Re: [PATCH] virtio-console: avoid DMA from vmalloc area

On Wed, Jul 28, 2021 at 10:28 AM Xianting Tian
[off-list ref] wrote:

在 2021/7/28 下午3:25, Arnd Bergmann 写道:

I checked several hvc backends, like drivers/tty/hvc/hvc_riscv_sbi.c,
drivers/tty/hvc/hvc_iucv.c, drivers/tty/hvc/hvc_rtas.c, they don't use dma.

I not finished all hvc backends check yet. But I think even if all hvc
backends don't use dma currently, it is still possible that the hvc
backend using dma will be added in the furture.

So I agree with you it should better be fixed in the hvc framework,
solve the issue in the first place.

Ok, sounds good to me, no need to check more backends then.
I see the hvc-console driver is listed as 'Odd Fixes' in the maintainer
list, with nobody assigned other than the ppc kernel list (added to Cc).

Once you come up with a fix in hvc_console.c, please send that to the
tty maintainers, the ppc list and me, and I'll review it.

        Arnd

在 2021/7/28 下午5:01, Arnd Bergmann 写道:

On Wed, Jul 28, 2021 at 10:28 AM Xianting Tian
[off-list ref] wrote:

quoted

在 2021/7/28 下午3:25, Arnd Bergmann 写道:

I checked several hvc backends, like drivers/tty/hvc/hvc_riscv_sbi.c,
drivers/tty/hvc/hvc_iucv.c, drivers/tty/hvc/hvc_rtas.c, they don't use dma.

I not finished all hvc backends check yet. But I think even if all hvc
backends don't use dma currently, it is still possible that the hvc
backend using dma will be added in the furture.

So I agree with you it should better be fixed in the hvc framework,
solve the issue in the first place.

Ok, sounds good to me, no need to check more backends then.
I see the hvc-console driver is listed as 'Odd Fixes' in the maintainer
list, with nobody assigned other than the ppc kernel list (added to Cc).

Once you come up with a fix in hvc_console.c, please send that to the
tty maintainers, the ppc list and me, and I'll review it.

OK, thanks, I will submit the patch ASAP :)

         Arnd