| Message ID | 20200529220600.225320-7-vaibhav@linux.ibm.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Add support for reporting papr nvdimm health | expand |
> -----Original Message----- > From: Vaibhav Jain <vaibhav@linux.ibm.com> > Sent: Friday, May 29, 2020 3:06 PM > To: linux-nvdimm@lists.01.org > Cc: Vaibhav Jain <vaibhav@linux.ibm.com>; Williams, Dan J > <dan.j.williams@intel.com>; Verma, Vishal L <vishal.l.verma@intel.com>; > Aneesh Kumar K . V <aneesh.kumar@linux.ibm.com>; Jeff Moyer > <jmoyer@redhat.com>; Oliver O'Halloran <oohall@gmail.com>; Santosh > Sivaraj <santosh@fossix.org>; Weiny, Ira <ira.weiny@intel.com> > Subject: [ndctl PATCH v5 6/6] libndctl,papr_scm: Implement support for > PAPR_PDSM_HEALTH > > Add support for reporting DIMM health and shutdown state by issuing > PAPR_PDSM_HEALTH request to papr_scm module. It returns an instance of > 'struct nd_papr_pdsm_health' as defined in 'papr_pdsm.h'. The patch > provides support for dimm-ops 'new_smart', 'smart_get_health' & > 'smart_get_shutdown_state' as newly introduced functions > papr_new_smart_health(), papr_smart_get_health() & > papr_smart_get_shutdown_state() respectively. These callbacks should > enable ndctl to report DIMM health. > > Also a new member 'struct dimm_priv.health' is introduced which holds the > current health status of the dimm. This member is set inside newly added > function 'update_dimm_health_v1()' which parses the v1 payload returned > by the kernel after servicing PAPR_PDSM_HEALTH. The function will also > update dimm-flags viz 'struct ndctl_dimm.flags.f_*' > based on the flags set in the returned payload. > > Signed-off-by: Vaibhav Jain <vaibhav@linux.ibm.com> > --- > Changelog: > > v4..v5: > * Updated patch description to reflect updated names of struct and > defines that have the term 'scm' removed. > > v3..v4: > * None > > v2..v3: > * None > > v1..v2: > * Squashed patch to report nvdimm bad shutdown state with this patch. > * Switched to new structs/enums as defined in papr_scm_pdsm.h > --- > ndctl/lib/papr.c | 90 > ++++++++++++++++++++++++++++++++++++++++++++++-- > 1 file changed, 87 insertions(+), 3 deletions(-) > > diff --git a/ndctl/lib/papr.c b/ndctl/lib/papr.c index > 1b7870beb631..cb7ff9e0d5bd 100644 > --- a/ndctl/lib/papr.c > +++ b/ndctl/lib/papr.c > @@ -42,7 +42,9 @@ > > /* Per dimm data. Holds per-dimm data parsed from the cmd_pkgs */ struct > dimm_priv { > - /* Empty for now */ > + > + /* Cache the dimm health status */ > + struct nd_papr_pdsm_health health; I don't understand this. The kernel is caching this, why does libndctl need to cache it?
"Williams, Dan J" <dan.j.williams@intel.com> writes: >> -----Original Message----- >> From: Vaibhav Jain <vaibhav@linux.ibm.com> >> Sent: Friday, May 29, 2020 3:06 PM >> To: linux-nvdimm@lists.01.org >> Cc: Vaibhav Jain <vaibhav@linux.ibm.com>; Williams, Dan J >> <dan.j.williams@intel.com>; Verma, Vishal L <vishal.l.verma@intel.com>; >> Aneesh Kumar K . V <aneesh.kumar@linux.ibm.com>; Jeff Moyer >> <jmoyer@redhat.com>; Oliver O'Halloran <oohall@gmail.com>; Santosh >> Sivaraj <santosh@fossix.org>; Weiny, Ira <ira.weiny@intel.com> >> Subject: [ndctl PATCH v5 6/6] libndctl,papr_scm: Implement support for >> PAPR_PDSM_HEALTH >> >> Add support for reporting DIMM health and shutdown state by issuing >> PAPR_PDSM_HEALTH request to papr_scm module. It returns an instance of >> 'struct nd_papr_pdsm_health' as defined in 'papr_pdsm.h'. The patch >> provides support for dimm-ops 'new_smart', 'smart_get_health' & >> 'smart_get_shutdown_state' as newly introduced functions >> papr_new_smart_health(), papr_smart_get_health() & >> papr_smart_get_shutdown_state() respectively. These