[PATCH v3] getrandom system call wrapper [BZ #17252]

[Zack Weinberg]

On Wed, Oct 5, 2016 at 3:18 AM, Florian Weimer <fweimer@redhat.com> wrote:
> On 10/04/2016 10:30 PM, Zack Weinberg wrote:
>> Fundamentally the problem I have with it is that getrandom must always
>> succeed; applications cannot realistically be expected to recover from
>> failure.
>
> The current patch is the consensus which emerged after the first patches.
> In particular, you wrote a while back, “I think it would be better to expose
> getrandom() as a completely unadorned syscall.”

But note that an unadorned syscall is *not* a cancellation point.

> As a result of this community-support decision, ENOSYS due to lack of kernel
> support is always a possibility.  I think this was pretty clear at the time,
> when the emulation based on device nodes was rejected.

I would have supported the emulation based on device nodes if I had
thought there was any way to make it reliable enough that it would not
cause more problems than it solved.

zw