DASD FC endpoint security
Subsystems (4)
Stefan Haberland
maintainerv2 cover DASD FC endpoint security
Hi Jens,
please apply the following patches that add a new DASD feature to show fibre channel endpoint security.
The patches apply on you for-next branch and are intended for 5.10.
The s390/cio patches should also go upstream through your tree this time since they are required for the DASD patches to build.
Best regards,
Stefan
v1->v2:
- fix memleak in "s390/dasd: Fix operational path inconsistency"
- add comment suggested by Cornelia Huck
Jan Höppner (7):
s390/dasd: Remove unused parameter from dasd_generic_probe() s390/dasd: Move duplicate code to separate function s390/dasd: Store path configuration data during path handling s390/dasd: Fix operational path inconsistency s390/dasd: Display FC Endpoint Security information via sysfs s390/dasd: Prepare for additional path event handling s390/dasd: Process FCES path event notification
Sebastian Ott (1):
s390/cio: Export information about Endpoint-Security Capability
Vineeth Vijayan (2):
s390/cio: Provide Endpoint-Security Mode per CU s390/cio: Add support for FCES status notification
arch/s390/include/asm/ccwdev.h | 2 + arch/s390/include/asm/cio.h | 1 + drivers/s390/block/dasd.c | 22 ++-- drivers/s390/block/dasd_devmap.c | 109 +++++++++++++++++++ drivers/s390/block/dasd_eckd.c | 175 +++++++++++++++++++++---------- drivers/s390/block/dasd_fba.c | 2 +- drivers/s390/block/dasd_int.h | 111 +++++++++++++++++++- drivers/s390/cio/chp.c | 15 +++ drivers/s390/cio/chp.h | 1 + drivers/s390/cio/chsc.c | 145 +++++++++++++++++++++++-- drivers/s390/cio/chsc.h | 3 +- drivers/s390/cio/device.c | 15 ++- 12 files changed, 523 insertions(+), 78 deletions(-)
Signature · 1 line
2.17.1
Stefan Haberland
maintainerPATCH v2 1/10 s390/cio: Export information about Endpoint-Security Capability
From: Sebastian Ott <[email protected]>
Add a new sysfs attribute 'esc' per chpid. This new attribute exports the Endpoint-Security-Capability byte of channel-path description block, which could be 0-None, 1-Authentication, 2 and 3-Encryption.
For example:
$ cat /sys/devices/css0/chp0.34/esc 0
--- drivers/s390/cio/chp.c | 15 +++++++++++++++ drivers/s390/cio/chsc.h | 3 ++- 2 files changed, 17 insertions(+), 1 deletion(-)
drivers/s390/cio/chp.c +15 -0
@@ -384,6 +384,20 @@ static ssize_t chp_chid_external_show(struct device *dev, } static DEVICE_ATTR(chid_external, 0444, chp_chid_external_show, NULL); +static ssize_t chp_esc_show (struct device *dev, +struct device_attribute *attr,char *buf) +{ +struct channel_path *chp =to_channelpath (dev); +ssize_t rc; + +mutex_lock (&chp->lock); + rc =sprintf (buf,"%x\n" , chp->desc_fmt1.esc); +mutex_unlock (&chp->lock); + +return rc; +} +static DEVICE_ATTR(esc, 0444, chp_esc_show, NULL); +static ssize_t util_string_read (struct file *filp,struct kobject *kobj,struct bin_attribute *attr,char *buf,loff_t off,size_t count)
@@ -414,6 +428,7 @@ static struct attribute *chp_attrs[] = { &dev_attr_shared.attr, &dev_attr_chid.attr, &dev_attr_chid_external.attr, + &dev_attr_esc.attr, NULL, }; static struct attribute_group chp_attr_group = {
drivers/s390/cio/chsc.h +2 -1
@@ -27,7 +27,8 @@ struct channel_path_desc_fmt1 {u8 lsn;u8 desc;u8 chpid; - u32:24 ; + u32:16 ; +u8 esc;u8 chpp;u32 unused[2 ];u16 chid;
Signature · 1 line
2.17.1
Stefan Haberland
maintainerPATCH v2 2/10 s390/cio: Provide Endpoint-Security Mode per CU
In reply to Stefan HaberlandFrom: Vineeth Vijayan <[email protected]>
Add an interface in the CIO layer to retrieve the information about the Endpoint-Security Mode (ESM) of the specified CU. The ESM values are defined as 0-None, 1-Authenticated or 2, 3-Encrypted.
--- arch/s390/include/asm/cio.h | 1 + drivers/s390/cio/chsc.c | 83 +++++++++++++++++++++++++++++++++++++ 2 files changed, 84 insertions(+)
arch/s390/include/asm/cio.h +1 -0
@@ -373,5 +373,6 @@ struct gen_pool *cio_gp_dma_create(struct device *dma_dev, int nr_pages);int chsc_sstpc (void *page,unsigned int op,u16 ctrl,u64 *clock_delta);int chsc_sstpi (void *page,void *result,size_t size);int chsc_sgib (u32 origin); +int chsc_scud (u16 cu,u64 *esm,u8 *esm_valid);#endif
drivers/s390/cio/chsc.c +83 -0
86 lines not on this page · raw message
Signature · 1 line
2.17.1
Stefan Haberland
maintainerPATCH v2 3/10 s390/cio: Add support for FCES status notification
In reply to Stefan HaberlandFrom: Vineeth Vijayan <[email protected]>
Fibre Channel Endpoint-Security event is received as an sei:nt0 type in the CIO layer. This information needs to be shared with the CCW device drivers using the path_events callback.
