On Mon, Apr 09, 2018 at 01:37:11PM +0300, Sagi Grimberg wrote:
>
> > --- a/drivers/nvme/host/fabrics.c
> > +++ b/drivers/nvme/host/fabrics.c
> > @@ -493,7 +493,7 @@ EXPORT_SYMBOL_GPL(nvmf_should_reconnect)
> > */
> > int nvmf_register_transport(struct nvmf_transport_ops *ops)
> > {
> > - if (!ops->create_ctrl)
> > + if (!ops->create_ctrl || !ops->module)
> > return -EINVAL;
> > down_write(&nvmf_transports_rwsem);
>
> Hi Greg,
>
> I think that this part broke builtin compilation of nvme over fabrics
> code.
>
> This was later fixed by Christoph in:
> --
> commit 5a1e59533380a3fd04593e4ab2d4633ebf7745c1
> Author: Christoph Hellwig <hch@xxxxxx>
> Date: Thu Feb 22 07:24:08 2018 -0800
>
> nvme-fabrics: don't check for non-NULL module in nvmf_register_transport
>
> THIS_MODULE evaluates to NULL when used from code built into the kernel,
> thus breaking built-in transport modules. Remove the bogus check.
>
> Fixes: 0de5cd36 ("nvme-fabrics: protect against module unload during
> create_ctrl")
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> Reviewed-by: Sagi Grimberg <sagi@xxxxxxxxxxx>
> Reviewed-by: Johannes Thumshirn <jthumshirn@xxxxxxx>
> Signed-off-by: Keith Busch <keith.busch@xxxxxxxxx>
>
> diff --git a/drivers/nvme/host/fabrics.c b/drivers/nvme/host/fabrics.c
> index 5dd4ceefed8f..a1c58e35075e 100644
> --- a/drivers/nvme/host/fabrics.c
> +++ b/drivers/nvme/host/fabrics.c
> @@ -493,7 +493,7 @@ EXPORT_SYMBOL_GPL(nvmf_should_reconnect);
> */
> int nvmf_register_transport(struct nvmf_transport_ops *ops)
> {
> - if (!ops->create_ctrl || !ops->module)
> + if (!ops->create_ctrl)
> return -EINVAL;
>
> down_write(&nvmf_transports_rwsem);
> --
>
> So I'd suggest taking that as well.
Many thanks for letting me know, I've now queued up that patch as well.
greg k-h