[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: why chipset needn't PERR#?



** Proprietary **

Hi Liu,

I dont know why the perticulat chepset dont have PERR# pin,
but the chipset can detect polling, if there is any parity error found by the traget 
See, in PCI header connfiguration, there is one bit in Status Register "Detected Parity Error"
the chepset might be polling on fixed time to all devices to see whether any divice detected PERR??

OK, but this is not as effective as using PERR# and also it will affect the PCI bandwidth..
have u got any other solution?

regards,
---sushant

>>> " $(AÁõÔÆ·É" <liuyunfei@routerd.com> 11/22/02 08:07AM >>>
Hi,experts
We're using a chipset and I find that there is no PERR# pin within the PCI interface.And in the system the only PCI slot's PERR# pin is left unconnected(but with pull up resistor).
I read in PCI spec that chipset used in mainboard is excluded from the requirement to implement the PERR# signal. According to the explanation of the PCI spec and Tom Shanley's "PCI system architecture",the chipset surely can learn whether parity error occurs when performing reading. But when initiating a write  ,how can the chipset get to know that a parity error is detected by the target? 
The author of "PCI system architecture" explains this by assuming no PCI slots in system and no PERR# generated by PCI device. Obviously,it's just an assumption!Can anybody tell me how does a chipset get the information of detecting of parity errors by other devices?
I'm confused about that and any light threw on me will be appreciated!
Thanks in advance.

Liu Yunfei