callbacks should >> enable ndctl to report DIMM health. >> >> Also a new member 'struct dimm_priv.health' is introduced which holds the >> current health status of the dimm. This member is set inside newly added >> function 'update_dimm_health_v1()' which parses the v1 payload returned >> by the kernel after servicing PAPR_PDSM_HEALTH. The function will also >> update dimm-flags viz 'struct ndctl_dimm.flags.f_*' >> based on the flags set in the returned payload. >> >> Signed-off-by: Vaibhav Jain <vaibhav@linux.ibm.com> >> --- >> Changelog: >> >> v4..v5: >> * Updated patch description to reflect updated names of struct and >> defines that have the term 'scm' removed. >> >> v3..v4: >> * None >> >> v2..v3: >> * None >> >> v1..v2: >> * Squashed patch to report nvdimm bad shutdown state with this patch. >> * Switched to new structs/enums as defined in papr_scm_pdsm.h >> --- >> ndctl/lib/papr.c | 90 >> ++++++++++++++++++++++++++++++++++++++++++++++-- >> 1 file changed, 87 insertions(+), 3 deletions(-) >> >> diff --git a/ndctl/lib/papr.c b/ndctl/lib/papr.c index >> 1b7870beb631..cb7ff9e0d5bd 100644 >> --- a/ndctl/lib/papr.c >> +++ b/ndctl/lib/papr.c >> @@ -42,7 +42,9 @@ >> >> /* Per dimm data. Holds per-dimm data parsed from the cmd_pkgs */ struct >> dimm_priv { >> - /* Empty for now */ >> + >> + /* Cache the dimm health status */ >> + struct nd_papr_pdsm_health health; > [.] > I don't understand this. The kernel is caching this, why does libndctl > need to cache it? Was caching it here as the returned nvdimm health payload from kernel might be of different version than what was requested and reconstituting 'struct nd_papr_pdsm_health' from 'struct ndctl_cmd' in each dimm-ops callback would have be costly. However with payload versioning scheme not used any more this caching wont be needed anymore.
diff --git a/ndctl/lib/papr.c b/ndctl/lib/papr.c index 1b7870beb631..cb7ff9e0d5bd 100644 --- a/ndctl/lib/papr.c +++ b/ndctl/lib/papr.c @@ -42,7 +42,9 @@ /* Per dimm data. Holds per-dimm data parsed from the cmd_pkgs */ struct dimm_priv { - /* Empty for now */ + + /* Cache the dimm health status */ + struct nd_papr_pdsm_health health; }; static bool papr_cmd_is_supported(struct ndctl_dimm *dimm, int cmd) @@ -97,6 +99,43 @@ static bool cmd_is_valid(struct ndctl_dimm *dimm, struct ndctl_cmd *cmd) return true; } +/* + * Parse the nd_papr_pdsm_health_v1 payload embedded in ndctl_cmd and + * update dimm health/flags + */ +static int update_dimm_health_v1(struct ndctl_dimm *dimm, struct ndctl_cmd *cmd) +{ + struct nd_pdsm_cmd_pkg *pcmd = nd_to_pdsm_cmd_pkg(cmd->pkg); + struct dimm_priv *p = dimm->dimm_user_data; + const struct nd_papr_pdsm_health_v1 *health = + pdsm_cmd_to_payload(pcmd); + + /* Update the dimm flags */ + dimm->flags.f_arm = health->dimm_unarmed; + dimm->flags.f_flush = health->dimm_bad_shutdown; + dimm->flags.f_restore = health->dimm_bad_restore; + dimm->flags.f_smart = (health->dimm_health != 0); + + /* Cache the dimm health information */ + memcpy(&p->health, health, sizeof(*health)); + return 0; +} + +/* Check payload version returned and pass the packet to appropriate handler */ +static int update_dimm_health(struct ndctl_dimm *dimm, struct ndctl_cmd *cmd) +{ + const struct nd_pdsm_cmd_pkg *pcmd = nd_to_pdsm_cmd_pkg(cmd->pkg); + + if (pcmd->payload_version == 1) + return update_dimm_health_v1(dimm, cmd); + + /* unknown version */ + papr_err(dimm, "Unknown payload version for dimm_health.