--- arch/s390/include/asm/ccwdev.h | 2 ++ drivers/s390/cio/chp.h | 1 + drivers/s390/cio/chsc.c | 62 +++++++++++++++++++++++++++++----- drivers/s390/cio/device.c | 15 +++++++- 4 files changed, 70 insertions(+), 10 deletions(-)
arch/s390/include/asm/ccwdev.h +2 -0
@@ -104,6 +104,8 @@ struct ccw_device {was successfully verified. */ #define PE_PATHGROUP_ESTABLISHED 0x4/* A pathgroup was reset and had to be established again. */ +#define PE_PATH_FCES_EVENT 0x8/* The FCES Status of a path has +* changed. */ /* * Possible CIO actions triggered by the unit check handler.
drivers/s390/cio/chp.h +1 -0
@@ -23,6 +23,7 @@#define CHP_OFFLINE 1#define CHP_VARY_ON 2#define CHP_VARY_OFF 3 +#define CHP_FCES_EVENT 4struct chp_link {struct chp_id chpid;
drivers/s390/cio/chsc.c +53 -9
98 lines not on this page · raw message
drivers/s390/cio/device.c +14 -1
@@ -1170,7 +1170,8 @@ static int io_subchannel_chp_event(struct subchannel *sch, struct chp_link *link, int event) {struct ccw_device *cdev =sch_get_cdev (sch); -int mask; +int mask, chpid, valid_bit; +int path_event[8 ]; mask =chp_ssd_get_mask (&sch->ssd_info, link);if (!mask)
@@ -1205,6 +1206,18 @@ static int io_subchannel_chp_event(struct subchannel *sch, cdev->private->path_new_mask |= mask;io_subchannel_verify (sch);break ; +case CHP_FCES_EVENT : +/* Forward Endpoint Security event */ +for (chpid =0 , valid_bit =0x80 ; chpid <8 ; chpid++, + valid_bit >>=1 ) { +if (mask & valid_bit) + path_event[chpid] =PE_PATH_FCES_EVENT ; +else + path_event[chpid] =PE_NONE ; + } +if (cdev) + cdev->drv->path_event (cdev, path_event); +break ; }return 0 ; }
Signature · 1 line
2.17.1
Stefan Haberland
maintainerPATCH v2 4/10 s390/dasd: Remove unused parameter from dasd_generic_probe()
In reply to Stefan HaberlandFrom: Jan Höppner <[email protected]>
The discipline argument in dasd_generic_probe() isn't used and there is no history how it was used in the past. Remove it.
--- drivers/s390/block/dasd.c | 3 +-- drivers/s390/block/dasd_eckd.c | 2 +- drivers/s390/block/dasd_fba.c | 2 +- drivers/s390/block/dasd_int.h | 2 +- 4 files changed, 4 insertions(+), 5 deletions(-)
drivers/s390/block/dasd.c +1 -2
@@ -3465,8 +3465,7 @@ static void dasd_generic_auto_online(void *data, async_cookie_t cookie)* Initial attempt at a probe function. this can be simplified once * the other detection code is gone. */ -int dasd_generic_probe (struct ccw_device *cdev, -struct dasd_discipline *discipline) +int dasd_generic_probe (struct ccw_device *cdev) {int ret;
drivers/s390/block/dasd_eckd.c +1 -1
@@ -143,7 +143,7 @@ dasd_eckd_probe (struct ccw_device *cdev) "ccw-device options");return ret; } - ret =dasd_generic_probe (cdev, &dasd_eckd_discipline); + ret =dasd_generic_probe (cdev);return ret; }
drivers/s390/block/dasd_fba.c +1 -1
@@ -58,7 +58,7 @@ static struct ccw_driver dasd_fba_driver; /* see below */static int dasd_fba_probe (struct ccw_device *cdev) { -return dasd_generic_probe (cdev, &dasd_fba_discipline); +return dasd_generic_probe (cdev); } static int
drivers/s390/block/dasd_int.h +1 -1
@@ -774,7 +774,7 @@ void dasd_block_set_timer(struct dasd_block *, int);void dasd_block_clear_timer (struct dasd_block *);int dasd_cancel_req (struct dasd_ccw_req *);int dasd_flush_device_queue (struct dasd_device *); -int dasd_generic_probe (struct ccw_device *,struct dasd_discipline *); +int dasd_generic_probe (struct ccw_device *);void dasd_generic_free_discipline (struct dasd_device *);void dasd_generic_remove (struct ccw_device *cdev);int dasd_generic_set_online (struct ccw_device *,struct dasd_discipline *);
Signature · 1 line
2.17.1
Re: [PATCH v2 04/10] s390/dasd: Remove unused parameter from dasd_generic_probe()
Quoted 8 lines · On Thu, 8 Oct 2020 15:13:30 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
The discipline argument in
dasd_generic_probe()isn't used and there is no history how it was used in the past. Remove it.Signed-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>Signed-off-by Stefan Haberland <[email protected]>
Quoted 6 lines
--- drivers/s390/block/dasd.c | 3 +-- drivers/s390/block/dasd_eckd.c | 2 +- drivers/s390/block/dasd_fba.c | 2 +- drivers/s390/block/dasd_int.h | 2 +- 4 files changed, 4 insertions(+), 5 deletions(-)
Stefan Haberland
maintainerPATCH v2 5/10 s390/dasd: Move duplicate code to separate function
In reply to Stefan HaberlandFrom: Jan Höppner <[email protected]>
For storing retrieved path information both the if and else block in dasd_eckd_read_conf() use the same code. To avoid duplicate code this should be done after the if/else block. To further increase readability, move the code to a new function, dasd_eckd_store_conf_data().
