Messages in this thread | | | Date | Fri, 21 Nov 2008 18:03:21 +0100 | From | Ingo Molnar <> | Subject | Re: [PATCH 0/10] DMA-API debugging facility |
| |
* Joerg Roedel <joerg.roedel@amd.com> wrote:
> On Fri, Nov 21, 2008 at 04:54:52PM +0000, David Woodhouse wrote: > > On Fri, 2008-11-21 at 17:26 +0100, Joerg Roedel wrote: > > > this patchset introduces code to debug drivers usage of the DMA-API. > > > Tests with hardware IOMMUs have shown several bugs in drivers > > > regarding the usage of that API. > > > Problems were found especially in network card drivers. > > > > This is really useful -- but surely it shouldn't be x86-specific? > > > > All the code except the hooks in the architecture's dma_map_single() et > > al functions could be generic, couldn't it? > > Yes, in principle we could move most of it to generic code. There is > nothing architecture specific in it. Anybody who prefers this to be > arch/x86 before moving it to lib/?
yeah, we want to make it generic once it works.
but my comments about the allocation needs to be addressed (see my comments on [03/10]), and solving that will likely impact the structure of the approach in a way that will generalize it anyway, as a side-effect.
Ingo
| |