- 27 Mar, 2020 15 commits
-
-
Nitin Rawat authored
As a part of sysfs reading of descriptors/attributes/flags, query commands should only be executed when hba's power runtime status is active. To guarantee this, add pm_runtime_get/put_sync() to those paths where query commands are sent. Link: https://lore.kernel.org/r/f712a4f7bdb0ae32e0d83634731e7aaa1b3a6cdd.1585009663.git.asutoshd@codeaurora.orgReviewed-by:
Avri Altman <avri.altman@wdc.com> Signed-off-by:
Nitin Rawat <nitirawa@codeaurora.org> Signed-off-by:
Asutosh Das <asutoshd@codeaurora.org> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
Stanley Chu authored
MediaTek platform and UFS controller can dynamically customize the delay for host enabling according to different scenarios. For example, if UniPro enters lower-power mode, such delay can be minimized, otherwise longer delay shall be expected. Link: https://lore.kernel.org/r/20200318104016.28049-8-stanley.chu@mediatek.comReviewed-by:
Stanley Chu <stanley.chu@mediatek.com> Signed-off-by:
-
Stanley Chu authored
Reduce the waiting period between each HCE (Host Controller Enable) polling from 5 ms to 1 ms. Also increase the maximum polling times to make "total polling time" roughly the same. This change could make HCE initialization faster to improve latency of ufshcd initialization, error recovery, and resume behaviors. Link: https://lore.kernel.org/r/20200318104016.28049-7-stanley.chu@mediatek.comReviewed-by:
Can Guo <cang@codeaurora.org> Signed-off-by:
-
Stanley Chu authored
Currently a 1 ms delay is applied before polling CONTROLLER_ENABLE bit. This delay may not be required or can be changed in different controllers. Make the delay as a changeable value in struct ufs_hba to allow it customized by vendors. Link: https://lore.kernel.org/r/20200318104016.28049-6-stanley.chu@mediatek.comReviewed-by:
-
Stanley Chu authored
A common delay function is introduced in UFS core driver, thus ufs-mediatek can use it instead of the private delay function. Link: https://lore.kernel.org/r/20200318104016.28049-5-stanley.chu@mediatek.comReviewed-by:
-
Stanley Chu authored
Introduce a common delay function to provide flexible way for users to take choices of udelay and usleep_range into consideration according to the required delay time. Link: https://lore.kernel.org/r/20200318104016.28049-4-stanley.chu@mediatek.comReviewed-by:
-
Stanley Chu authored
Use an enum to specify the host capabilities instead of #defines inside the structure definition. Link: https://lore.kernel.org/r/20200318104016.28049-3-stanley.chu@mediatek.comReviewed-by:
Bean Huo <beanhuo@micron.com> Reviewed-by:
-
Stanley Chu authored
In ufshcd_disable_tx_lcc(), if ufshcd_dme_get() or ufshcd_dme_peer_get() get fail, uninitialized variable "tx_lanes" may be used as unexpected lane ID for DME configuration. Fix this issue by initializing "tx_lanes". Link: https://lore.kernel.org/r/20200318104016.28049-2-stanley.chu@mediatek.comReviewed-by:
-
Gabriel Krisman Bertazi authored
If an iSCSI connection happens to fail while the daemon isn't running (due to a crash or for another reason), the kernel failure report is not received. When the daemon restarts, there is insufficient kernel state in sysfs for it to know that this happened. open-iscsi tries to reopen every connection, but on different initiators, we'd like to know which connections have failed. There is session->state, but that has a different lifetime than an iSCSI connection, so it doesn't directly reflect the connection state. [mkp: typos] Link: https://lore.kernel.org/r/20200317233422.532961-1-krisman@collabora.com Cc: Khazhismel Kumykov <khazhy@google.com> Suggested-by:
Junho Ryu <jayr@google.com> Reviewed-by:
Lee Duncan <lduncan@suse.com> Signed-off-by:
Gabriel Krisman Bertazi <krisman@collabora.com> Signed-off-by:
-
Viacheslav Dubeyko authored
Trace events target_sequencer_start and target_cmd_complete (include/trace/events/target.h) are ready to show NAA identifier, LUN ID, and many other important command details in the system log: TP_printk("%s -> LUN %03u %s data_length %6u CDB %s (TA:%s C:%02x)", However, it's still hard to identify command on the initiator and command on the target in the real life output of system log. For that purpose SCSI provides a command identifier or task tag (term used in previous standards). This patch adds tag ID in the system log's output: TP_printk("%s -> LUN %03u tag %#llx %s data_length %6u CDB %s (TA:%s C:%02x)", kworker/1:1-35 [001] .... 1392.989452: target_sequencer_start: naa.5001405ec1ba6364 -> LUN 001 tag 0x1 SERVICE_ACTION_IN_16 data_length 32 CDB 9e 10 00 00 00 00 00 00 00 00 00 00 00 20 00 00 (TA:SIMPLE C:00) kworker/1:1-35 [001] .... 1392.989456: target_cmd_complete: naa.5001405ec1ba6364 <- LUN 001 tag 0x1 status GOOD (sense len 0) SERVICE_ACTION_IN_16 data_length 32 CDB 9e 10 00 00 00 00 00 00 00 00 00 00 00 20 00 00 (TA:SIMPLE C:00) Link: https://lore.kernel.org/r/226e01deaa9baf46d6ff3b8698bc9fe881f7dfc1.camel@dubeyko.comReviewed-by:Roman Bolshakov <r.bolshakov@yadro.com> Reviewed-by:
Konstantin Shelekhin <k.shelekhin@yadro.com> Reviewed-by:
Bart van Assche <bvanassche@acm.org> Signed-off-by:
Viacheslav Dubeyko <v.dubeiko@yadro.com> Signed-off-by:
-
Maurizio Lombardi authored
iscsit_close_session() can only be called when nconn is zero (otherwise a kernel panic is triggered). If nconn is zero then iscsit_stop_session() does nothing and exits, so calling it makes no sense. We still need to call iscsit_check_session_usage_count() because this function will sleep if the session's refcount is not zero and we don't want to destroy the session structure if it's still being referenced. Link: https://lore.kernel.org/r/20200313170656.9716-4-mlombard@redhat.comTested-by:
Rahul Kundu <rahul.kundu@chelsio.com> Signed-off-by:
Maurizio Lombardi <mlombard@redhat.com> Signed-off-by:
-
Maurizio Lombardi authored
A number of hangs have been reported against the target driver; they are due to the fact that multiple threads may try to destroy the iscsi session at the same time. This may be reproduced for example when a "targetcli iscsi/iqn.../tpg1 disable" command is executed while a logout operation is underway. When this happens, two or more threads may end up sleeping and waiting for iscsit_close_connection() to execute "complete(session_wait_comp)". Only one of the threads will wake up and proceed to destroy the session structure, the remaining threads will hang forever. Note that if the blocked threads are somehow forced to wake up with complete_all(), they will try to free the same iscsi session structure destroyed by the first thread, causing double frees, memory corruptions etc... With this patch, the threads that want to destroy the iscsi session will increase the session refcount and will set the "session_close" flag to 1; then they wait for the driver to close the remaining active connections. When the last connection is closed, iscsit_close_connection() will wake up all the threads and will wait for the session's refcount to reach zero; when this happens, iscsit_close_connection() will destroy the session structure because no one is referencing it anymore. INFO: task targetcli:5971 blocked for more than 120 seconds. Tainted: P OE 4.15.0-72-generic #81~16.04.1 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. targetcli D 0 5971 1 0x00000080 Call Trace: __schedule+0x3d6/0x8b0 ? vprintk_func+0x44/0xe0 schedule+0x36/0x80 schedule_timeout+0x1db/0x370 ? __dynamic_pr_debug+0x8a/0xb0 wait_for_completion+0xb4/0x140 ? wake_up_q+0x70/0x70 iscsit_free_session+0x13d/0x1a0 [iscsi_target_mod] iscsit_release_sessions_for_tpg+0x16b/0x1e0 [iscsi_target_mod] iscsit_tpg_disable_portal_group+0xca/0x1c0 [iscsi_target_mod] lio_target_tpg_enable_store+0x66/0xe0 [iscsi_target_mod] configfs_write_file+0xb9/0x120 __vfs_write+0x1b/0x40 vfs_write+0xb8/0x1b0 SyS_write+0x5c/0xe0 do_syscall_64+0x73/0x130 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 Link: https://lore.kernel.org/r/20200313170656.9716-3-mlombard@redhat.comReported-by:Matt Coleman <mcoleman@datto.com> Tested-by:
-
Maurizio Lombardi authored
iscsit_free_session() is equivalent to iscsit_stop_session() followed by a call to iscsit_close_session(). Link: https://lore.kernel.org/r/20200313170656.9716-2-mlombard@redhat.comTested-by:
-
Christophe JAILLET authored
If 'dma_map_single()' fails, the ref counted 'shpnt' will be decremented twice because 'scsi_host_put()' is called in the if block, and in the error handling path. Axe one of these calls. Link: https://lore.kernel.org/r/20200228215948.7473-1-christophe.jaillet@wanadoo.fr Fixes: 1dc09e12 ("scsi: aha1740: stop using scsi_unregister") Signed-off-by:
Christophe JAILLET <christophe.jaillet@wanadoo.fr> Signed-off-by:
-
Daniel Wagner authored
Remove code which has no functional use anymore since commit 3c75ad1d ("scsi: qla2xxx: Remove defer flag to indicate immeadiate port loss"). While at it remove also the stale function documentation. Link: https://lore.kernel.org/r/20200206135443.110701-1-dwagner@suse.deReviewed-by:
Arun Easi <aeasi@marvell.com> Reviewed-by:
Daniel Wagner <dwagner@suse.de> Signed-off-by:
-
- 17 Mar, 2020 25 commits
-
-
Viswas G authored
Removed the common length and introduce read and write length for IOCTL payload structure. [mkp: fixed SoB ordering] Link: https://lore.kernel.org/r/20200316074906.9119-7-deepak.ukey@microchip.comAcked-by:
Jack Wang <jinpu.wang@cloud.ionos.com> Signed-off-by:
Viswas G <viswas.g@microchip.com> Signed-off-by:
Deepak Ukey <deepak.ukey@microchip.com> Signed-off-by:
Radha Ramachandran <radha@google.com> Signed-off-by:
-
Deepak Ukey authored
Added the sysfs attribute for non fatal log so that management utility can get the non fatal dump from driver. The non-fatal error is an error condition or abnormal behavior detected by the host, or detected and reported by the controller to the host.The non-fatal error does not stop the controller firmware and enables it to still respond to host requests. A typical example of a non-fatal error is an I/O timeout or an unusual error notification from the controller. Since the firmware is operational, the error dump information is pushed to host memory (by firmware) upon request from the host. Link: https://lore.kernel.org/r/20200316074906.9119-6-deepak.ukey@microchip.comAcked-by:
-
Peter Chang authored
1) Move the instance tracking down after we think the instance is good to go. Avoids having a use-after free. 2) There are goto targets for trying to cleanup if the hw fails to initialize, but there's some overlap depending on who thinks they own the sub-structures. Link: https://lore.kernel.org/r/20200316074906.9119-5-deepak.ukey@microchip.comAcked-by:
Peter Chang <dpf@google.com> Signed-off-by:
-
yuuzheng authored
In pm80xx driver, the command mpi_set_phy_profile_req is sent by host during boot to configure the phy profile such as analog setting page, rate control page. However, the tag is not freed when its response is received. As a result, 16 tags are missing for each HBA after boot. When NCQ is enabled with queue depth 16, it needs at least, 15 * 16 = 240 tags for each HBA to achieve the best performance. In current pm80xx driver with setting CCB_MAX = 256, the total number of tags in each HBA is 255 for data IO. Hence, without returning those tags to the pool after boot, some device will finally be forced to non-ncq mode by ATA layer due to excessive errors (i.e. LLDD cannot allocate tag for queued task). Link: https://lore.kernel.org/r/20200316074906.9119-4-deepak.ukey@microchip.comAcked-by:
yuuzheng <yuuzheng@google.com> Signed-off-by:
-
Vikram Auradkar authored
A kexec reboot causes the controller fw to assert. This assertion shows up in two ways, the controller doesn't show up as ready and an interrupt is waiting as soon as the handler is registered. To resolve this added below fix: - Split the interrupt handling setup into two parts, setup and request. - If the controller ready register indicates not-ready, but that the not readiness is only on the IOC units we can still try a reset to bring the system back to the pre-reboot state. Link: https://lore.kernel.org/r/20200316074906.9119-3-deepak.ukey@microchip.comAcked-by:
Vikram Auradkar <auradkar@google.com> Signed-off-by:
-
Peter Chang authored
Increasing the per-request size maximum (max_sectors_kb) runs into the per-device DMA scatter gather list limit (max_segments) for users of the io vector system calls (eg, readv and writev). This is because the kernel combines io vectors into DMA segments when possible, but it doesn't work for our user because the vectors in the buffer cache get scrambled. This change bumps the advertised max scatter gather length to 528 to cover 2M w/ x86's 4k pages and some extra for the user checksum. It trims the size of some of the tables we don't care about and exposes all of the command slots upstream to the SCSI layer. Also reduced the PM8001_MAX_CCB to 256 as pm8001 driver has memory limit depend on machine capability. If we increase the sg length, we need to trade-off it by decreasing PM8001_MAX_CCB. PM8001_MAX_CCB = 256 does not have any influence on normal use Link: https://lore.kernel.org/r/20200316074906.9119-2-deepak.ukey@microchip.comReported-by:
kbuild test robot <lkp@intel.com> Acked-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Link: https://lore.kernel.org/r/20200315094241.9086-9-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: Don Brace <don.brace@microsemi.com> Cc: linux-scsi@vger.kernel.org Signed-off-by:
Takashi Iwai <tiwai@suse.de> Signed-off-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Link: https://lore.kernel.org/r/20200315094241.9086-8-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: linux-scsi@vger.kernel.org Reviewed-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Also corrected the wrongly passed limit size. The remaining buffer size must be decremented. Link: https://lore.kernel.org/r/20200315094241.9086-7-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: Kashyap Desai <kashyap.desai@broadcom.com> Cc: Sumit Saxena <sumit.saxena@broadcom.com> Cc: Shivasharan S <shivasharan.srikanteshwara@broadcom.com> Cc: linux-scsi@vger.kernel.org Signed-off-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Link: https://lore.kernel.org/r/20200315094241.9086-6-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: Brian King <brking@us.ibm.com> Cc: linux-scsi@vger.kernel.org Signed-off-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). [mkp: checkpatch fix] Link: https://lore.kernel.org/r/20200315094241.9086-5-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: Achim Leubner <achim_leubner@adaptec.com> Cc: linux-scsi@vger.kernel.org Signed-off-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Link: https://lore.kernel.org/r/20200315094241.9086-4-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: Satish Kharat <satishkh@cisco.com> Cc: Sesidhar Baddela <sebaddel@cisco.com> Cc: Karan Tilak Kumar <kartilak@cisco.com> Cc: linux-scsi@vger.kernel.org Signed-off-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Link: https://lore.kernel.org/r/20200315094241.9086-3-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: Subbu Seetharaman <subbu.seetharaman@broadcom.com> Cc: Ketan Mukadam <ketan.mukadam@broadcom.com> Cc: Jitendra Bhivare <jitendra.bhivare@broadcom.com> Cc: linux-scsi@vger.kernel.org Signed-off-by:
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Link: https://lore.kernel.org/r/20200315094241.9086-2-tiwai@suse.de Cc: "James E . J . Bottomley" <jejb@linux.ibm.com> Cc: "Martin K . Petersen" <martin.petersen@oracle.com> Cc: Adaptec OEM Raid Solutions <aacraid@microsemi.com> Cc: linux-scsi@vger.kernel.org Acked-by:
Balsundar P <Balsundar.P@microchip.com> Signed-off-by:
-
Jens Remus authored
Log any FC Endpoint Security errors to the kernel ring buffer with rate- limiting. Link: https://lore.kernel.org/r/20200312174505.51294-11-maier@linux.ibm.comReviewed-by:
Steffen Maier <maier@linux.ibm.com> Signed-off-by:
Jens Remus <jremus@linux.ibm.com> Signed-off-by:
-
Jens Remus authored
Enable for explicit FCP channel FC Endpoint Security error reporting and handle any FSF security errors according to specification. Take the following recovery actions when a FSF_SECURITY_ERROR is reported for the specified FSF commands: - Open Port: Retry the command if possible - Send FCP : Physically close the remote port and reopen For Open Port the command status is set to error, which triggers a retry. For Send FCP the command status is set to error and recovery is triggered to physically reopen the remote port. Link: https://lore.kernel.org/r/20200312174505.51294-10-maier@linux.ibm.comReviewed-by:
-
Jens Remus authored
Trace changes in Fibre Channel Endpoint Security capabilities of FCP devices as well as changes in Fibre Channel Endpoint Security state of their connections to FC remote ports as FC Endpoint Security changes with trace level 3 in HBA DBF. A change in FC Endpoint Security capabilities of FCP devices is traced as response to FSF command FSF_QTCB_EXCHANGE_PORT_DATA with a trace tag of "fsfcesa" and a WWPN of ZFCP_DBF_INVALID_WWPN = 0x0000000000000000 (see FC-FS-4 §18 "Name_Identifier Formats", NAA field). A change in FC Endpoint Security state of connections between FCP devices and FC remote ports is traced as response to FSF command FSF_QTCB_OPEN_PORT_WITH_DID with a trace tag of "fsfcesp". Example trace record of FC Endpoint Security capability change of FCP device formatted with zfcpdbf from s390-tools: Timestamp : ... Area : HBA Subarea : 00 Level : 3 Exception : - CPU ID : ... Caller : 0x... Record ID : 5 ZFCP_DBF_HBA_FCES Tag : fsfcesa FSF FC Endpoint Security adapter Request ID : 0x... Request status : 0x00000010 FSF cmnd : 0x0000000e FSF_QTCB_EXCHANGE_PORT_DATA FSF sequence no: 0x... FSF issued : ... FSF stat : 0x00000000 FSF_GOOD FSF stat qual : n/a Prot stat : n/a Prot stat qual : n/a Port handle : 0x00000000 none (invalid) LUN handle : n/a WWPN : 0x0000000000000000 ZFCP_DBF_INVALID_WWPN FCES old : 0x00000000 old FC Endpoint Security FCES new : 0x00000007 new FC Endpoint Security Example trace record of FC Endpoint Security change of connection to FC remote port formatted with zfcpdbf from s390-tools: Timestamp : ... Area : HBA Subarea : 00 Level : 3 Exception : - CPU ID : ... Caller : 0x... Record ID : 5 ZFCP_DBF_HBA_FCES Tag : fsfcesp FSF FC Endpoint Security port Request ID : 0x... Request status : 0x00000010 FSF cmnd : 0x00000005 FSF_QTCB_OPEN_PORT_WITH_DID FSF sequence no: 0x... FSF issued : ... FSF stat : 0x00000000 FSF_GOOD FSF stat qual : n/a Prot stat : n/a Prot stat qual : n/a Port handle : 0x... WWPN : 0x500507630401120c WWPN FCES old : 0x00000000 old FC Endpoint Security FCES new : 0x00000004 new FC Endpoint Security Link: https://lore.kernel.org/r/20200312174505.51294-9-maier@linux.ibm.comReviewed-by:
-
Jens Remus authored
Log the usage of and subsequent changes in FC Endpoint Security of connections between FCP devices and FC remote ports to the kernel ring buffer. Activation of FC Endpoint Security is logged as informational. Change and deactivation are logged as warning. No logging takes place, if FC Endpoint Security is not used (i.e. never activated) on a connection or if it does not change during reopen of a port (e.g. due to adapter or port recovery). Link: https://lore.kernel.org/r/20200312174505.51294-8-maier@linux.ibm.comReviewed-by:
Fedor Loshakov <loshakov@linux.ibm.com> Signed-off-by:
-
Jens Remus authored
Add an interface to read Fibre Channel Endpoint Security information of FCP channels and their connections to FC remote ports. It comes in the form of new sysfs attributes that are attached to the CCW device representing the FCP device and its zfcp port objects. The read-only sysfs attribute "fc_security" of a CCW device representing a FCP device shows the FC Endpoint Security capabilities of the device. Possible values are: "unknown", "unsupported", "none", or a comma- separated list of one or more mnemonics and/or one hexadecimal value representing the supported FC Endpoint Security: Authentication: Authentication supported Encryption : Encryption supported The read-only sysfs attribute "fc_security" of a zfcp port object shows the FC Endpoint Security used on the connection between its parent FCP device and the FC remote port. Possible values are: "unknown", "unsupported", "none", or a mnemonic or hexadecimal value representing the FC Endpoint Security used: Authentication: Connection has been authenticated Encryption : Connection is encrypted Both sysfs attributes may return hexadecimal values instead of mnemonics, if the mnemonic lookup table does not contain an entry for the FC Endpoint Security reported by the FCP device. Link: https://lore.kernel.org/r/20200312174505.51294-7-maier@linux.ibm.comReviewed-by:
Benjamin Block <bblock@linux.ibm.com> Signed-off-by:
-
Jens Remus authored
Introduce automatic variables for adapter and QTCB bottom in zfcp_fsf_open_port_handler(). This facilitates subsequent changes to meet the 80 character per line limit. Link: https://lore.kernel.org/r/20200312174505.51294-6-maier@linux.ibm.comReviewed-by:
-
Steffen Maier authored
When we get an unsolicited notification on local link went down, zfcp_fsf_status_read_link_down() calls zfcp_fsf_link_down_info_eval(). This only blocks rports, and sets ZFCP_STATUS_ADAPTER_LINK_UNPLUGGED and ZFCP_STATUS_COMMON_ERP_FAILED. Only the fc_host port_state changes to "Linkdown", because zfcp_scsi_get_host_port_state() is an active callback and uses the adapter status. Other fc_host attributes model, port_id, port_type, speed, fabric_name (and zfcp device attributes card_version, peer_wwpn, peer_wwnn, peer_d_id) which depend on a local link, continued to show their last known "good" value. Only if something triggered an exchange config data, some values were updated to their unknown equivalent via case FSF_EXCHANGE_CONFIG_DATA_INCOMPLETE due to local link down. Triggers for exchange config data are adapter recovery, or reading any of the following zfcp-specific scsi host sysfs attributes "requests", "megabytes", or "seconds_active" in /sys/devices/css*/*.*.*/*.*.