On Mon, May 31, 2021 at 09:39:20PM +0200, Klaus Jensen wrote:
> On May 31 15:42, Niklas Cassel wrote:
> > On Fri, May 28, 2021 at 01:22:38PM +0200, Klaus Jensen wrote:
> > > On May 28 11:05, Niklas Cassel wrote:
> > > > From: Niklas Cassel <niklas.cas...@wdc.com>
> > > >
> > > > In the Zoned Namespace Command Set Specification, chapter
> > > > 2.5.1 Managing resources
> > > >
> > > > "The controller may transition zones in the ZSIO:Implicitly Opened state
> > > > to the ZSC:Closed state for resource management purposes."
> > > >
> > > > The word may in this sentence means that automatically transitioning
> > > > an implicitly opened zone to closed is completely optional.
> > > >
> > > > Add a new parameter so that the user can control if this automatic
> > > > transitioning should be performed or not.
> > > >
> > > > Being able to control this can help with verifying that e.g. a
> > > > user-space
> > > > program behaves properly even without this optional ZNS feature.
> > > >
> > > > The default value is set to true, in order to not change the existing
> > > > behavior.
> > > >
> > > > Signed-off-by: Niklas Cassel <niklas.cas...@wdc.com>
> > > > ---
> > > > hw/nvme/ctrl.c | 9 ++++++++-
> > > > hw/nvme/ns.c | 2 ++
> > > > hw/nvme/nvme.h | 1 +
> > > > 3 files changed, 11 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
> > > > index 40a7efcea9..d00f0297a5 100644
> > > > --- a/hw/nvme/ctrl.c
> > > > +++ b/hw/nvme/ctrl.c
> > > > @@ -141,6 +141,11 @@
> > > > *
> > > > * zoned.cross_read=<enable RAZB, default: false>
> > > > * Setting this property to true enables Read Across Zone
> > > > Boundaries.
> > > > + *
> > > > + * zoned.auto_transition=<enable auto resource management,
> > > > default: true>
> > > > + * Indicates if zones in zone state implicitly opened can be
> > > > + * automatically transitioned to zone state closed for resource
> > > > + * management purposes.
> > > > */
> > > >
> > > > #include "qemu/osdep.h"
> > > > @@ -1699,7 +1704,9 @@ static uint16_t nvme_zrm_open_flags(NvmeNamespace
> > > > *ns, NvmeZone *zone,
> > > > /* fallthrough */
> > > >
> > > > case NVME_ZONE_STATE_CLOSED:
> > > > - nvme_zrm_auto_transition_zone(ns);
> > > > + if (ns->params.auto_transition_zones) {
> > > > + nvme_zrm_auto_transition_zone(ns);
> > > > + }
> > > > status = nvme_aor_check(ns, act, 1);
> > > > if (status) {
> > > > return status;
> > > > diff --git a/hw/nvme/ns.c b/hw/nvme/ns.c
> > > > index 3fec9c6273..31dee43d30 100644
> > > > --- a/hw/nvme/ns.c
> > > > +++ b/hw/nvme/ns.c
> > > > @@ -531,6 +531,8 @@ static Property nvme_ns_props[] = {
> > > > params.max_open_zones, 0),
> > > > DEFINE_PROP_UINT32("zoned.descr_ext_size", NvmeNamespace,
> > > > params.zd_extension_size, 0),
> > > > + DEFINE_PROP_BOOL("zoned.auto_transition", NvmeNamespace,
> > > > + params.auto_transition_zones, true),
> > > > DEFINE_PROP_END_OF_LIST(),
> > > > };
> > > >
> > > > diff --git a/hw/nvme/nvme.h b/hw/nvme/nvme.h
> > > > index 81a35cda14..bd86054db2 100644
> > > > --- a/hw/nvme/nvme.h
> > > > +++ b/hw/nvme/nvme.h
> > > > @@ -100,6 +100,7 @@ typedef struct NvmeNamespaceParams {
> > > > uint32_t max_active_zones;
> > > > uint32_t max_open_zones;
> > > > uint32_t zd_extension_size;
> > > > + bool auto_transition_zones;
> > > > } NvmeNamespaceParams;
> > > >
> > > > typedef struct NvmeNamespace {
> > > > --
> > > > 2.31.1
> > > >
> > >
> > > Looks good Niklas!
