| Message ID | 20171220221325.GA56741@jaegeuk-macbookpro.roam.corp.google.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
On Wed, Dec 20, 2017 at 02:13:25PM -0800, Jaegeuk Kim wrote: > This patch adds a new sysfs group, namely health, via: > > /sys/devices/soc/X.ufshc/health/ > > This directory contains the below entries, each of which shows an 8-bytes > hex number representing different meanings defined by JEDEC specfication. > > Users can simply read these entries to check how their underlying flash > storage is getting reached out to its end of life. For example, if > lifetimeA shows 0xb, it would be the right time to consider device swap. > > - length > : must be 25h > > - type > : must be 09h > > - eol > 00h: Not defined > 01h: Normal > 02h: Warning > 03h: Critical > > - lifetimeA/B > 00h: Not defined > 01h: 0% ~ 10% device life time used > 02h: 10% ~ 20% device life time used > 03h: 20% ~ 30% device life time used > 04h: 30% ~ 40% device life time used > 05h: 40% ~ 50% device life time used > 06h: 50% ~ 60% device life time used > 07h: 60% ~ 70% device life time used > 08h: 70% ~ 80% device life time used > 09h: 80% ~ 90% device life time used > 0Ah: 90% ~ 100% device life time used > 0Bh: Exceeded its maximum estimated device life time > > Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> > Signed-off-by: Jaegeuk Kim <jaegeuk@google.com> > --- Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> -----Original Message----- > From: linux-scsi-owner@vger.kernel.org [mailto:linux-scsi- > owner@vger.kernel.org] On Behalf Of Greg Kroah-Hartman > Sent: Thursday, December 21, 2017 10:00 AM > To: Jaegeuk Kim <jaegeuk@kernel.org> > Cc: linux-kernel@vger.kernel.org; linux-scsi@vger.kernel.org; Jaegeuk Kim > <jaegeuk@google.com> > Subject: Re: [PATCH 2/2 v4] scsi: ufs: introduce sysfs entries exposing UFS > health info > > On Wed, Dec 20, 2017 at 02:13:25PM -0800, Jaegeuk Kim wrote: > > This patch adds a new sysfs group, namely health, via: > > > > /sys/devices/soc/X.ufshc/health/ As device health is just one piece of information out of the device management, I think that you should address this in a more comprehensive way, And set hooks for much more device info: Allow access to device descriptors, attributes and flags. The attributes and flags should be placed in separate subfolders The LUN specific descriptors and attributes should be placed in a luns subfolder, and then per descriptor / attribute type You might also would like to consider differentiating read and write - to control those type of accesses as well. Cheers, Avri
On Wed, Dec 27, 2017 at 09:00:10AM +0000, Avri Altman wrote: > > > > -----Original Message----- > > From: linux-scsi-owner@vger.kernel.org [mailto:linux-scsi- > > owner@vger.kernel.org] On Behalf Of Greg Kroah-Hartman > > Sent: Thursday, December 21, 2017 10:00 AM > > To: Jaegeuk Kim <jaegeuk@kernel.org> > > Cc: linux-kernel@vger.kernel.org; linux-scsi@vger.kernel.org; Jaegeuk Kim > > <jaegeuk@google.com> > > Subject: Re: [PATCH 2/2 v4] scsi: ufs: introduce sysfs entries exposing UFS > > health info > > > > On Wed, Dec 20, 2017 at 02:13:25PM -0800, Jaegeuk Kim wrote: > > > This patch adds a new sysfs group, namely health, via: > > > > > > /sys/devices/soc/X.ufshc/health/ > As device health is just one piece of information out of the device management, > I think that you should address this in a more comprehensive way, > And set hooks for much more device info: > Allow access to device descriptors, attributes and flags. Add on patches are easy to create for this if people really want and need it :) > The attributes and flags should be placed in separate subfolders Why? What is that going to help with? > The LUN specific descriptors and attributes should be placed in a luns > subfolder, and then per descriptor / attribute type Again, why? > You might also would like to consider differentiating read and write - > to control those type of accesses as well. What do you mean by this exactly? As it is, this is a step forward in getting attributes that people are asking for and already using, into the kernel tree. Please don't object because not all attributes that are possible are being added here, it should be trivial to add more as needed, right? I'm really tired of seeing all of the various out-of-tree forks of this driver, it's about time that someone works to get those features merged, right? thanks, greg k-h
