[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 08/28] mips_malta: do not use isa_mmio
From: |
Paolo Bonzini |
Subject: |
Re: [Qemu-devel] [PATCH 08/28] mips_malta: do not use isa_mmio |
Date: |
Wed, 28 Aug 2013 13:30:46 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130805 Thunderbird/17.0.8 |
Il 28/08/2013 13:13, Aurelien Jarno ha scritto:
> On Wed, Aug 28, 2013 at 01:03:01PM +0200, Aurelien Jarno wrote:
>> On Mon, Jul 22, 2013 at 03:54:18PM +0200, Paolo Bonzini wrote:
>>> This fixes endianness bugs in I/O port access.
>>
>> It looks like it actually did the reverse, ie introducing endianness
>> bugs. With this patch, the pcnet-pci NIC (default NIC card) doesn't work
>> any more on big endian Malta, while it still works on little endian
>> Malta.
>>
>> Reverting this commit fixes the issue for the pcnet card, but it makes
>> the IDE controller to fail, likely due to endianness issues.
>>
>> This is reproducible using the following kernel and the following line:
>>
>> http://ftp.debian.org/debian/dists/wheezy/main/installer-mips/current/images/malta/netboot/vmlinux-3.2.0-4-4kc-malta
>> qemu-system-mips -kernel vmlinux-3.2.0-4-4kc-malta -nographic
>>
>> In that case the boot log is:
>>
>> | [ 0.464000] pcnet32: pcnet32.c:v1.35 21.Apr.2008 address@hidden
>> | [ 0.464000] PCI: Enabling device 0000:00:0b.0 (0000 -> 0003)
>> | [ 0.468000] pcnet32: No access methods
>>
>> Without this patch, the boot log is:
>>
>> | [ 0.524000] pcnet32: pcnet32.c:v1.35 21.Apr.2008 address@hidden
>> | [ 0.524000] PCI: Enabling device 0000:00:0b.0 (0000 -> 0003)
>> | [ 0.524000] pcnet32: PCnet/PCI II 79C970A at 0x1020, 52:54:00:12:34:56
>> assigned IRQ 10
>> | [ 0.528000] pcnet32: eth0: registered as PCnet/PCI II 79C970A
>> | [ 0.532000] pcnet32: 1 cards_found
>>
>
> It seems to be due to the fact that the pcnet-pci device is declared as
> NATIVE_ENDIAN. I don't really understand why. Changing it to
> LITTLE_ENDIAN as in the following patch also fixes the problem, but I am
> not sure it is correct.
Yes, it is. See patch 15 for an example of doing the same thing.
Paolo
> diff --git a/hw/net/pcnet-pci.c b/hw/net/pcnet-pci.c
> index a893165..865f2f0 100644
> --- a/hw/net/pcnet-pci.c
> +++ b/hw/net/pcnet-pci.c
> @@ -134,7 +134,7 @@ static void pcnet_ioport_write(void *opaque, hwaddr addr,
> static const MemoryRegionOps pcnet_io_ops = {
> .read = pcnet_ioport_read,
> .write = pcnet_ioport_write,
> - .endianness = DEVICE_NATIVE_ENDIAN,
> + .endianness = DEVICE_LITTLE_ENDIAN,
> };
>
> static void pcnet_mmio_writeb(void *opaque, hwaddr addr, uint32_t val)
> @@ -256,7 +256,7 @@ static const MemoryRegionOps pcnet_mmio_ops = {
> .read = { pcnet_mmio_readb, pcnet_mmio_readw, pcnet_mmio_readl },
> .write = { pcnet_mmio_writeb, pcnet_mmio_writew, pcnet_mmio_writel },
> },
> - .endianness = DEVICE_NATIVE_ENDIAN,
> + .endianness = DEVICE_LITTLE_ENDIAN,
> };
>
> static void pci_physical_memory_write(void *dma_opaque, hwaddr addr,
>