--- drivers/s390/block/dasd_eckd.c | 42 ++++++++++++++++------------------ 1 file changed, 20 insertions(+), 22 deletions(-)
drivers/s390/block/dasd_eckd.c +20 -22
@@ -1000,6 +1000,22 @@ static unsigned char dasd_eckd_path_access(void *conf_data, int conf_len)return 0 ; } +static void dasd_eckd_store_conf_data (struct dasd_device *device, +struct dasd_conf_data *conf_data,int chp) +{ +struct channel_path_desc_fmt0 *chp_desc; +struct subchannel_id sch_id; + +ccw_device_get_schid (device->cdev, &sch_id); + device->path[chp].conf_data = conf_data; + device->path[chp].cssid = sch_id.cssid; + device->path[chp].ssid = sch_id.ssid; + chp_desc =ccw_device_get_chp_desc (device->cdev, chp); +if (chp_desc) + device->path[chp].chpid = chp_desc->chpid; +kfree (chp_desc); +} +static void dasd_eckd_clear_conf_data (struct dasd_device *device) {struct dasd_eckd_private *private = device->private;
@@ -1016,7 +1032,6 @@ static void dasd_eckd_clear_conf_data(struct dasd_device *device) } } -static int dasd_eckd_read_conf (struct dasd_device *device) {void *conf_data;
@@ -1026,12 +1041,9 @@ static int dasd_eckd_read_conf(struct dasd_device *device)struct dasd_eckd_private *private, path_private;struct dasd_uid *uid;char print_path_uid[60 ], print_device_uid[60 ]; -struct channel_path_desc_fmt0 *chp_desc; -struct subchannel_id sch_id; private = device->private; opm =ccw_device_get_path_mask (device->cdev); -ccw_device_get_schid (device->cdev, &sch_id); conf_data_saved =0 ; path_err =0 ;/* get configuration data per operational path */
@@ -1066,15 +1078,6 @@ static int dasd_eckd_read_conf(struct dasd_device *device)kfree (conf_data);continue ; } - pos =pathmask_to_pos (lpm); -/* store per path conf_data */ - device->path[pos].conf_data = conf_data; - device->path[pos].cssid = sch_id.cssid; - device->path[pos].ssid = sch_id.ssid; - chp_desc =ccw_device_get_chp_desc (device->cdev, pos); -if (chp_desc) - device->path[pos].chpid = chp_desc->chpid; -kfree (chp_desc);/* * build device UID that other path data * can be compared to it
@@ -1132,18 +1135,13 @@ static int dasd_eckd_read_conf(struct dasd_device *device)dasd_path_add_cablepm (device, lpm);continue ; } - pos =pathmask_to_pos (lpm); -/* store per path conf_data */ - device->path[pos].conf_data = conf_data; - device->path[pos].cssid = sch_id.cssid; - device->path[pos].ssid = sch_id.ssid; - chp_desc =ccw_device_get_chp_desc (device->cdev, pos); -if (chp_desc) - device->path[pos].chpid = chp_desc->chpid; -kfree (chp_desc); path_private.conf_data =NULL ; path_private.conf_len =0 ; } + + pos =pathmask_to_pos (lpm); +dasd_eckd_store_conf_data (device, conf_data, pos); +switch (dasd_eckd_path_access (conf_data, conf_len)) {case 0x02 :dasd_path_add_nppm (device, lpm);
Signature · 1 line
2.17.1
Re: [PATCH v2 05/10] s390/dasd: Move duplicate code to separate function
Quoted 7 lines · On Thu, 8 Oct 2020 15:13:31 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
For storing retrieved path information both the if and else block in
dasd_eckd_read_conf()use the same code. To avoid duplicate code this should be done after the if/else block. To further increase readability, move the code to a new function,dasd_eckd_store_conf_data().
Quoted 6 lines
Signed-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>Signed-off-by Stefan Haberland <[email protected]>--- drivers/s390/block/dasd_eckd.c | 42 ++++++++++++++++------------------ 1 file changed, 20 insertions(+), 22 deletions(-)
Stefan Haberland
maintainerPATCH v2 6/10 s390/dasd: Store path configuration data during path handling
In reply to Stefan HaberlandFrom: Jan Höppner <[email protected]>
Currently, the configuration data for a path is retrieved during a path verification and used only temporarily. If a path is newly added to the I/O setup after a boot, no configuration data will be stored for this particular path.
However, this data is required for later use and should be present for a valid I/O path anyway. Store this data during the path verification so that newly added paths can provide all information necessary.
--- drivers/s390/block/dasd_eckd.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-)
drivers/s390/block/dasd_eckd.c +15 -1
@@ -1007,6 +1007,11 @@ static void dasd_eckd_store_conf_data(struct dasd_device *device,struct subchannel_id sch_id;ccw_device_get_schid (device->cdev, &sch_id); +/* +* path handling and read_conf allocate data +* free it before replacing the pointer +*/ +kfree (device->path[chp].conf_data); device->path[chp].conf_data = conf_data; device->path[chp].cssid = sch_id.cssid; device->path[chp].ssid = sch_id.ssid;
@@ -1263,9 +1268,10 @@ static void do_path_verification_work(struct work_struct *work)struct dasd_uid *uid;__u8 path_rcd_buf[DASD_ECKD_RCD_DATA_SIZE ];__u8 lpm, opm, npm, ppm, epm, hpfpm, cablepm; +struct dasd_conf_data *conf_data;unsigned long flags;char print_uid[60 ]; -int rc; +int rc, pos; data = container_of(work, struct path_verification_work_data, worker); device = data->device;
@@ -1395,6 +1401,14 @@ static void do_path_verification_work(struct work_struct *work) } } + conf_data =kzalloc (DASD_ECKD_RCD_DATA_SIZE ,GFP_KERNEL ); +if (conf_data) { +memcpy (conf_data, data->rcd_buffer, +DASD_ECKD_RCD_DATA_SIZE ); + } + pos =pathmask_to_pos (lpm); +dasd_eckd_store_conf_data (device, conf_data, pos); +/* * There is a small chance that a path is lost again between * above path verification and the following modification of
Signature · 1 line
2.17.1
Re: [PATCH v2 06/10] s390/dasd: Store path configuration data during path handling
Quoted 10 lines · On Thu, 8 Oct 2020 15:13:32 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
Currently, the configuration data for a path is retrieved during a path verification and used only temporarily. If a path is newly added to the I/O setup after a boot, no configuration data will be stored for this particular path.
However, this data is required for later use and should be present for a valid I/O path anyway. Store this data during the path verification so that newly added paths can provide all information necessary.