On 12/27, Greg Kroah-Hartman wrote: > On Wed, Dec 27, 2017 at 09:00:10AM +0000, Avri Altman wrote: > > > > > > > -----Original Message----- > > > From: linux-scsi-owner@vger.kernel.org [mailto:linux-scsi- > > > owner@vger.kernel.org] On Behalf Of Greg Kroah-Hartman > > > Sent: Thursday, December 21, 2017 10:00 AM > > > To: Jaegeuk Kim <jaegeuk@kernel.org> > > > Cc: linux-kernel@vger.kernel.org; linux-scsi@vger.kernel.org; Jaegeuk Kim > > > <jaegeuk@google.com> > > > Subject: Re: [PATCH 2/2 v4] scsi: ufs: introduce sysfs entries exposing UFS > > > health info > > > > > > On Wed, Dec 20, 2017 at 02:13:25PM -0800, Jaegeuk Kim wrote: > > > > This patch adds a new sysfs group, namely health, via: > > > > > > > > /sys/devices/soc/X.ufshc/health/ > > As device health is just one piece of information out of the device management, > > I think that you should address this in a more comprehensive way, > > And set hooks for much more device info: > > Allow access to device descriptors, attributes and flags. > > Add on patches are easy to create for this if people really want and > need it :) > > > The attributes and flags should be placed in separate subfolders > > Why? What is that going to help with? > > > The LUN specific descriptors and attributes should be placed in a luns > > subfolder, and then per descriptor / attribute type > > Again, why? > > > You might also would like to consider differentiating read and write - > > to control those type of accesses as well. > > What do you mean by this exactly? > > As it is, this is a step forward in getting attributes that people are > asking for and already using, into the kernel tree. Please don't object > because not all attributes that are possible are being added here, it > should be trivial to add more as needed, right? > > I'm really tired of seeing all of the various out-of-tree forks of this > driver, it's about time that someone works to get those features merged, > right? Indeed, it seems someone has already done similar work before. Let me drop my patches and dive into that, although I have to clean up all the mess once again. :( Thanks, > > thanks, > > greg k-h
diff --git a/Documentation/ABI/testing/sysfs-devices-soc-ufs b/Documentation/ABI/testing/sysfs-devices-soc-ufs new file mode 100644 index 000000000000..313771a383e4 --- /dev/null +++ b/Documentation/ABI/testing/sysfs-devices-soc-ufs @@ -0,0 +1,25 @@ +What: /sys/devices/soc/X.ufshc/health +Date: September 2017 +contact: Jaegeuk Kim <jaegeuk@google.com> +Description: + This directory contains health information reported by UFS. + - length must be 25h + - type must be 09h + - eol represent + 00h: Not defined + 01h: Normal + 02h: Warning + 03h: Critical + - lifetimeA/B + 00h: Not defined + 01h: 0% ~ 10% device life time used + 02h: 10% ~ 20% device life time used + 03h: 20% ~ 30% device life time used + 04h: 30% ~ 40% device life time used + 05h: 40% ~ 50% device life time used + 06h: 50% ~ 60% device life time used + 07h: 60% ~ 70% device life time used + 08h: 70% ~ 80% device life time used + 09h: 80% ~ 90% device life time used + 0Ah: 90% ~ 100% device life time used + 0Bh: Exceeded its maximum estimated device life time diff --git a/MAINTAINERS b/MAINTAINERS index aa71ab52fd76..947034319bb4 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -13999,6 +13999,7 @@ M: Vinayak Holikatti <vinholikatti@gmail.com> L: linux-scsi@vger.kernel.org S: Supported F: Documentation/scsi/ufs.txt +F: Documentation/ABI/testing/sysfs-devices-soc-ufs F: drivers/scsi/ufs/ UNIVERSAL FLASH STORAGE HOST CONTROLLER DRIVER DWC HOOKS diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h index 54deeb754db5..1af541d56c7d 100644 --- a/drivers/scsi/ufs/ufs.h +++ b/drivers/scsi/ufs/ufs.h @@ -154,6 +154,7 @@ enum desc_idn { QUERY_DESC_IDN_RFU_1 = 0x6, QUERY_DESC_IDN_GEOMETRY = 0x7, QUERY_DESC_IDN_POWER = 0x8, + QUERY_DESC_IDN_HEALTH = 0x9, QUERY_DESC_IDN_MAX, }; @@ -169,6 +170,7 @@ enum ufs_desc_def_size { QUERY_DESC_INTERCONNECT_DEF_SIZE = 0x06, QUERY_DESC_GEOMETRY_DEF_SIZE = 0x44, QUERY_DESC_POWER_DEF_SIZE = 0x62, + QUERY_DESC_HEALTH_DEF_SIZE = 0x25, }; /* Unit descriptor parameters offsets in bytes*/ diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c index fc93cd808a93..17b898ca820b 100644 --- a/drivers/scsi/ufs/ufshcd.c +++ b/drivers/scsi/ufs/ufshcd.c @@ -2991,6 +2991,9 @@ int ufshcd_map_desc_id_to_length(struct ufs_hba *hba, case QUERY_DESC_IDN_RFU_1: *desc_len = 0; break; + case QUERY_DESC_IDN_HEALTH: + *desc_len = hba->desc_size.health_desc; + break; default: *desc_len = 0; return -EINVAL; @@ -6298,6 +6301,11 @@ static void ufshcd_init_desc_sizes(struct ufs_hba *hba) &hba->desc_size.geom_desc); if (err) hba->desc_size.geom_desc = QUERY_DESC_GEOMETRY_DEF_SIZE; + + err = ufshcd_read_desc_length(hba, QUERY_DESC_IDN_HEALTH, 0, + &hba->desc_size.health_desc); + if (err) + hba->desc_size.health_desc = QUERY_DESC_HEALTH_DEF_SIZE; } static void ufshcd_def_desc_sizes(struct ufs_hba *hba) @@ -6308,6 +6316,7 @@ static void ufshcd_def_desc_sizes(struct ufs_hba *hba) hba->desc_size.conf_desc = QUERY_DESC_CONFIGURATION_DEF_SIZE; hba->desc_size.unit_desc = QUERY_DESC_UNIT_DEF_SIZE; hba->desc_size.geom_desc = QUERY_DESC_GEOMETRY_DEF_SIZE; + hba->desc_size.health_desc = QUERY_DESC_HEALTH_DEF_SIZE; } /** @@ -7684,7 +7693,64 @@ static struct attribute *ufshcd_attrs[] = { NULL }; -ATTRIBUTE_GROUPS(ufshcd); +static const struct attribute_group ufshcd_default_group = { + .attrs = ufshcd_attrs, +}; + +static u8 ufshcd_read_desc_health(struct device *dev, int byte_offset) +{ + struct ufs_hba *hba = dev_get_drvdata(dev); + int buff_len = hba->desc_size.health_desc; + u8 desc_buf[hba->desc_size.health_desc]; + int err; + + if (byte_offset >= buff_len) + return 0; + + pm_runtime_get_sync(hba->dev); + err = ufshcd_read_desc(hba, QUERY_DESC_IDN_HEALTH, 0, + desc_buf, buff_len); + pm_runtime_put_sync(hba->dev); + if (err) + return 0; + + return desc_buf[byte_offset]; +} + +#define HEALTH_ATTR_RO(_name, _byte_offset) \ +static ssize_t _name##_show(struct device *_dev, \ + struct device_attribute *attr, char *buf) \ +{ \ + u8 byte = ufshcd_read_desc_health(_dev, _byte_offset); \ + return scnprintf(buf, PAGE_SIZE, "0x%02x\n", byte); \ +} \ +static DEVICE_ATTR_RO(_name); + +HEALTH_ATTR_RO(length, 0); +HEALTH_ATTR_RO(type, 1); +HEALTH_ATTR_RO(eol, 2); +HEALTH_ATTR_RO(lifetimeA, 3); +HEALTH_ATTR_RO(lifetimeB, 4); + +static struct attribute *ufshcd_health_attrs[] = { + &dev_attr_length.attr, + &dev_attr_type.attr, + &dev_attr_eol.attr, + &dev_attr_lifetimeA.attr, + &dev_attr_lifetimeB.attr, + NULL +}; + +static const struct attribute_group ufshcd_health_group = { + .name = "health", + .attrs = ufshcd_health_attrs, +}; + +static const struct attribute_group *ufshcd_groups[] = { + &ufshcd_default_group, + &ufshcd_health_group, + NULL, +}; static inline void ufshcd_add_sysfs_nodes(struct ufs_hba *hba) { diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h index 56e26eb15185..ba44cbcf755e 100644 --- a/drivers/scsi/ufs/ufshcd.h +++ b/drivers/scsi/ufs/ufshcd.h @@ -229,6 +229,7 @@ struct ufs_desc_size { int interc_desc; int unit_desc; int conf_desc; + int health_desc; }; /**
This patch adds a new sysfs group, namely health, via: /sys/devices/soc/X.ufshc/health/ This directory contains the below entries, each of which shows an 8-bytes hex number representing different meanings defined by JEDEC specfication. Users can simply read these entries to check how their underlying flash storage is getting reached out to its end of life. For example, if lifetimeA shows 0xb, it would be the right time to consider device swap. - length : must be 25h - type : must be 09h - eol 00h: Not defined 01h: Normal 02h: Warning 03h: Critical - lifetimeA/B 00h: Not defined 01h: 0% ~ 10% device life time used 02h: 10% ~ 20% device life time used 03h: 20% ~ 30% device life time used 04h: 30% ~ 40% device life time used 05h: 40% ~ 50% device life time used 06h: 50% ~ 60% device life time used 07h: 60% ~ 70% device life time used 08h: 70% ~ 80% device life time used 09h: 80% ~ 90% device life time used 0Ah: 90% ~ 100% device life time used 0Bh: Exceeded its maximum estimated device life time Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Jaegeuk Kim <jaegeuk@google.com> --- Change log from v3: - add '\n' at the end of print-out Documentation/ABI/testing/sysfs-devices-soc-ufs | 25 +++++++++ MAINTAINERS | 1 + drivers/scsi/ufs/ufs.h | 2 + drivers/scsi/ufs/ufshcd.c | 68 ++++++++++++++++++++++++- drivers/scsi/ufs/ufshcd.h | 1 + 5 files changed, 96 insertions(+), 1 deletion(-) create mode 100644 Documentation/ABI/testing/sysfs-devices-soc-ufs