On Thu, Oct 27, 2022 at 11:29:48PM +0100, Catalin Marinas wrote: > On Wed, Oct 26, 2022 at 10:46:46AM -0700, Linus Torvalds wrote: > > I think we should just stop bending over backwards over this, and say > > "if your DMA isn't coherent, it's on your driver to mark its > > allocations". > [...] > > That hardware may then be one of the one-off strange cases, but those > > people with their masochistic tendencies can take the pain of "oh, now > > I need to mark my broken driver with dma_alloc()". > > The driver is not necessarily broken. The same small kmalloc() in a USB > driver can work fine on a fully coherent platform but if that chip ends > up on a SoC that doesn't support coherent DMA, it needs bigger kmalloc() > alignment. The driver could check if it's coherent but that's more of an > arch detail that the driver shouldn't care about. If we define a new API > like dma_alloc() and drivers don't use it, that's when we can claim they > are broken. > > A further optimisation would be for dma_alloc() to take a struct device > pointer and check dev_is_dma_coherent() before deciding to align the > size, though this doesn't work when the allocation place cannot tell the > destination device (e.g. alloc_skb(), though these buffers are > cacheline-aligned already). > > Reading up on coccinelle to see if I can make this transition easier. If > not, I'll probably go back to bouncing. bouncing? sparse is your friend here, here's a tiny patch that if you apply and then build the kernel with sparse will show up all the USB driver changes that are needed. (note, sample code only, does not fully work yet as there are no .c changes made). I suggest we add something like this now, work on fixing up all of the drivers for 6.2-rc1, and then you can add the backend allocator changes after that. A few rounds of 'make allmodconfig' will show us the places needing to be fixed up and 0-day will help out with that as well. Yes it's a lot, but it gives us a fighting chance to do the right thing going forward with regards to knowing what "type" of memory needs to be allocated. thanks, greg k-h