Quoted 7 lines
Signed-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>[[email protected]: fix conf_data memleak]Signed-off-by Stefan Haberland <[email protected]>--- drivers/s390/block/dasd_eckd.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-)
Stefan Haberland
maintainerPATCH v2 7/10 s390/dasd: Fix operational path inconsistency
In reply to Stefan HaberlandFrom: Jan Höppner <[email protected]>
During online processing and setting up a DASD device, the configuration data for operational paths is read and validated two times (dasd_eckd_read_conf()). The first time to provide information that are necessary for the LCU setup. A second time after the LCU setup as a device might report different configuration data then.
When the configuration setup for each operational path is being validated, an initial call to dasd_eckd_clear_conf_data() is issued. This call wipes all previously available configuration data and path information for each path.
However, the operational path mask is not updated during this process.
As a result, the stored operational path mask might no longer correspond to the operational paths mask reported by the CIO layer, as several paths might be gone between the two dasd_eckd_read_conf() calls.
This inconsistency leads to more severe issues in later path handling changes. Fix this by removing the channel paths from the operational path mask during the dasd_eckd_clear_conf_data() call.
--- drivers/s390/block/dasd_eckd.c | 1 + 1 file changed, 1 insertion(+)
drivers/s390/block/dasd_eckd.c +1 -0
@@ -1034,6 +1034,7 @@ static void dasd_eckd_clear_conf_data(struct dasd_device *device) device->path[i].cssid =0 ; device->path[i].ssid =0 ; device->path[i].chpid =0 ; +dasd_path_notoper (device, i); } }
Signature · 1 line
2.17.1
Re: [PATCH v2 07/10] s390/dasd: Fix operational path inconsistency
Quoted 34 lines · On Thu, 8 Oct 2020 15:13:33 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
During online processing and setting up a DASD device, the configuration data for operational paths is read and validated two times (
dasd_eckd_read_conf()). The first time to provide information that are necessary for the LCU setup. A second time after the LCU setup as a device might report different configuration data then.When the configuration setup for each operational path is being validated, an initial call to
dasd_eckd_clear_conf_data()is issued. This call wipes all previously available configuration data and path information for each path.
However, the operational path mask is not updated during this process.As a result, the stored operational path mask might no longer correspond to the operational paths mask reported by the CIO layer, as several paths might be gone between the two
dasd_eckd_read_conf()calls.This inconsistency leads to more severe issues in later path handling changes. Fix this by removing the channel paths from the operational path mask during the
dasd_eckd_clear_conf_data()call.Signed-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>Signed-off-by Stefan Haberland <[email protected]>--- drivers/s390/block/dasd_eckd.c | 1 + 1 file changed, 1 insertion(+)drivers/s390/block/dasd_eckd.c +0
-0
drivers/s390/block/dasd_eckd.c static void dasd_eckd_clear_conf_data(struct dasd_device *device) · line 1040
Quoted 7 lines
drivers/s390/block/dasd_eckd.c +1
-0@@ -1034,6 +1034,7 @@ static void dasd_eckd_clear_conf_data(struct dasd_device *device) device->path[i].cssid =0 ; device->path[i].ssid =0 ; device->path[i].chpid =0 ; +dasd_path_notoper (device, i); } }
Stefan Haberland
maintainerPATCH v2 8/10 s390/dasd: Display FC Endpoint Security information via sysfs
In reply to Stefan HaberlandFrom: Jan Höppner <[email protected]>
Add a new sysfs attribute (fc_security) per device and per operational channel path. The information of the current FC Endpoint Security state is received through the CIO layer.
The state of the FC Endpoint Security can be either "Unsupported", "Authentication", or "Encryption".
For example:
$ cat /sys/bus/ccw/devices/0.0.c600/fc_security Encryption
If any of the operational paths is in a state different from all others, the device sysfs attribute will display the additional state "Inconsistent".
The sysfs attributes per paths are organised in a new directory called "paths_info" with subdirectories for each path.
/sys/bus/ccw/devices/0.0.c600/paths_info/
├── 0.38
│ └── fc_security
├── 0.39
│ └── fc_security
├── 0.3a
│ └── fc_security
└── 0.3b
└── fc_security--- drivers/s390/block/dasd_devmap.c | 109 +++++++++++++++++++++++++++++++ drivers/s390/block/dasd_eckd.c | 30 +++++++++ drivers/s390/block/dasd_int.h | 68 +++++++++++++++++++ 3 files changed, 207 insertions(+)
drivers/s390/block/dasd_devmap.c +109 -0
@@ -576,6 +576,11 @@ dasd_create_device(struct ccw_device *cdev)dev_set_drvdata (&cdev->dev, device);spin_unlock_irqrestore (get_ccwdev_lock (cdev), flags); + device->paths_info =kset_create_and_add ("paths_info" ,NULL , + &device->cdev->dev.kobj); +if (!device->paths_info) +dev_warn (&cdev->dev,"Could not create paths_info kset\n" ); +return device; }
@@ -622,6 +627,9 @@ dasd_delete_device(struct dasd_device *device)wait_event (dasd_delete_wq,atomic_read (&device->ref_count) ==0 );dasd_generic_free_discipline (device); + +kset_unregister (device->paths_info); +/* Disconnect dasd_device structure from ccw_device structure. */ cdev = device->cdev; device->cdev =NULL ;
