel? Can someone come up with 
a patch?

> BTW you should:
>
>     /* GPU */
> -    dev = DEVICE(pci_create(pci_bus, -1, "ati-vga"));
> +    dev = DEVICE(pci_create(pci_bus, PCI_DEVFN(6, 0), "ati-vga"));

I was trying to find info on where the hardware has the ati-vga chip but 
could not find any logs showing this on-line. Now I've noticed there are 
some defines at the beginning of mips_fulong2e.c I haven't found before:

/* PCI SLOT in fulong 2e */
#define FULONG2E_VIA_SLOT        5
#define FULONG2E_ATI_SLOT        6
#define FULONG2E_RTL8139_SLOT    7

so that's probably correct and we could use this define.

Thank you,
BALATON Zoltan