Re: [PATCH V2 05/20] nvdimm/region_label: Add region label updation routine

[Neeraj Kumar <s.neeraj@samsung.com>]

[-- Attachment #1: Type: text/plain, Size: 10530 bytes --]

On 15/08/25 02:55PM, Dave Jiang wrote:
>
>
>On 7/30/25 5:11 AM, Neeraj Kumar wrote:
>> Added __pmem_region_label_update region label update routine to update
>> region label.
>>
>> Also used guard(mutex)(&nd_mapping->lock) in place of mutex_lock() and
>> mutex_unlock()
>>
>> Signed-off-by: Neeraj Kumar <s.neeraj@samsung.com>
>
>Subject, s/updation/update/ ?

Thanks Dave, Sure. Will fix it in next patch-set

>
>> ---
>>  drivers/nvdimm/label.c          | 171 +++++++++++++++++++++++++++++---
>>  drivers/nvdimm/label.h          |   2 +
>>  drivers/nvdimm/namespace_devs.c |  12 +++
>>  drivers/nvdimm/nd.h             |  20 ++++
>>  include/linux/libnvdimm.h       |   8 ++
>>  5 files changed, 198 insertions(+), 15 deletions(-)
>>
>> diff --git a/drivers/nvdimm/label.c b/drivers/nvdimm/label.c
>> index 3f8a6bdb77c7..94f2d0ba7aca 100644
>> --- a/drivers/nvdimm/label.c
>> +++ b/drivers/nvdimm/label.c
>> @@ -381,6 +381,16 @@ static void nsl_calculate_checksum(struct nvdimm_drvdata *ndd,
>>  	nsl_set_checksum(ndd, nd_label, sum);
>>  }
>>
>> +static void rgl_calculate_checksum(struct nvdimm_drvdata *ndd,
>
>region_label_checksum()? rgl/rg is just a bit jarring to read even though I get nvdimm already has nd and nsl etc.
>
>> +				   struct cxl_region_label *rg_label)
>
>Prefer just spell out region_label

Sure Dave, I will fix it wherever its rgl/rg to region_label

>
>> +{
>> +	u64 sum;
>> +
>> +	rgl_set_checksum(rg_label, 0);
>> +	sum = nd_fletcher64(rg_label, sizeof_namespace_label(ndd), 1);
>> +	rgl_set_checksum(rg_label, sum);
>> +}
>> +
>>  static bool slot_valid(struct nvdimm_drvdata *ndd,
>>  		struct nd_lsa_label *lsa_label, u32 slot)
>>  {
>> @@ -960,7 +970,7 @@ static int __pmem_label_update(struct nd_region *nd_region,
>>  		return rc;
>>
>>  	/* Garbage collect the previous label */
>> -	mutex_lock(&nd_mapping->lock);
>> +	guard(mutex)(&nd_mapping->lock);
>>  	list_for_each_entry(label_ent, &nd_mapping->labels, list) {
>>  		if (!label_ent->label)
>>  			continue;
>> @@ -972,20 +982,20 @@ static int __pmem_label_update(struct nd_region *nd_region,
>>  	/* update index */
>>  	rc = nd_label_write_index(ndd, ndd->ns_next,
>>  			nd_inc_seq(__le32_to_cpu(nsindex->seq)), 0);
>> -	if (rc == 0) {
>> -		list_for_each_entry(label_ent, &nd_mapping->labels, list)
>> -			if (!label_ent->label) {
>> -				label_ent->label = lsa_label;
>> -				lsa_label = NULL;
>> -				break;
>> -			}
>> -		dev_WARN_ONCE(&nspm->nsio.common.dev, lsa_label,
>> -				"failed to track label: %d\n",
>> -				to_slot(ndd, lsa_label));
>> -		if (lsa_label)
>> -			rc = -ENXIO;
>> -	}
>> -	mutex_unlock(&nd_mapping->lock);
>> +	if (rc)
>> +		return rc;
>> +
>> +	list_for_each_entry(label_ent, &nd_mapping->labels, list)
>> +		if (!label_ent->label) {
>> +			label_ent->label = lsa_label;
>> +			lsa_label = NULL;
>> +			break;
>> +		}
>
>Would've preferred the original code to look like:
>
>list_for_each_entry(label_ent, &nd_mapping->labels, list) {
>	if (label_ent->label)
>		continue;
>
>	label_ent->label = lsa_label;
>	lsa_label = NULL;
>	break;
>}
>
>But ah well....

Thanks, I will fix it here

>
>> +	dev_WARN_ONCE(&nspm->nsio.common.dev, lsa_label,> +			"failed to track label: %d\n",
>> +			to_slot(ndd, lsa_label));
>> +	if (lsa_label)
>> +		rc = -ENXIO;
>
>Just return here as Jonathan already mentioned. guard() helps with cleaning that up.