@@ -1641,6 +1649,39 @@ dasd_path_interval_store(struct device *dev, struct device_attribute *attr, static DEVICE_ATTR(path_interval, 0644, dasd_path_interval_show, dasd_path_interval_store); +static ssize_t +dasd_device_fcs_show (struct device *dev,struct device_attribute *attr, +char *buf) +{ +struct dasd_device *device; +int fc_sec; +int rc; + + device =dasd_device_from_cdev (to_ccwdev (dev)); +if (IS_ERR (device)) +return -ENODEV ; + fc_sec =dasd_path_get_fcs_device (device); +if (fc_sec == -EINVAL ) + rc =snprintf (buf,PAGE_SIZE ,"Inconsistent\n" ); +else + rc =snprintf (buf,PAGE_SIZE ,"%s\n" ,dasd_path_get_fcs_str (fc_sec)); +dasd_put_device (device); + +return rc; +} +static DEVICE_ATTR(fc_security, 0444, dasd_device_fcs_show, NULL); + +static ssize_t +dasd_path_fcs_show (struct kobject *kobj,struct kobj_attribute *attr,char *buf) +{ +struct dasd_path *path =to_dasd_path (kobj); +unsigned int fc_sec = path->fc_security; + +return snprintf (buf,PAGE_SIZE ,"%s\n" ,dasd_path_get_fcs_str (fc_sec)); +} + +static struct kobj_attribute path_fcs_attribute = +__ATTR (fc_security,0444 , dasd_path_fcs_show,NULL );#define DASD_DEFINE_ATTR (_name, _func) \ static ssize_t dasd_##_name##_show(struct device *dev, \
@@ -1697,6 +1738,7 @@ static struct attribute * dasd_attrs[] = { &dev_attr_path_reset.attr, &dev_attr_hpf.attr, &dev_attr_ese.attr, + &dev_attr_fc_security.attr, NULL, };
@@ -1777,6 +1819,73 @@ dasd_set_feature(struct ccw_device *cdev, int feature, int flag) }EXPORT_SYMBOL (dasd_set_feature); +static struct attribute *paths_info_attrs[] = { + &path_fcs_attribute.attr, +NULL , +}; + +static struct kobj_type path_attr_type = { + .release = dasd_path_release, + .default_attrs = paths_info_attrs, + .sysfs_ops = &kobj_sysfs_ops, +}; + +static void dasd_path_init_kobj (struct dasd_device *device,int chp) +{ + device->path[chp].kobj.kset = device->paths_info; +kobject_init (&device->path[chp].kobj, &path_attr_type); +} + +void dasd_path_create_kobj (struct dasd_device *device,int chp) +{ +int rc; + +if (test_bit (DASD_FLAG_OFFLINE , &device->flags)) +return ; +if (!device->paths_info) { +dev_warn (&device->cdev->dev,"Unable to create paths objects\n" ); +return ; + } +if (device->path[chp].in_sysfs) +return ; +if (!device->path[chp].conf_data) +return ; + +dasd_path_init_kobj (device, chp); + + rc =kobject_add (&device->path[chp].kobj,NULL ,"%x.%02x" , + device->path[chp].cssid, device->path[chp].chpid); +if (rc) +kobject_put (&device->path[chp].kobj); + device->path[chp].in_sysfs = true; +} +EXPORT_SYMBOL (dasd_path_create_kobj); + +void dasd_path_create_kobjects (struct dasd_device *device) +{ +u8 lpm, opm; + + opm =dasd_path_get_opm (device); +for (lpm =0x80 ; lpm; lpm >>=1 ) { +if (!(lpm & opm)) +continue ; +dasd_path_create_kobj (device,pathmask_to_pos (lpm)); + } +} +EXPORT_SYMBOL (dasd_path_create_kobjects); + +/* +* As we keep kobjects for the lifetime of a device, this function must not be +* called anywhere but in the context of offlining a device. +*/ +void dasd_path_remove_kobj (struct dasd_device *device,int chp) +{ +if (device->path[chp].in_sysfs) { +kobject_put (&device->path[chp].kobj); + device->path[chp].in_sysfs = false; + } +} +EXPORT_SYMBOL (dasd_path_remove_kobj);int dasd_add_sysfs_files (struct ccw_device *cdev) {
drivers/s390/block/dasd_eckd.c +30 -0
@@ -1035,6 +1035,30 @@ static void dasd_eckd_clear_conf_data(struct dasd_device *device) device->path[i].ssid =0 ; device->path[i].chpid =0 ;dasd_path_notoper (device, i); +dasd_path_remove_kobj (device, i); + } +} + +static void dasd_eckd_read_fc_security (struct dasd_device *device) +{ +struct dasd_eckd_private *private = device->private; +u8 esm_valid; +u8 esm[8 ]; +int chp; +int rc; + + rc =chsc_scud (private->uid.ssid, (u64 *)esm, &esm_valid); +if (rc) { +for (chp =0 ; chp <8 ; chp++) + device->path[chp].fc_security =0 ; +return ; + } + +for (chp =0 ; chp <8 ; chp++) { +if (esm_valid & (0x80 >> chp)) + device->path[chp].fc_security = esm[chp]; +else + device->path[chp].fc_security =0 ; } }
@@ -1164,6 +1188,8 @@ static int dasd_eckd_read_conf(struct dasd_device *device) } } +dasd_eckd_read_fc_security (device); +return path_err; }
@@ -1430,6 +1456,8 @@ static void do_path_verification_work(struct work_struct *work)dasd_path_add_cablepm (device, cablepm);dasd_path_add_nohpfpm (device, hpfpm);spin_unlock_irqrestore (get_ccwdev_lock (device->cdev), flags); + +dasd_path_create_kobj (device, pos); }clear_bit (DASD_FLAG_PATH_VERIFY , &device->flags);dasd_put_device (device);
@@ -2069,6 +2097,8 @@ dasd_eckd_check_characteristics(struct dasd_device *device)if (rc) goto out_err3; +dasd_path_create_kobjects (device); +/* Read Feature Codes */ dasd_eckd_read_features (device);
drivers/s390/block/dasd_int.h +68 -0
100 lines not on this page · raw message
Signature · 1 line
2.17.1
Re: [PATCH v2 08/10] s390/dasd: Display FC Endpoint Security information via sysfs
Quoted 38 lines · On Thu, 8 Oct 2020 15:13:34 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
Add a new sysfs attribute (
fc_security) per device and per operational channel path. The information of the current FC Endpoint Security state is received through the CIO layer.The state of the FC Endpoint Security can be either "Unsupported", "Authentication", or "Encryption".