> > >
> > > Reviewed-by: Klaus Jensen <k.jen...@samsung.com>
> >
> > In reality, it is the controller that does the auto transitioning.
> >
> > In theory, one namespace could be attached to two different controllers,
> > and I guess, in that case, it depends on if the controller that we used
> > when doing the write supports auto transitioning or not, that determines
> > if a zone will be auto transitioned or not.
> >
> > If we were to change this to be a parameter of the controller instead
> > of a parameter of the namespace, we would require to refactor a lot of
> > code in the regular write path. As we currently don't have any NvmeRequest
> > object in nvme_zrm_open_flags().
> >
> > Thoughts?
> >
>
> I think you are right. This should be controller-specific behavior. I took
> the liberty of moving the parameter; the refactor is minimal.
>
>
> From: Niklas Cassel <niklas.cas...@wdc.com>
>
> In the Zoned Namespace Command Set Specification, chapter
> 2.5.1 Managing resources
>
> "The controller may transition zones in the ZSIO:Implicitly Opened state
> to the ZSC:Closed state for resource management purposes."
>
> The word may in this sentence means that automatically transitioning
> an implicitly opened zone to closed is completely optional.
>
> Add a new parameter so that the user can control if this automatic
> transitioning should be performed or not.
>
> Being able to control this can help with verifying that e.g. a user-space
> program behaves properly even without this optional ZNS feature.
>
> The default value is set to true, in order to not change the existing
> behavior.
>
> Signed-off-by: Niklas Cassel <niklas.cas...@wdc.com>
> [k.jensen: moved parameter to controller]
> Signed-off-by: Klaus Jensen <k.jen...@samsung.com>
> ---
> hw/nvme/nvme.h | 1 +
> hw/nvme/ctrl.c | 32 ++++++++++++++++++++++----------
> 2 files changed, 23 insertions(+), 10 deletions(-)
>
> diff --git a/hw/nvme/nvme.h b/hw/nvme/nvme.h
> index 81a35cda142b..93a7e0e5380e 100644
> --- a/hw/nvme/nvme.h
> +++ b/hw/nvme/nvme.h
> @@ -382,6 +382,7 @@ typedef struct NvmeParams {
> uint8_t vsl;
> bool use_intel_id;
> uint8_t zasl;
> + bool auto_transition_zones;
> bool legacy_cmb;
> } NvmeParams;
> diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
> index 40a7efcea914..8dd9cb2ccbf3 100644
> --- a/hw/nvme/ctrl.c
> +++ b/hw/nvme/ctrl.c
> @@ -34,6 +34,7 @@
> * aerl=<N[optional]>,aer_max_queued=<N[optional]>, \
> * mdts=<N[optional]>,vsl=<N[optional]>, \
> * zoned.zasl=<N[optional]>, \
> + * zoned.auto_transition=<on|off[optional]>, \
> * subsys=<subsys_id>
> * -device nvme-ns,drive=<drive_id>,bus=<bus_name>,nsid=<nsid>,\
> * zoned=<true|false[optional]>, \
> @@ -100,6 +101,11 @@
> * the minimum memory page size (CAP.MPSMIN). The default value is 0 (i.e.
> * defaulting to the value of `mdts`).
> *
> + * - `zoned.auto_transition`
> + * Indicates if zones in zone state implicitly opened can be automatically
> + * transitioned to zone state closed for resource management purposes.
> + * Defaults to 'on'.
