keyctl_capabilities and buflen

keyctl_capabilities and buflen

[Ben Boeckel]

Hi,

I'm looking at the implementation of `keyctl_capabilities` in the `next`
branch and it's not clear if `buflen` is the size of the memory pointed
to by `buffer` in array size or in byte size by its documentation. It
seems to be the size in bytes by the code (the &3) could probably
warrant a comment to make it clearer that it is related to
sizeof(unsigned int) here). This seems very odd to me, or is it just a
common kernel pattern I haven't seen before?

Thanks,

--Ben

Re: keyctl_capabilities and buflen

[Ben Boeckel]

On Wed, Jun 05, 2019 at 09:02:47 -0400, Ben Boeckel wrote:
> I'm looking at the implementation of `keyctl_capabilities` in the `next`
> branch and it's not clear if `buflen` is the size of the memory pointed
> to by `buffer` in array size or in byte size by its documentation. It
> seems to be the size in bytes by the code (the &3) could probably
> warrant a comment to make it clearer that it is related to
> sizeof(unsigned int) here). This seems very odd to me, or is it just a
> common kernel pattern I haven't seen before?

I see that this is now `unsigned char*` instead of `unsigned int` on
`next` now and in the kernel by the newest patch submission.

Thanks,

--Ben

Re: keyctl_capabilities and buflen

[David Howells]

Ben Boeckel <mathstuf@gmail.com> wrote:

> I see that this is now `unsigned char*` instead of `unsigned int` on
> `next` now and in the kernel by the newest patch submission.

If you're okay with the patch, could you reply to it with a Reviewed-by?

Thanks,
David

Re: keyctl_capabilities and buflen

[Ben Boeckel]

On Wed, Jun 19, 2019 at 15:56:15 +0100, David Howells wrote:
> If you're okay with the patch, could you reply to it with a Reviewed-by?

I'm not confident enough in the kernel side to give such a tag, sorry.
I'm more interested in the libkeyutils patches which tend to just show
up in the repo rather than for review on the list (where I'd happily
give such tags where I can).

--Ben

Re: keyctl_capabilities and buflen

[David Howells]

Ben Boeckel <mathstuf@gmail.com> wrote:

> I'm not confident enough in the kernel side to give such a tag, sorry.
> I'm more interested in the libkeyutils patches which tend to just show
> up in the repo rather than for review on the list (where I'd happily
> give such tags where I can).

Okay, thanks.

David