For example:
$ cat /sys/bus/ccw/devices/0.0.c600/fc_security EncryptionIf any of the operational paths is in a state different from all others, the device sysfs attribute will display the additional state "Inconsistent".
The sysfs attributes per paths are organised in a new directory called "
paths_info" with subdirectories for each path.
/sys/bus/ccw/devices/0.0.c600/paths_info/├── 0.38 │ └── fc_security ├── 0.39 │ └── fc_security ├── 0.3a │ └── fc_security └── 0.3b └── fc_securitySigned-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>Signed-off-by Stefan Haberland <[email protected]>--- drivers/s390/block/dasd_devmap.c | 109 +++++++++++++++++++++++++++++++ drivers/s390/block/dasd_eckd.c | 30 +++++++++ drivers/s390/block/dasd_int.h | 68 +++++++++++++++++++ 3 files changed, 207 insertions(+)
(...)
Quoted 4 lines
+/*
+ * As we keep kobjects for the lifetime of a device, this function must not be
+ * called anywhere but in the context of offlining a device.
+ */
Works for me :)
Quoted 8 lines
drivers/s390/block/dasd_devmap.c +8
-0+void dasd_path_remove_kobj (struct dasd_device *device,int chp) +{ +if (device->path[chp].in_sysfs) { +kobject_put (&device->path[chp].kobj); + device->path[chp].in_sysfs = false; + } +} +EXPORT_SYMBOL (dasd_path_remove_kobj);
(...)
Stefan Haberland
maintainerPATCH v2 9/10 s390/dasd: Prepare for additional path event handling
In reply to Stefan HaberlandFrom: Jan Höppner <[email protected]>
As more path events need to be handled for ECKD the current path verification infrastructure can be reused. Rename all path verifcation code to fit the more broadly based task of path event handling and put the path verification in a new separate function.
--- drivers/s390/block/dasd.c | 4 +- drivers/s390/block/dasd_eckd.c | 78 +++++++++++++++++++--------------- drivers/s390/block/dasd_int.h | 1 + 3 files changed, 47 insertions(+), 36 deletions(-)
drivers/s390/block/dasd.c +2 -2
@@ -2115,8 +2115,8 @@ static void __dasd_device_check_path_events(struct dasd_device *device)if (device->stopped & ~(DASD_STOPPED_DC_WAIT |DASD_UNRESUMED_PM ))return ; - rc = device->discipline->verify_path (device, -dasd_path_get_tbvpm (device)); + rc = device->discipline->pe_handler (device, +dasd_path_get_tbvpm (device));if (rc)dasd_device_set_timer (device,50 );else
drivers/s390/block/dasd_eckd.c +44 -34
162 lines not on this page · raw message
drivers/s390/block/dasd_int.h +1 -0
@@ -298,6 +298,7 @@ struct dasd_discipline {* configuration. */ int (*verify_path)(struct dasd_device *,__u8 ); +int (*pe_handler)(struct dasd_device *,__u8 );/* * Last things to do when a device is set online, and first things
Signature · 1 line
2.17.1
Re: [PATCH v2 09/10] s390/dasd: Prepare for additional path event handling
Quoted 7 lines · On Thu, 8 Oct 2020 15:13:35 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
As more path events need to be handled for ECKD the current path verification infrastructure can be reused. Rename all path verifcation code to fit the more broadly based task of path event handling and put the path verification in a new separate function.
Quoted 8 lines
Signed-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>Signed-off-by Stefan Haberland <[email protected]>--- drivers/s390/block/dasd.c | 4 +- drivers/s390/block/dasd_eckd.c | 78 +++++++++++++++++++--------------- drivers/s390/block/dasd_int.h | 1 + 3 files changed, 47 insertions(+), 36 deletions(-)
Stefan Haberland
maintainerPATCH v2 10/10 s390/dasd: Process FCES path event notification
In reply to Stefan HaberlandFrom: Jan Höppner <[email protected]>
If the Fibre Channel Endpoint-Security status of a path changes, a corresponding path event is received from the CIO layer.
Process this event by re-reading the FCES information.
As the information is retrieved for all paths on a single CU in one call, the internal status can also be updated for all paths and no processing per path is necessary.
--- drivers/s390/block/dasd.c | 19 +++++++++++---- drivers/s390/block/dasd_eckd.c | 12 +++++++--- drivers/s390/block/dasd_int.h | 42 +++++++++++++++++++++++++++++++++- 3 files changed, 64 insertions(+), 9 deletions(-)
drivers/s390/block/dasd.c +14 -5
@@ -2107,20 +2107,25 @@ static void __dasd_device_start_head(struct dasd_device *device)static void __dasd_device_check_path_events (struct dasd_device *device) { +__u8 tbvpm, fcsecpm;int rc; -if (!dasd_path_get_tbvpm (device)) + tbvpm =dasd_path_get_tbvpm (device); + fcsecpm =dasd_path_get_fcsecpm (device); + +if (!tbvpm && !fcsecpm)return ;if (device->stopped & ~(DASD_STOPPED_DC_WAIT |DASD_UNRESUMED_PM ))return ; - rc = device->discipline->pe_handler (device, -dasd_path_get_tbvpm (device)); -if (rc) + rc = device->discipline->pe_handler (device, tbvpm, fcsecpm); +if (rc) {dasd_device_set_timer (device,50 ); -else + }else {dasd_path_clear_all_verify (device); +dasd_path_clear_all_fcsec (device); + } };/*
@@ -3868,6 +3873,10 @@ void dasd_generic_path_event(struct ccw_device *cdev, int *path_event)if (device->discipline->kick_validate) device->discipline->kick_validate (device); } +if (path_event[chp] &PE_PATH_FCES_EVENT ) { +dasd_path_fcsec_update (device, chp); +dasd_schedule_device_bh (device); + } } hpfpm =dasd_path_get_hpfpm (device); ifccpm =dasd_path_get_ifccpm (device);
