Re: [PATCH 1/1] Userspace I/O (UIO): Add support for userspace DMA (corrected)

On Wed, Dec 03, 2008 at 05:51:30PM -0800, Edward Estabrook wrote:
From: Edward Estabrook <Edward_Estabrook@xxxxxxxxxxx>

Here is a patch that adds the ability to dynamically allocate and use
coherent DMA
from userspace by extending the Userspace IO driver. This patch applies against
2.6.28-rc6.

Hi Edward,
sorry I didn't answer sooner. It was not lack of interest in your work, I
just had to think about it (and had quite a lot of other work, too).
I consider your idea of having DMA coherent memory allocatable from userspace
a valid demand. However, the interface should be implemented differently.

As a general requirement, all information you need to pass to userspace has
to go through sysfs. This is a UIO design decision and should not be changed
lightly. Please don't fill information into the allocated memory.

We shouldn't mix dynamically allocated DMA regions with the already existing
static memory mappings. There should be an extra directory, maybe called
/sys/class/uio/uioN/dyn_dma_mem/ containing information about this kind of
DMA memory. This directory would not exist by default, it's only created if
some user actually allocates such memory.

In the UIO core, these mappings should be handled with a dynamic list, not
a static array. The number of possible mappings has to be limited to a
sensible maximum, maybe 50 or 100. The code handling the existing static
mappings could mostly stay as-is.

Userspace can allocate new DMA memory by using some magic number as the offset
parameter of mmap() like you suggested. However, this magic should be just one
of the bits 16..31 (e.g. 0x00010000) to have other bits available for future
extensions.

I think you should also reconsider the way you unmap the memory. It's certainly
right to unmap everything when the device is unregistered, but I think there's
more to think about:
- Should the memory be freed as soon as all users have the device closed?
- What happens if a user calls munmap() on such a mapping?

I'm looking forward to your next version.

Thanks,
Hans

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Relevant Pages

  • Re: Sharing memory between kernelspace and userspace
    ... deallocate, on a totally dynamic basis, userspace ... Let userspace allocate shared memory visible to multiple ... and pass that into the kernel for it to write to. ...
    (Linux-Kernel)
  • Re: NdisAllocateMemoryWithTag(Priority) -> Does it ensure Congtiguous physical memory by default
    ... I believe in W2K8 if your device does 64 bit DMA, when you call NdisMAllocateSharedMemory you will get high physical addresses before getting low addresses. ... NDIS does not provide an API to allocate physically contiguous memory outside what HAL DMA APIs provide. ...
    (microsoft.public.development.device.drivers)
  • Re: [PATCH 1/1] Userspace I/O (UIO): Add support for userspace DMA (corrected)
    ... Here is a patch that adds the ability to dynamically allocate and use ... from userspace by extending the Userspace IO driver. ... and map a new DMA region on demand. ... functions in the core that can help a driver doing this. ...
    (Linux-Kernel)
  • Re: 1G DMA buffer under XP - how?
    ... system memory, and what is address 0 to the controller could be something ... so only the HAL's DMA routines can allocate ... common buffer correctly in all situations. ... My card is happy with any physical address> below 4G, and I can tell this call to only allocate below the given> HighestAcceptableAddress. ...
    (microsoft.public.development.device.drivers)
  • How to handle a specific DMA configuration ?
    ... My physical memory mapping is a bit special: ... last 128MB are usable for DMA. ... My problem is the allocation of skbuff. ... even if a I allow Linux to allocate memory between ...
    (Linux-Kernel)