*/host*/scsi_host/host*/. The other fc_host attributes active_fc4s and permanent_port_name continued to show their last known "good" value. Only if something triggered an exchange port data, some values changed. Active_fc4s became all zeros as unknown equivalent during link down. Permanent_port_name does not depend on a local link. But for non-NPIV FCP devices, permanent_port_name erroneously became whatever value fc_host port_name had at that point in time (see previous paragraph). Triggers for exchange port data are the zfcp-specific scsi host sysfs attribute "utilization", or [{reset,get}_fc_host_stats] write anything into "reset_statistics" or read any of the other attributes under /sys/devices/css*/*.*.*/*.*.*/host*/fc_host/host*/statistics/. (cf. v4.9 commit bd77befa ("zfcp: fix fc_host port_type with NPIV")) This is particularly confusing when using "lszfcp -b <fcpdevbusid> -Ha" or dbginfo.sh which read fc_host attributes and also scsi_host attributes. After link down, the first invocation produces (abbreviated): Class = "fc_host" active_fc4s = "0x00 0x00 0x01 0x00 ..." ... fabric_name = "0x10000027f8e04c49" ... permanent_port_name = "0xc05076e4588059c1" port_id = "0x244800" port_state = "Linkdown" port_type = "NPort (fabric via point-to-point)" ... speed = "16 Gbit" Class = "scsi_host" ... megabytes = "0 0" ... requests = "0 0 0" seconds_active = "37" ... utilization = "0 0 0" The second and next invocations produce (abbreviated): Class = "fc_host" active_fc4s = "0x00 0x00 0x00 0x00 ..." ... fabric_name = "0x0" ... permanent_port_name = "0x0" port_id = "0x000000" port_state = "Linkdown" port_type = "Unknown" ... speed = "unknown" Class = "scsi_host" ... megabytes = "0 0" ... requests = "0 0 0" seconds_active = "38" ... utilization = "0 0 0" Factor out the resetting of local link dependent fc_host attributes from zfcp_fsf_exchange_config_data_handler() case FSF_EXCHANGE_CONFIG_DATA_INCOMPLETE into a new helper function zfcp_fsf_fc_host_link_down(). All code places that detect local link down (SRB, FSF_PROT_LINK_DOWN, xconf data/port incomplete) call zfcp_fsf_link_down_info_eval(). Call the new helper from there. This works because zfcp_fsf_link_down_info_eval() and thus the helper is called before zfcp_fsf_exchange_{config,port}_evaluate(). Port_name and node_name are always valid, so never reset them. Get the permanent_port_name from exchange port data unconditionally as it always has a valid known good value, even during link down. Note: Rather than hardcode in zfcp_fsf_exchange_config_evaluate(), fc_host supported_classes could theoretically get its value from fsf_qtcb_bottom_port.class_of_service in zfcp_fsf_exchange_port_evaluate(). When the link comes back, we get a different notification, perform adapter recovery, and this triggers an implicit exchange config data followed by exchange port data filling in the link dependent fc_host attributes with known good values again. Link: https://lore.kernel.org/r/20200312174505.51294-5-maier@linux.ibm.comReviewed-by: -
Steffen Maier authored
Manufacturer, HBA model, firmware version, and hardware version. Use the same value format as for the driver-specific attributes. Keep the driver-specific attributes for stable user space sysfs API. Link: https://lore.kernel.org/r/20200312174505.51294-4-maier@linux.ibm.comReviewed-by:
-
Steffen Maier authored
FICON Express8S or older, as well as card features newer than FICON Express16S+ have no certain firmware level requirement. FICON Express16S or FICON Express16S+ have the following minimum firmware level requirements to show a proper fabric name value: z13 machine FICON Express16S , MCL P08424.005 , LIC version 0x00000721 z14 machine FICON Express16S , MCL P42611.008 , LIC version 0x10200069 FICON Express16S+ , MCL P42625.010 , LIC version 0x10300147 Otherwise, the read value is not the fabric name. Each FCP channel of these card features might need one SAN fabric re-login after concurrent microcode update in order to show the proper fabric name. Possible ways to trigger a SAN fabric re-login are one of: Pull fibres between