On 5/5/20 8:15 PM, Jared Rossi wrote:
> Remove the explicit prefetch check when using vfio-ccw devices.
> This check does not trigger in practice as all Linux channel programs
> are intended to use prefetch.
>
> It is expected that all ORBs issued by Linux will request prefetch.
> Although non-prefetching ORBs are not rejected, they will prefetch
> nonetheless. A warning is issued up to once per 5 seconds when a
> forced prefetch occurs.
>
> A non-prefetch ORB does not necessarily result in an error, however
> frequent encounters with non-prefetch ORBs indicate that channel
> programs are being executed in a way that is inconsistent with what
> the guest is requesting. While there is currently no known case of an
> error caused by forced prefetch, it is possible in theory that forced
> prefetch could result in an error if applied to a channel program that
> is dependent on non-prefetch.
>
> Signed-off-by: Jared Rossi <jrossi@xxxxxxxxxxxxx>
> ---
> Documentation/s390/vfio-ccw.rst | 6 ++++++
> drivers/s390/cio/vfio_ccw_cp.c | 19 ++++++++++++-------
> 2 files changed, 18 insertions(+), 7 deletions(-)
>
> diff --git a/Documentation/s390/vfio-ccw.rst b/Documentation/s390/vfio-ccw.rst
> index fca9c4f5bd9c..23e7d136f8b4 100644
> --- a/Documentation/s390/vfio-ccw.rst
> +++ b/Documentation/s390/vfio-ccw.rst
> @@ -335,6 +335,12 @@ device.
> The current code allows the guest to start channel programs via
> START SUBCHANNEL, and to issue HALT SUBCHANNEL and CLEAR SUBCHANNEL.
>
> +Currently all channel programs are prefetched, regardless of the
> +p-bit setting in the ORB. As a result, self modifying channel
> +programs are not supported. For this reason, IPL has to be handled as
> +a special case by a userspace/guest program; this has been implemented
> +in QEMU's s390-ccw bios as of QEMU 4.1.
> +
> vfio-ccw supports classic (command mode) channel I/O only. Transport
> mode (HPF) is not supported.
>
> diff --git a/drivers/s390/cio/vfio_ccw_cp.c b/drivers/s390/cio/vfio_ccw_cp.c
> index 3645d1720c4b..d423ca934779 100644
> --- a/drivers/s390/cio/vfio_ccw_cp.c
> +++ b/drivers/s390/cio/vfio_ccw_cp.c
> @@ -8,6 +8,7 @@
> * Xiao Feng Ren <renxiaof@xxxxxxxxxxxxxxxxxx>
> */
>
> +#include <linux/ratelimit.h>
> #include <linux/mm.h>
> #include <linux/slab.h>
> #include <linux/iommu.h>
> @@ -625,23 +626,27 @@ static int ccwchain_fetch_one(struct ccwchain *chain,
> * the target channel program from @orb->cmd.iova to the new ccwchain(s).
> *
> * Limitations:
> - * 1. Supports only prefetch enabled mode.
> - * 2. Supports idal(c64) ccw chaining.
> - * 3. Supports 4k idaw.
> + * 1. Supports idal(c64) ccw chaining.
> + * 2. Supports 4k idaw.
> *
> * Returns:
> * %0 on success and a negative error value on failure.
> */
> int cp_init(struct channel_program *cp, struct device *mdev, union orb *orb)
> {
> + static DEFINE_RATELIMIT_STATE(ratelimit_state, 5 * HZ, 1);
This looks peculiar, being static and on the stack. But I guess it is
fine. And as Conny mentions, a comment about the boot messages would be
good.
> int ret;
>
> /*
> - * XXX:
> - * Only support prefetch enable mode now.
> + * We only support prefetching the channel program. We assume all channel
> + * programs executed by supported guests (i.e. Linux) likewise support
s/(i.e. Linux) //
> + * prefetching.
Even if prefetching is not specified the channel program
> + * is still executed using prefetch.
The above sentence seems redundant, and can be removed.
Executing a channel program that
> + * does not specify prefetching will typically not cause an error, but a
> + * warning is issued to help identify the problem if something does break.
> */
> - if (!orb->cmd.pfch)
> - return -EOPNOTSUPP;
> + if (!orb->cmd.pfch && __ratelimit(&ratelimit_state))
> + dev_warn(mdev, "executing channel program with prefetch, but prefetch isn't specified");
Works well enough (with QEMU patch, obviously).
Almost-r-b: me :)
>
> INIT_LIST_HEAD(&cp->ccwchain_list);
> memcpy(&cp->orb, orb, sizeof(*orb));
>