Sure, Will fix it in next patch-set

>
>>
>>  	return rc;
>>  }
>> @@ -1127,6 +1137,137 @@ int nd_pmem_namespace_label_update(struct nd_region *nd_region,
>>  	return 0;
>>  }
>>
>> +static int __pmem_region_label_update(struct nd_region *nd_region,
>> +		struct nd_mapping *nd_mapping, int pos, unsigned long flags)
>> +{
>> +	struct nd_interleave_set *nd_set = nd_region->nd_set;
>> +	struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
>> +	struct nd_lsa_label *nd_label;
>> +	struct cxl_region_label *rg_label;
>> +	struct nd_namespace_index *nsindex;> +	struct nd_label_ent *label_ent;
>> +	unsigned long *free;
>> +	u32 nslot, slot;
>> +	size_t offset;
>> +	int rc;
>> +	uuid_t tmp;
>
>uuid instead of tmp would make the variable clearer to read
>
>Also please arrange variable ordering in reverse xmas tree.

Sure Dave, I will fix it in next patch-set

>
>> +
>> +	if (!preamble_next(ndd, &nsindex, &free, &nslot))
>> +		return -ENXIO;
>> +
>> +	/* allocate and write the label to the staging (next) index */
>> +	slot = nd_label_alloc_slot(ndd);
>> +	if (slot == UINT_MAX)
>> +		return -ENXIO;
>> +	dev_dbg(ndd->dev, "allocated: %d\n", slot);
>> +
>> +	nd_label = to_label(ndd, slot);
>> +
>> +	memset(nd_label, 0, sizeof_namespace_label(ndd));
>> +	rg_label = &nd_label->rg_label;
>> +
>> +	/* Set Region Label Format identification UUID */
>> +	uuid_parse(CXL_REGION_UUID, &tmp);
>> +	export_uuid(nd_label->rg_label.type, &tmp);
>> +
>> +	/* Set Current Region Label UUID */
>> +	export_uuid(nd_label->rg_label.uuid, &nd_set->uuid);
>> +
>> +	rg_label->flags = __cpu_to_le32(flags);
>> +	rg_label->nlabel = __cpu_to_le16(nd_region->ndr_mappings);
>> +	rg_label->position = __cpu_to_le16(pos);
>> +	rg_label->dpa = __cpu_to_le64(nd_mapping->start);
>> +	rg_label->rawsize = __cpu_to_le64(nd_mapping->size);
>> +	rg_label->hpa = __cpu_to_le64(nd_set->res->start);
>> +	rg_label->slot = __cpu_to_le32(slot);
>> +	rg_label->ig = __cpu_to_le32(nd_set->interleave_granularity);
>> +	rg_label->align = __cpu_to_le16(0);
>> +
>> +	/* Update fletcher64 Checksum */
>> +	rgl_calculate_checksum(ndd, rg_label);
>> +
>> +	/* update label */
>> +	offset = nd_label_offset(ndd, nd_label);
>> +	rc = nvdimm_set_config_data(ndd, offset, nd_label,
>> +			sizeof_namespace_label(ndd));
>> +	if (rc < 0) {
>> +		nd_label_free_slot(ndd, slot);
>> +		return rc;
>> +	}
>> +
>> +	/* Garbage collect the previous label */
>> +	guard(mutex)(&nd_mapping->lock);
>> +	list_for_each_entry(label_ent, &nd_mapping->labels, list) {
>> +		if (!label_ent->label)
>> +			continue;
>> +		if (rgl_uuid_equal(&label_ent->label->rg_label, &nd_set->uuid))
>> +			reap_victim(nd_mapping, label_ent);
>> +	}
>> +
>> +	/* update index */
>> +	rc = nd_label_write_index(ndd, ndd->ns_next,
>> +			nd_inc_seq(__le32_to_cpu(nsindex->seq)), 0);
>> +	if (rc)
>> +		return rc;
>> +
>> +	list_for_each_entry(label_ent, &nd_mapping->labels, list)
>> +		if (!label_ent->label) {
>> +			label_ent->label = nd_label;
>> +			nd_label = NULL;
>> +			break;
>> +		}
>> +	dev_WARN_ONCE(&nd_region->dev, nd_label,
>> +			"failed to track label: %d\n",
>> +			to_slot(ndd, nd_label));
>> +	if (nd_label)
>> +		rc = -ENXIO;
>just return here

Sure, Will fix it

>
>> +
>> +	return rc;
>> +}
>> +
>> +int nd_pmem_region_label_update(struct nd_region *nd_region)
>> +{
>> +	int i, rc;
>> +
>> +	for (i = 0; i < nd_region->ndr_mappings; i++) {
>> +		struct nd_mapping *nd_mapping = &nd_region->mapping[i];
>> +		struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
>> +
>> +		/* No need to update region label for non cxl format */
>> +		if (!ndd->cxl)
>> +			continue;
>
>Would there be a mix of different nd mappings? I wonder if you can just 'return 0' if you find ndd->cxl on the first one and just skip everything.

