non cached memory for user space DMA

I am trying to write a MAC driver for Arm 11 MPcore in 2.6.23. The DMA'ble memory required is supposed to be non-cached for performance reasons.

A chunk of memory(~ 6MB) is allocated at startup using bigphys area patch. This physical area in the RAM is mmap'd to userspace as follows and thereafter carved into network buffers:

fd = open("/dev/mem", O_RDWR|O_SYNC);
uptr = mmap(0, sz, PROT_READ|PROT_WRITE, MAP_SHARED|MAP_LOCKED,fd, pptr);

Inspite of using the O_SYNC flag the user space mapping is cached and the DMA'd data is not visible. I tried writing up my own mmap and marked the vma non-cached without any results:

size_t size = vma->vm_end - vma->vm_start;
unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;

vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
if (remap_pfn_range(vma, vma->vm_start, offset, size, vma->vm_page_prot)) {
return -EAGAIN;
}

I found that the bigphys patch allocated memory from bootmem and bootmem maps the memory as MT_MEMORY. Is there a way I could reduce bootmem allocation at startup (using kernel boot cmd mem=x ?) ?

The idea is to map the rest of the available memory as MT_DEVICE(assuming it removes the BUFFERABLE & CACHEABLE bits while mapping) so that the kernel mapping is also non-cacheable. Hopefully, this will ease the userspace mapping to non-cached as well ?

thanks,
Deb Rupam Banerjee

Webcast
How to Build an Optimal Hadoop Cluster to Store and Maintain Unlimited Amounts of Data Using Microservers

Realizing the promise of Apache® Hadoop® requires the effective deployment of compute, memory, storage and networking to achieve optimal results. With its flexibility and multitude of options, it is easy to over or under provision the server infrastructure, resulting in poor performance and high TCO. Join us for an in depth, technical discussion with industry experts from leading Hadoop and server companies who will provide insights into the key considerations for designing and deploying an optimal Hadoop cluster.

Learn More

Sponsored by AMD

White Paper
Red Hat White Paper: Using an Open Source Framework to Catch the Bad Guy

Built-in forensics, incident response, and security with Red Hat Enterprise Linux 6

Every security policy provides guidance and requirements for ensuring adequate protection of information and data, as well as high-level technical and administrative security requirements for a system in a given environment. Traditionally, providing security for a system focuses on the confidentiality of the information on it. However, protecting the data integrity and system and data availability is just as important. For example, when processing United States intelligence information, there are three attributes that require protection: confidentiality, integrity, and availability.

Learn more about catching the bad guy in this free white paper.

Learn More

Sponsored by DLT Solutions