mmap'd physical memory not updating user virtual
I am writing a user space MAC driver in Linux 2.6.23 hosted on a Arm11 mpcore. A contiguous block of physical memory is allocated and mmap'd to userspace.
/dev/mem with O_SYNC flag is used to mmap so that the user virtual memory is uncached (with no much benefit).
Kernel
------
kptr = kmalloc(sz, GFP_DMA|GFP_KERNEL);
pptr = __pa(kptr);
User
----
fd = open("/dev/mem", O_RDWR|O_SYNC);
uptr = mmap(0, sz, PROT_READ|PROT_WRITE, MAP_SHARED|MAP_LOCKED,fd, pptr);
The map'd user virtual memory is carved into network buffers and used for DMA.The MAC is successfully able to DMA ethernet frames to the physical memory. From the Rx interrupt handler bh (kernel logical memory) the frames received could be viewed and the cache functions also work fine. However, the mmap'd memory in user space does not reflect the changes.
Why is the physical memory mapped to user virtual does not reflect the chages ?
Or any suggestions on what could be going wrong here ?
thanks,
Deb Rupam Banerjee
P.S: Please note that all buffer alignment's are taken care of.
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.
Sponsored by AMD
If you already use virtualized infrastructure, you are well on your way to leveraging the power of the cloud. Virtualization offers the promise of limitless resources, but how do you manage that scalability when your DevOps team doesn’t scale? In today’s hypercompetitive markets, fast results can make a difference between leading the pack vs. obsolescence. Organizations need more benefits from cloud computing than just raw resources. They need agility, flexibility, convenience, ROI, and control.
Stackato private Platform-as-a-Service technology from ActiveState extends your private cloud infrastructure by creating a private PaaS to provide on-demand availability, flexibility, control, and ultimately, faster time-to-market for your enterprise.
Sponsored by ActiveState
| Non-Linux FOSS: libnotify, OS X Style | Jun 18, 2013 |
| Containers—Not Virtual Machines—Are the Future Cloud | Jun 17, 2013 |
| Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer | Jun 12, 2013 |
| Weechat, Irssi's Little Brother | Jun 11, 2013 |
| One Tail Just Isn't Enough | Jun 07, 2013 |
| Introduction to MapReduce with Hadoop on Linux | Jun 05, 2013 |
- Containers—Not Virtual Machines—Are the Future Cloud
- Non-Linux FOSS: libnotify, OS X Style
- Linux Systems Administrator
- Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer
- Validate an E-Mail Address with PHP, the Right Way
- Senior Perl Developer
- Technical Support Rep
- UX Designer
- Introduction to MapReduce with Hadoop on Linux
- RSS Feeds
- user namespaces
1 hour 17 min ago - yea
1 hour 43 min ago - One advantage with VMs
4 hours 12 min ago - about info
4 hours 45 min ago - info
4 hours 46 min ago - info
4 hours 47 min ago - info
4 hours 49 min ago - info
4 hours 50 min ago - abut info
4 hours 52 min ago - info
4 hours 53 min ago
Featured Jobs
| Linux Systems Administrator | Houston and Austin, Texas | Host Gator |
| Senior Perl Developer | Austin, Texas | Host Gator |
| Technical Support Rep | Houston and Austin, Texas | Host Gator |
| UX Designer | Austin, Texas | Host Gator |
| Web & UI Developer (JavaScript & j Query) | Austin, Texas | Host Gator |
Free Webinar: Hadoop
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.
Some of key questions to be discussed are:
- What is the “typical” Hadoop cluster and what should be installed on the different machine types?
- Why should you consider the typical workload patterns when making your hardware decisions?
- Are all microservers created equal for Hadoop deployments?
- How do I plan for expansion if I require more compute, memory, storage or networking?



Hi Deb, How did you solve
Hi Deb,
How did you solve this issue? I am also facing similar problem with ARM9. I have reserved 4M Bytes of memory at boot time, for using with the DMA and using the following code:
#define DMAMEM_START 0x73C00000
dma_fd = open("/dev/mem", O_RDWR | O_SYNC);
dma_mem = mmap(0, MSP2_DEV_SIZE, PROT_READ | PROT_WRITE, MAP_SHARED, dma_fd, DMAMEM_START);
DMA_write(DMAMEM_START, INST_Add, 16);
err = msync(dma_mem, (size_t) 64, MS_SYNC | MS_INVALIDATE);
DMA_read(INST_Add, DMAMEM_START+32, 16);
Using DMA_READ, I can see the data at DMAMEM_START, but sometime it is correct sometime incorrect.
I appreciate your step by step help!
Thanks,
Aa
DEVMEM
Mitch,
Thanks for your comments. I was able to solve the problem as follows:
I was using the cache (flush/inv) routines used on the kernel logical address and was able to see the DMA'd data. The same idea I used for the user virtual mapping and the DMA'd data was visible in the user space mapping.
thanks,
Deb
DEVMEM
Presumably mmap() is returning a valid pointer, correct? And you're not seeing any other errors occur, correct? You just never see the data change in the mapped area.
Does your kernel have CONFIG_STRICT_DEVMEM enabled?
Mitch Frazier is an Associate Editor for Linux Journal.