When we create cxl region with two mem device, then we will have two separate
nd_mapping for both mem devices. But Yes, I don't see difference in both device
nd_mapping characters. So instead of "continue", I will just "return 0".

>
>> +
>> +		/* Init labels to include region label */
>> +		rc = init_labels(nd_mapping, 1);
>> +
>> +		if (rc < 0)
>> +			return rc;
>> +
>> +		rc = __pmem_region_label_update(nd_region, nd_mapping, i,
>> +					NSLABEL_FLAG_UPDATING);
>> +
>> +		if (rc)
>> +			return rc;
>> +	}
>> +
>> +	/* Clear the UPDATING flag per UEFI 2.7 expectations */
>> +	for (i = 0; i < nd_region->ndr_mappings; i++) {
>> +		struct nd_mapping *nd_mapping = &nd_region->mapping[i];
>> +		struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
>> +
>> +		/* No need to update region label for non cxl format */
>> +		if (!ndd->cxl)
>> +			continue;
>> +> +		rc = __pmem_region_label_update(nd_region, nd_mapping, i, 0);
>> +
>> +		if (rc)
>> +			return rc;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>>  int __init nd_label_init(void)
>>  {
>>  	WARN_ON(guid_parse(NVDIMM_BTT_GUID, &nvdimm_btt_guid));
>> diff --git a/drivers/nvdimm/label.h b/drivers/nvdimm/label.h
>> index 4883b3a1320f..0f428695017d 100644
>> --- a/drivers/nvdimm/label.h
>> +++ b/drivers/nvdimm/label.h
>> @@ -190,6 +190,7 @@ struct nd_namespace_label {
>>  struct nd_lsa_label {
>>  	union {
>>  		struct nd_namespace_label ns_label;
>> +		struct cxl_region_label rg_label;
>>  	};
>>  };
>>
>> @@ -233,4 +234,5 @@ struct nd_region;
>>  struct nd_namespace_pmem;
>>  int nd_pmem_namespace_label_update(struct nd_region *nd_region,
>>  		struct nd_namespace_pmem *nspm, resource_size_t size);
>> +int nd_pmem_region_label_update(struct nd_region *nd_region);
>>  #endif /* __LABEL_H__ */
>> diff --git a/drivers/nvdimm/namespace_devs.c b/drivers/nvdimm/namespace_devs.c
>> index 5b73119dc8fd..02ae8162566c 100644
>> --- a/drivers/nvdimm/namespace_devs.c
>> +++ b/drivers/nvdimm/namespace_devs.c
>> @@ -232,6 +232,18 @@ static ssize_t __alt_name_store(struct device *dev, const char *buf,
>>  	return rc;
>>  }
>>
>> +int nd_region_label_update(struct nd_region *nd_region)
>> +{
>> +	int rc;
>> +
>> +	nvdimm_bus_lock(&nd_region->dev);
>> +	rc = nd_pmem_region_label_update(nd_region);
>> +	nvdimm_bus_unlock(&nd_region->dev);
>> +
>> +	return rc;
>> +}
>> +EXPORT_SYMBOL_GPL(nd_region_label_update);
>> +
>>  static int nd_namespace_label_update(struct nd_region *nd_region,
>>  		struct device *dev)
>>  {
>> diff --git a/drivers/nvdimm/nd.h b/drivers/nvdimm/nd.h
>> index 651847f1bbf9..15d94e3937f0 100644
>> --- a/drivers/nvdimm/nd.h
>> +++ b/drivers/nvdimm/nd.h
>> @@ -322,6 +322,26 @@ static inline void nsl_set_region_uuid(struct nvdimm_drvdata *ndd,
>>  		export_uuid(ns_label->cxl.region_uuid, uuid);
>>  }
>>
>> +static inline bool rgl_uuid_equal(struct cxl_region_label *rg_label,
>> +				  const uuid_t *uuid)
>
>region_label_uuid_equal() and region_label
>
>> +{
>> +	uuid_t tmp;
>
>tmp_uuid
>
>> +
>> +	import_uuid(&tmp, rg_label->uuid);
>> +	return uuid_equal(&tmp, uuid);
>> +}
>> +
>> +static inline u64 rgl_get_checksum(struct cxl_region_label *rg_label)
>
>region_label_get_checksum()
>
>> +{
>> +	return __le64_to_cpu(rg_label->checksum);
>> +}
>> +
>> +static inline void rgl_set_checksum(struct cxl_region_label *rg_label,
>region_label_set_checksum()

Sure, I will fix them in next patch-set


Regards,
Neeraj

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]