[add some people who might know] On Tue, Dec 13, 2016 at 4:20 AM, David Laight <David.Laight@xxxxxxxxxx> wrote: > From: Andy Lutomirski >> Sent: 12 December 2016 20:53 >> The driver put a constant buffer of all zeros on the stack and >> pointed a scatterlist entry at it in two places. This doesn't work >> with virtual stacks. Use a static 16-byte buffer of zeros instead. > ... > > I didn't think you could dma from static data either. According to lib/dma-debug.c, you can't dma to or from kernel text or rodata, but you can dma to or from kernel bss or data. So empty_zero_page should be okay, because it's not rodata right now. But I think this is rather silly. Joerg, Linus, etc: would it be okay to change lib/dma-debug.c to allow DMA *from* rodata? After all, rodata is ordinary memory, is backed by struct page, etc. And DMA from the zero page had better be okay because I think it happens if you mmap some zeros, don't write to them, and then direct I/O them to a device. Then I could also move empty_zero_page to rodata. --Andy -- To unsubscribe from this list: send the line "unsubscribe linux-crypto" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html