\n"); + papr_dbg(dimm, "dimm_health payload Ver=%d, Supported=%d\n", + pcmd->payload_version, ND_PAPR_PDSM_HEALTH_VERSION); + return -EINVAL; +} + /* Parse a command payload and update dimm flags/private data */ static int update_dimm_stats(struct ndctl_dimm *dimm, struct ndctl_cmd *cmd) { @@ -122,6 +161,8 @@ static int update_dimm_stats(struct ndctl_dimm *dimm, struct ndctl_cmd *cmd) /* Get the pdsm request and handle it */ pcmd = nd_to_pdsm_cmd_pkg(cmd->pkg); switch (pcmd_to_pdsm(pcmd)) { + case PAPR_PDSM_HEALTH: + return update_dimm_health(dimm, cmd); default: papr_err(dimm, "Unhandled pdsm-request 0x%016llx\n", pcmd_to_pdsm(pcmd)); @@ -166,14 +207,54 @@ static struct ndctl_cmd *allocate_cmd(struct ndctl_dimm *dimm, return cmd; } +static struct ndctl_cmd *papr_new_smart_health(struct ndctl_dimm *dimm) +{ + struct ndctl_cmd *cmd_ret; + + cmd_ret = allocate_cmd(dimm, PAPR_PDSM_HEALTH, + sizeof(struct nd_papr_pdsm_health), + ND_PAPR_PDSM_HEALTH_VERSION); + if (!cmd_ret) { + papr_err(dimm, "Unable to allocate smart_health command\n"); + return NULL; + } + + cmd_ret->pkg[0].nd_size_out = ND_PDSM_ENVELOPE_CONTENT_SIZE( + struct nd_papr_pdsm_health); + + return cmd_ret; +} + +static unsigned int papr_smart_get_health(struct ndctl_cmd *cmd) +{ + struct dimm_priv *p = cmd->dimm->dimm_user_data; + + /* + * Update the dimm stats and use some math to return one of + * defined ND_SMART_*_HEALTH values + */ + if (update_dimm_stats(cmd->dimm, cmd) || !p->health.dimm_health) + return 0; + else + return 1 << (p->health.dimm_health - 1); +} + +static unsigned int papr_smart_get_shutdown_state(struct ndctl_cmd *cmd) +{ + struct dimm_priv *p = cmd->dimm->dimm_user_data; + + /* Update dimm state and return f_flush */ + return update_dimm_stats(cmd->dimm, cmd) ? + 0 : p->health.dimm_bad_shutdown; +} + static unsigned int papr_smart_get_flags(struct ndctl_cmd *cmd) { /* In case of error return empty flags * */ if (update_dimm_stats(cmd->dimm, cmd)) return 0; - /* Return empty flags for now as no DSM support */ - return 0; + return ND_SMART_HEALTH_VALID | ND_SMART_SHUTDOWN_VALID; } static int papr_dimm_init(struct ndctl_dimm *dimm) @@ -214,4 +295,7 @@ struct ndctl_dimm_ops * const papr_dimm_ops = &(struct ndctl_dimm_ops) { .dimm_uninit = papr_dimm_uninit, .smart_get_flags = papr_smart_get_flags, .get_firmware_status = papr_get_firmware_status, + .new_smart = papr_new_smart_health, + .smart_get_health = papr_smart_get_health, + .smart_get_shutdown_state = papr_smart_get_shutdown_state, };
Add support for reporting DIMM health and shutdown state by issuing PAPR_PDSM_HEALTH request to papr_scm module. It returns an instance of 'struct nd_papr_pdsm_health' as defined in 'papr_pdsm.h'. The patch provides support for dimm-ops 'new_smart', 'smart_get_health' & 'smart_get_shutdown_state' as newly introduced functions papr_new_smart_health(), papr_smart_get_health() & papr_smart_get_shutdown_state() respectively. These callbacks should enable ndctl to report DIMM health. Also a new member 'struct dimm_priv.health' is introduced which holds the current health status of the dimm. This member is set inside newly added function 'update_dimm_health_v1()' which parses the v1 payload returned by the kernel after servicing PAPR_PDSM_HEALTH. The function will also update dimm-flags viz 'struct ndctl_dimm.flags.f_*' based on the flags set in the returned payload. Signed-off-by: Vaibhav Jain <vaibhav@linux.ibm.com> --- Changelog: v4..v5: * Updated patch description to reflect updated names of struct and defines that have the term 'scm' removed. v3..v4: * None v2..v3: * None v1..v2: * Squashed patch to report nvdimm bad shutdown state with this patch. * Switched to new structs/enums as defined in papr_scm_pdsm.h --- ndctl/lib/papr.c | 90 ++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 87 insertions(+), 3 deletions(-)