Hi,
For the communication between user space and kernel
space, which one would give more performance - netlink
or ioctl.
When used ioremap_nocache to map from user space to
kernel space does it do cache flush ?
Any help is highly appreciated.
thanks in advance
regards
aryan
__________________________________________________
Do You Yahoo!?
Yahoo! Sports - sign up for Fantasy Baseball
http://sports.yahoo.com
aryan aru wrote:
> When used ioremap_nocache to map from user space to
> kernel space does it do cache flush ?
I'd guess ioremap_nocache marks the area as uncacheable,
so that all memory accesses bypass the cache and go
directly to physical memory. best way to find out
is to check the code ;)
-Bryan
Hi,
I am new to this device driver area. I have a question
on PCI DMA usage. Could you plz correct me.
I have two processors, processor(pci device) pA,
processor(pci device) pB. I want to use DMA pull
mechanism for transfering the messages between them. I
need to follow the mechanism for this.
When pA wants to send a pkt to pB:
pA places the address of the src_buff in one of the
common registers. pB, by accessing that register, will
come to know the location of the src_buffer. Now pB,
using its DMA controller has to pull the message to
its local buffer, say dest_loc_buff. For setting up
the dma controller on pB, I need the pci_addr of
src_buuf and dest_loc_buff( DMA controller of
MPC82xx).
How can I get pci_addr of src_buff?
Can I get this by using pci_map_single?
>From my understating pci_map_single takes local buffer
as the argument.
In this transfer, do both the DMA controllers (pA and
pB) involve in message transfer or only pB pulls the
pkt.
Any driver example available in that net for this
"pull" mechanism.
Any help is highly appreciated.
thans and regards
Aryan
__________________________________________________
Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup
http://fifaworldcup.yahoo.com