drivers/s390/block/dasd_eckd.c +9 -3
@@ -111,6 +111,7 @@ struct pe_handler_work_data {__u8 rcd_buffer[DASD_ECKD_RCD_DATA_SIZE ];int isglobal;__u8 tbvpm; +__u8 fcsecpm; };static struct pe_handler_work_data *pe_handler_worker;static DEFINE_MUTEX (dasd_pe_handler_mutex);
@@ -1466,7 +1467,10 @@ static void do_pe_handler_work(struct work_struct *work)return ; } -dasd_eckd_path_available_action (device, data); +if (data->tbvpm) +dasd_eckd_path_available_action (device, data); +if (data->fcsecpm) +dasd_eckd_read_fc_security (device);clear_bit (DASD_FLAG_PATH_VERIFY , &device->flags);dasd_put_device (device);
@@ -1476,7 +1480,8 @@ static void do_pe_handler_work(struct work_struct *work)kfree (data); } -static int dasd_eckd_pe_handler (struct dasd_device *device,__u8 lpm) +static int dasd_eckd_pe_handler (struct dasd_device *device, +__u8 tbvpm,__u8 fcsecpm) {struct pe_handler_work_data *data;
@@ -1495,7 +1500,8 @@ static int dasd_eckd_pe_handler(struct dasd_device *device, __u8 lpm)INIT_WORK (&data->worker, do_pe_handler_work);dasd_get_device (device); data->device = device; - data->tbvpm = lpm; + data->tbvpm = tbvpm; + data->fcsecpm = fcsecpm;schedule_work (&data->worker);return 0 ; }
drivers/s390/block/dasd_int.h +41 -1
72 lines not on this page · raw message
Signature · 1 line
2.17.1
Re: [PATCH v2 10/10] s390/dasd: Process FCES path event notification
Quoted 11 lines · On Thu, 8 Oct 2020 15:13:36 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
If the Fibre Channel Endpoint-Security status of a path changes, a corresponding path event is received from the CIO layer.
Process this event by re-reading the FCES information.
As the information is retrieved for all paths on a single CU in one call, the internal status can also be updated for all paths and no processing per path is necessary.
Quoted 8 lines
Signed-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>Signed-off-by Stefan Haberland <[email protected]>--- drivers/s390/block/dasd.c | 19 +++++++++++---- drivers/s390/block/dasd_eckd.c | 12 +++++++--- drivers/s390/block/dasd_int.h | 42 +++++++++++++++++++++++++++++++++- 3 files changed, 64 insertions(+), 9 deletions(-)
Jan Höppner
maintainerRe: [PATCH v2 10/10] s390/dasd: Process FCES path event notification
Quoted 23 lines · Cornelia Huck
Quoted 11 lines · On Thu, 8 Oct 2020 15:13:36 +0200 Stefan Haberland <[email protected]> wrote:
From: Jan Höppner <[email protected]>
If the Fibre Channel Endpoint-Security status of a path changes, a corresponding path event is received from the CIO layer.
Process this event by re-reading the FCES information.
As the information is retrieved for all paths on a single CU in one call, the internal status can also be updated for all paths and no processing per path is necessary.
Quoted 8 lines
Signed-off-by Jan Höppner <[email protected]>Reviewed-by Stefan Haberland <[email protected]>Signed-off-by Stefan Haberland <[email protected]>--- drivers/s390/block/dasd.c | 19 +++++++++++---- drivers/s390/block/dasd_eckd.c | 12 +++++++--- drivers/s390/block/dasd_int.h | 42 +++++++++++++++++++++++++++++++++- 3 files changed, 64 insertions(+), 9 deletions(-)
Quoted 2 lines
Reviewed-by Cornelia Huck <[email protected]>
Thanks a lot for reviewing the whole set, much appreciated!
regards,
Jan
Stefan Haberland
maintainerHi Jens,
quick ping. Are you going to apply this for 5.10?
Best regards,
Stefan
Stefan Haberland <[email protected]> schrieb am Thu, 08. Oct 15:13:
Quoted 50 lines
Hi Jens,
please apply the following patches that add a new DASD feature to show fibre channel endpoint security.
The patches apply on you for-next branch and are intended for 5.10.
The s390/cio patches should also go upstream through your tree this time since they are required for the DASD patches to build.
Best regards,
Stefan
v1->v2:
- fix memleak in "s390/dasd: Fix operational path inconsistency"
- add comment suggested by Cornelia HuckJan Höppner (7):
s390/dasd: Remove unused parameter fromdasd_generic_probe()s390/dasd: Move duplicate code to separate function s390/dasd: Store path configuration data during path handling s390/dasd: Fix operational path inconsistency s390/dasd: Display FC Endpoint Security information via sysfs s390/dasd: Prepare for additional path event handling s390/dasd: Process FCES path event notificationSebastian Ott (1):
s390/cio: Export information about Endpoint-Security CapabilityVineeth Vijayan (2):
s390/cio: Provide Endpoint-Security Mode per CU s390/cio: Add support for FCES status notificationarch/s390/include/asm/ccwdev.h | 2 + arch/s390/include/asm/cio.h | 1 + drivers/s390/block/dasd.c | 22 ++-- drivers/s390/block/dasd_devmap.c | 109 +++++++++++++++++++ drivers/s390/block/dasd_eckd.c | 175 +++++++++++++++++++++---------- drivers/s390/block/dasd_fba.c | 2 +- drivers/s390/block/dasd_int.h | 111 +++++++++++++++++++- drivers/s390/cio/chp.c | 15 +++ drivers/s390/cio/chp.h | 1 + drivers/s390/cio/chsc.c | 145 +++++++++++++++++++++++-- drivers/s390/cio/chsc.h | 3 +- drivers/s390/cio/device.c | 15 ++- 12 files changed, 523 insertions(+), 78 deletions(-)Signature · 1 line
2.17.1
Jens Axboe
maintainerQuoted 3 lines · Stefan Haberland
Hi Jens,
quick ping. Are you going to apply this for 5.10?
