Experimental Physics and
Industrial Control System

"Rees, NP (Nick)" <[email protected]> · Mon, 22 Jun 2009 09:34:39 +0100

LIU Ping,

We use a lot of Hytec on our 5500's and have the fololowing changes in mv5500A.h and config.h:

#define VME_A32_MSTR_BUS        0x00800000  /* This is in mv5500A.h */

#define VME_A32_MSTR_BUS      0x00800000 /* This is in config.h */

I hope this helps

Cheers,

Nick Rees
Principal Software Engineer           Phone: +44 (0)1235-778430
Diamond Light Source                  Fax:   +44 (0)1235-446713 

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of liuping
> Sent: 22 June 2009 04:42
> To: [email protected]
> Subject: re: "A32 RegisterAddress error" about Hytec8401
> 
> Yes, it is the first time to use a card that responds in the 
> VME A32 space
> for me. My VME CPU is mvme5500. I find out some information about
> VME_A32_MSTR_BUS and VME_A32_MSTR_SIZE in the files config.h 
> and mv5500A.h .
> 
> The information showed below was taken from mv5500A.h:
> /* VME A32 outbound (master)  - note ...MSTR_SIZE defined in 
> "config.h" */
> 
> #define VME_A32_MSTR_LOCAL      ((VME_A24_MSTR_LOCAL - 
> VME_A32_MSTR_SIZE)
> &\				  0xfff00000)
> 
> #define VME_A32_MSTR_BUS	0x08000000
> 
> /* VME A32 Inbound (slave) mapping for node 0 only - maps all 
> of DRAM */
> 
> #define VME_A32_SLV_BUS		VME_A32_MSTR_BUS
> #define VME_A32_SLV_LOCAL	LOCAL_MEM_LOCAL_ADRS
> #define VME_A32_SLV_SIZE  (((ULONG)sysPhysMemTop()) -
> (LOCAL_MEM_LOCAL_ADRS))
> 
> #define VME_MEM_LOCAL_START	VME_A32_MSTR_LOCAL
> #define VME_MEM_LOCAL_SIZE	(VME_MEM_LOCAL_END - 
> VME_MEM_LOCAL_START)
> 
> The information showed below was taken from config.h:
> /** VME Memory Mapping size Definitions  - note:  See "mv5500A.h" for
> commentary
>  * regarding PCI-VME and VME-PCI mappings. */
> 
> #define VME_A32_MSTR_SIZE       0x08000000      /* 128 MB */
> #define VME_A24_MSTR_SIZE       0x01000000      /* 16 MB */
> #define VME_A16_MSTR_SIZE	0x00010000	/* 64k */
> 
> Is there anything wrong?
> By the way, without the experience of building vxWorks, 
> setting up EPICS
> base and writing any device' driver, how can I do deeply to 
> understand the
> working mechanism of EPICS? My biggest aim is to learn how to 
> write driver
> codes for new hardware, even I know it is a very difficult road for a
> beginner.
> 
> Best wishes,
> 
> LIU Ping 
> --------------------------------------------------------------
> -----------
> From: Andrew Johnson [mailto:[email protected]] 
> Send: 2009年6月16日 22:28
> to: [email protected]
> to: Michael Abbott; liuping
> subject: Re: "A32 RegisterAddress error" about Hytec8401
> 
> Hi Liu Ping,
> 
> On Tuesday 16 June 2009 03:39:57 Michael Abbott wrote:
> > However, the problem is nothing to do with the 8401...
> 
> I agree with Michael, this is not a problem with the 8401 
> driver and it may 
> not be one with the 8002 driver either.
> 
> Is this the first time you have tried to use a card that 
> responds in the VME
> 
> A32 space?  If so I would guess that the problem is that your 
> vxWorks BSP 
> needs to be configured to map the relevent portion of VME A32 
> space into the
> 
> CPU's memory map.  The error message is occurring because the 
> driver is 
> asking the vxWorks BSP to give it the CPU address corresponding to a 
> particular VME A32 address, but the BSP is replying that 
> there is no such
> CPU 
> address because the BSP is not configuring the VME Bridge 
> chip with any A32 
> Master windows that cover that range.
> 
> Read the target.txt file in your target/config/<bsp> 
> directory, it should 
> document the window mappings.  Look for the VME_A32_MSTR_BUS and 
> VME_A32_MSTR_SIZE variables; you may be able to just modify 
> them and rebuild
> 
> the BSP, but don't just make the window as large as possible 
> because that
> can 
> significantly slow down the startup of the vxWorks image at 
> boot time and 
> also use up memory in MMU page tables.
> 
> I can't give you any specific advice because I don't know 
> what VME CPU board
> 
> you are using, and I only know about a limited number of BSPs anyway.
> 
> - Andrew
> -- 
> The best FOSS code is written to be read by other humans -- 
> Harold Welte
> 
> 
> 
-- 
Scanned by iCritical.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System