FCP channel port and SAN switch port on either side and re-plug, disable SAN switch port adjacent to FCP channel port and re-enable switch port, or at Service Element toggle off all CHPIDs of FCP channel over all LPARs and toggle CHPIDs on again. Zfcp operating subchannels (FCP devices) on such FCP channel recovers a fabric re-login. Initialize fabric name for any topology and have it an invalid WWPN 0x0 for anything but fabric topology. Otherwise for e.g. point-to-point topology one could see the initial -1 from fc_host_setup() and after a link unplug our fabric name would turn to 0x0 (with subsequent commit ("zfcp: fix fc_host attributes that should be unknown on local link down") and stay 0x0 on link replug. I did not initialize to 0x0 somewhere even earlier in the code path such that it would not flap from real to 0x0 to real on e.g. an exchange config data with fabric topology. Link: https://lore.kernel.org/r/20200312174505.51294-3-maier@linux.ibm.comReviewed-by: -
Steffen Maier authored
v2.6.27 commit cc8c2829 ("[SCSI] zfcp: Automatically attach remote ports") introduced zfcp automatic port scan. Before that, the user had to use the sysfs attribute "port_add" of an FCP device (adapter) to add and open remote (target) ports, even for the remote peer port in point-to-point topology. That code path did a proper port open recovery trigger taking the erp_lock. Since above commit, a new helper function zfcp_erp_open_ptp_port() performed an UNlocked port open recovery trigger. This can race with other parallel recovery triggers. In zfcp_erp_action_enqueue() this could corrupt e.g. adapter->erp_total_count or adapter->erp_ready_head. As already found for fabric topology in v4.17 commit fa89adba ("scsi: zfcp: fix infinite iteration on ERP ready list"), there was an endless loop during tracing of rport (un)block. A subsequent v4.18 commit 9e156c54 ("scsi: zfcp: assert that the ERP lock is held when tracing a recovery trigger") introduced a lockdep assertion for that case. As a side effect, that lockdep assertion now uncovered the unlocked code path for PtP. It is from within an adapter ERP action: zfcp_erp_strategy[1479] intentionally DROPs erp lock around zfcp_erp_strategy_do_action() zfcp_erp_strategy_do_action[1441] NO erp lock zfcp_erp_adapter_strategy[876] NO erp lock zfcp_erp_adapter_strategy_open[855] NO erp lock zfcp_erp_adapter_strategy_open_fsf[806]NO erp lock zfcp_erp_adapter_strat_fsf_xconf[772] erp lock only around zfcp_erp_action_to_running(), BUT *_not_* around zfcp_erp_enqueue_ptp_port() zfcp_erp_enqueue_ptp_port[728] BUG: *_not_* taking erp lock _zfcp_erp_port_reopen[432] assumes to be called with erp lock zfcp_erp_action_enqueue[314] assumes to be called with erp lock zfcp_dbf_rec_trig[288] _checks_ to be called with erp lock: lockdep_assert_held(&adapter->erp_lock); It causes the following lockdep warning: WARNING: CPU: 2 PID: 775 at drivers/s390/scsi/zfcp_dbf.c:288 zfcp_dbf_rec_trig+0x16a/0x188 no locks held by zfcperp0.0.17c0/775. Fix this by using the proper locked recovery trigger helper function. Link: https://lore.kernel.org/r/20200312174505.51294-2-maier@linux.ibm.com Fixes: cc8c2829 ("[SCSI] zfcp: Automatically attach remote ports") Cc: <stable@vger.kernel.org> #v2.6.27+ Reviewed-by:
-
Bart Van Assche authored
This makes the SCSI tracing code slightly easier to read. Link: https://lore.kernel.org/r/20200313203102.16613-6-bvanassche@acm.org Fixes: bf816235 ("[SCSI] add scsi trace core functions and put trace points") Cc: Christoph Hellwig <hch@lst.de> Cc: James E.J. Bottomley <jejb@linux.ibm.com> Cc: Martin K. Petersen <martin.petersen@oracle.com> Reported-by:
Colin Ian King <colin.king@canonical.com> Reviewed-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
-