I actually wasn't planning on it - it arrived a bit late, and it seemed like one of those things that needed a bit more review talk before being able to be applied.
Signature · 1 line
Jens Axboe
Stefan Haberland
maintainerAm 12.10.20 um 21:33 schrieb Jens Axboe:
Quoted 8 lines · Stefan Haberland
Quoted 3 lines
Hi Jens,
quick ping. Are you going to apply this for 5.10?
I actually wasn't planning on it - it arrived a bit late, and it seemed like one of those things that needed a bit more review talk before being able to be applied.
OK, too bad. I had hoped that this was still OK.
The patches have been tested and reviewed internally for quite a while. Which actually was the reason for the late submission. Cornelia also gave her RB last week.
But OK, if you think this needs some more review we will have to wait for 5.11.
Jens Axboe
maintainerQuoted 10 lines · Stefan Haberland
Am 12.10.20 um 21:33 schrieb Jens Axboe:
Quoted 8 lines · Stefan Haberland
Quoted 3 lines
Hi Jens,
quick ping. Are you going to apply this for 5.10?
I actually wasn't planning on it - it arrived a bit late, and it seemed like one of those things that needed a bit more review talk before being able to be applied.
Quoted 4 lines
OK, too bad. I had hoped that this was still OK.
The patches have been tested and reviewed internally for quite a while. Which actually was the reason for the late submission. Cornelia also gave her RB last week.
I'm not worried about the stability of it as much as whether the special feature is warranted. From the former point of view, it's probably fine to go in now.
Quoted 2 lines
But OK, if you think this needs some more review we will have to wait for 5.11.
I'd definitely feel more comfortable with that.
Signature · 1 line
Jens Axboe
Stefan Haberland
maintainerAm 13.10.20 um 21:40 schrieb Jens Axboe:
Quoted 14 lines · Stefan Haberland
Quoted 5 lines
Am 12.10.20 um 21:33 schrieb Jens Axboe:
Quoted 4 lines · Stefan Haberland
Quoted 3 lines
Hi Jens,
quick ping. Are you going to apply this for 5.10?
Quoted 8 lines
Quoted 4 lines
I actually wasn't planning on it - it arrived a bit late, and it seemed like one of those things that needed a bit more review talk before being able to be applied.
OK, too bad. I had hoped that this was still OK.
The patches have been tested and reviewed internally for quite a while. Which actually was the reason for the late submission. Cornelia also gave her RB last week.
Quoted 7 lines
I'm not worried about the stability of it as much as whether the special feature is warranted. From the former point of view, it's probably fine to go in now.
Quoted 2 lines
But OK, if you think this needs some more review we will have to wait for 5.11.
I'd definitely feel more comfortable with that.
OK, I will take care that features will be sent earlier next time.
So, instead could you please apply the patches for 5.11 as soon as it is suitable?
Jens Axboe
maintainerQuoted 18 lines · Stefan Haberland
Am 13.10.20 um 21:40 schrieb Jens Axboe:
Quoted 17 lines · Stefan Haberland
Quoted 5 lines
Am 12.10.20 um 21:33 schrieb Jens Axboe:
Quoted 4 lines · Stefan Haberland
Quoted 3 lines
Hi Jens,
quick ping. Are you going to apply this for 5.10?
Quoted 8 lines
Quoted 4 lines
I actually wasn't planning on it - it arrived a bit late, and it seemed like one of those things that needed a bit more review talk before being able to be applied.
OK, too bad. I had hoped that this was still OK.
The patches have been tested and reviewed internally for quite a while. Which actually was the reason for the late submission. Cornelia also gave her RB last week.I'm not worried about the stability of it as much as whether the special feature is warranted. From the former point of view, it's probably fine to go in now.
Quoted 6 lines
Quoted 4 lines
Quoted 2 lines
But OK, if you think this needs some more review we will have to wait for 5.11.
I'd definitely feel more comfortable with that.
OK, I will take care that features will be sent earlier next time.
Thanks, ideally I like to have new stuff like that in my tree (and for-next) for at least a week prior to the merge window opening.
Quoted 2 lines
So, instead could you please apply the patches for 5.11 as soon as it is suitable?
I will - I have it queued up, won't create anything public until we get past the merge window.
Signature · 1 line
Jens Axboe
Stefan Haberland
maintainerAm 14.10.20 um 03:19 schrieb Jens Axboe:
Quoted 6 lines
Quoted 3 lines
So, instead could you please apply the patches for 5.11 as soon as it is suitable?
I will - I have it queued up, won't create anything public until we get past the merge window.
Sorry to bother you again with this.
Is there any outlook when you are going to push the patches to your for-next branch?
Or shall I resend the patches?
Jens Axboe
maintainerQuoted 8 lines · Stefan Haberland
Am 14.10.20 um 03:19 schrieb Jens Axboe:
Quoted 6 lines
Quoted 3 lines
So, instead could you please apply the patches for 5.11 as soon as it is suitable?
I will - I have it queued up, won't create anything public until we get past the merge window.
Quoted 5 lines
Sorry to bother you again with this.
Is there any outlook when you are going to push the patches to your for-next branch?Or shall I resend the patches?
I already did apply it, just don't push out next release branches that early. Pushed out now, so it is in.
Signature · 1 line
Jens Axboe
Stefan Haberland
maintainerAm 11.11.20 um 17:20 schrieb Jens Axboe:
Quoted 7 lines · Stefan Haberland
Quoted 6 lines
Am 14.10.20 um 03:19 schrieb Jens Axboe:
Quoted 5 lines
Quoted 2 lines
So, instead could you please apply the patches for 5.11 as soon as it is suitable?
I will - I have it queued up, won't create anything public until we get past the merge window.
Quoted 8 lines
Quoted 5 lines
Sorry to bother you again with this.
Is there any outlook when you are going to push the patches to your for-next branch?Or shall I resend the patches?
I already did apply it, just don't push out next release branches that early. Pushed out now, so it is in.
Thanks a lot.