2011/9/3 Wolfgang Grandegger <[email protected]>

> Hi Willy,
>
> On 08/30/2011 01:36 PM, Willy Lambert wrote:
> > 2011/8/30 Wolfgang Grandegger <[email protected]>
> >
> >> On 08/30/2011 11:18 AM, Willy Lambert wrote:
> >>> 2011/8/18 Marc Kleine-Budde <[email protected]>
> >>>
> >>>> On 08/18/2011 07:55 AM, Oliver Hartkopp wrote:
> >>>>> On 16.08.2011 21:45, Willy Lambert wrote:
> >>>>>
> >>>>>> Hi all,
> >>>>>>
> >>>>>> I used sja1000 on a 32bits system and I am trying to switch to a 64b
> >>>> system.
> >>>>>> It not possible to select CONFIG_SJA1000_ISA when CONFIG_ISA is not
> >> set,
> >>>> which
> >>>>>> is not available when CONFIG_X86_64 is set.
> >>>>>>
> >>>>>> As in xenomai "socket" drivers I am able to use my board, I wonder
> why
> >>>> this is
> >>>>>> not possible under a vanilla linux. I access my HW by iomem (the
> >>>> equivalent
> >>>>>> xenomai driver is xeno_can_mem), is there an equivalent socket can
> >>>> driver ?
> >>>>>
> >>>>>
> >>>>> Hi Willy,
> >>>>>
> >>>>> i've looked into
> >>>>>
> >>>>> http://lxr.linux.no/#linux+v3.0.3/drivers/tty/serial/8250.c
> >>>>>
> >>>>> which should have a similar problem, as these old-style serial
> hardware
> >>>> is
> >>>>> also using the ISA bus.
> >>>>>
> >>>>> But the serial drivers are now depending on HAS_IOMEM
> >>>>>
> >>>>> http://lxr.linux.no/#linux+v3.0.3/drivers/tty/serial/Kconfig#L4
> >>>>>
> >>>>> and they use platform calls
> >>>>>
> >>>>> http://lxr.linux.no/#linux+v3.0.3/drivers/tty/serial/8250.c#L3177
> >>>>>
> >>>>> like platform_device_alloc(), platform_device_add(),
> >>>>> platform_driver_register() and friends ...
> >>>>>
> >>>>> http://lxr.linux.no/#linux+v3.0.3/drivers/tty/serial/8250.c#L3349
> >>>>>
> >>>>> The sja1000_isa.c still uses the ISA pendants
> >>>>>
> >>>>>
> >>>>
> >>
> http://lxr.linux.no/#linux+v3.0.3/drivers/net/can/sja1000/sja1000_isa.c#L253
> >>>>>
> >>>>> So IMO the sja1000_isa driver 'just' needs to be converted like they
> >> did
> >>>> with
> >>>>> the serial drivers.
> >>>>>
> >>>>> @Wolfgang/Marc: Or is it better to create some platform bindings for
> >> x86
> >>>> and
> >>>>> use the 'sja1000_platform' driver??
> >>>>
> >>>> A platform driver is of course sexier :)
> >>>>
> >>>> Marc
> >>>>
> >>>
> >>> Does someone know what is to do in order to make this sja1000_platform
> >>> driver in my case ?
> >>
> >> I think we have first to fix the HAS_ISA vs. HAS_IOMEM issue. The
> >> sja1000 driver supports both, I/O port and memory mapped accesses. For
> >> the first we need HAS_ISA and for the latter HAS_IOMEM... that's my
> >> current understanding. I will have a closer look later today or
> tomorrow.
> >>
> >
> > Thanks in advance, I am using IOMEM, so this fix should be directly
> usefull
> > for me.
>
> What CAN hardware are you using?


Motherboad and CAN stack :
http://www.advantech.com/products/PCM-3362/mod_2B69DA4C-D506-4AEA-8CC3-F6E58E3CBC87.aspx
http://www.ixxat.com/pc_i_04_104_en.html



> And what bus is it
> connected to? Why do I ask? In "arch/x86/Kconfig" there is:
>
>  # x86_64 have no ISA slots, but can have ISA-style DMA.
>  config ISA_DMA_API
>        bool "ISA-style DMA support" if (X86_64 && EXPERT)
>        default y
>        help
>          Enables ISA-style DMA support for devices requiring such
> controllers.
>          If unsure, say Y.
>
>  if X86_32
>
>  config ISA
>        bool "ISA support"
>        ---help---
>          Find out whether you have ISA slots on your motherboard.  ISA is
> the
>          name of a bus system, i.e. the way the CPU talks to the other
> stuff
>          inside your box.  Other bus systems are PCI, EISA, MicroChannel
>          (MCA) or VESA.  ISA is an older system, now being displaced by
> PCI;
>          newer boards don't support it.  If you have ISA, say Y, otherwise
> N.
>
>
> Maybe the assumption that the x86_64 does not have ISA slots is wrong and
> we
> should fix that instead of moving to a platform driver like Oliver
> suggested.
>
> Wolfgang.
>
_______________________________________________
Socketcan-users mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/socketcan-users

Reply via email to