> + *
> * nvme namespace device parameters
> * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> * - `shared`
> @@ -1686,8 +1692,8 @@ enum {
> NVME_ZRM_AUTO = 1 << 0,
> };
> -static uint16_t nvme_zrm_open_flags(NvmeNamespace *ns, NvmeZone *zone,
> - int flags)
> +static uint16_t nvme_zrm_open_flags(NvmeCtrl *n, NvmeNamespace *ns,
> + NvmeZone *zone, int flags)
> {
> int act = 0;
> uint16_t status;
> @@ -1699,7 +1705,9 @@ static uint16_t nvme_zrm_open_flags(NvmeNamespace *ns,
> NvmeZone *zone,
> /* fallthrough */
> case NVME_ZONE_STATE_CLOSED:
> - nvme_zrm_auto_transition_zone(ns);
> + if (n->params.auto_transition_zones) {
> + nvme_zrm_auto_transition_zone(ns);
> + }
> status = nvme_aor_check(ns, act, 1);
> if (status) {
> return status;
> @@ -1735,14 +1743,16 @@ static uint16_t nvme_zrm_open_flags(NvmeNamespace
> *ns, NvmeZone *zone,
> }
> }
> -static inline uint16_t nvme_zrm_auto(NvmeNamespace *ns, NvmeZone *zone)
> +static inline uint16_t nvme_zrm_auto(NvmeCtrl *n, NvmeNamespace *ns,
> + NvmeZone *zone)
> {
> - return nvme_zrm_open_flags(ns, zone, NVME_ZRM_AUTO);
> + return nvme_zrm_open_flags(n, ns, zone, NVME_ZRM_AUTO);
> }
> -static inline uint16_t nvme_zrm_open(NvmeNamespace *ns, NvmeZone *zone)
> +static inline uint16_t nvme_zrm_open(NvmeCtrl *n, NvmeNamespace *ns,
> + NvmeZone *zone)
> {
> - return nvme_zrm_open_flags(ns, zone, 0);
> + return nvme_zrm_open_flags(n, ns, zone, 0);
> }
> static void nvme_advance_zone_wp(NvmeNamespace *ns, NvmeZone *zone,
> @@ -2283,7 +2293,7 @@ static void nvme_copy_in_complete(NvmeRequest *req)
> goto invalid;
> }
> - status = nvme_zrm_auto(ns, zone);
> + status = nvme_zrm_auto(nvme_ctrl(req), ns, zone);
> if (status) {
> goto invalid;
> }
> @@ -3080,7 +3090,7 @@ static uint16_t nvme_do_write(NvmeCtrl *n, NvmeRequest
> *req, bool append,
> goto invalid;
> }
> - status = nvme_zrm_auto(ns, zone);
> + status = nvme_zrm_auto(n, ns, zone);
> if (status) {
> goto invalid;
> }
> @@ -3169,7 +3179,7 @@ enum NvmeZoneProcessingMask {
> static uint16_t nvme_open_zone(NvmeNamespace *ns, NvmeZone *zone,
> NvmeZoneState state, NvmeRequest *req)
> {
> - return nvme_zrm_open(ns, zone);
> + return nvme_zrm_open(nvme_ctrl(req), ns, zone);
> }
> static uint16_t nvme_close_zone(NvmeNamespace *ns, NvmeZone *zone,
> @@ -6259,6 +6269,8 @@ static Property nvme_props[] = {
> DEFINE_PROP_BOOL("use-intel-id", NvmeCtrl, params.use_intel_id, false),
> DEFINE_PROP_BOOL("legacy-cmb", NvmeCtrl, params.legacy_cmb, false),
> DEFINE_PROP_UINT8("zoned.zasl", NvmeCtrl, params.zasl, 0),
> + DEFINE_PROP_BOOL("zoned.auto_transition", NvmeCtrl,
> + params.auto_transition_zones, true),
> DEFINE_PROP_END_OF_LIST(),
> };
> --
> 2.31.1
>
Thanks a lot Klaus! I really appreciate it.
My initial thought was to add a new flag in the enum where NVME_ZRM_AUTO is.
But I think that would just make the code harder to read.
You simply check the parameter directly, which is more obvious to the reader,
so I think patch looks good!
Kind regards,
Niklas
