- 13 Feb, 2011 4 commits
-
-
James Bottomley authored
Right at the moment, the libata error handler is incredibly monolithic. This makes it impossible to use from composite drivers like libsas and ipr which have to handle error themselves in the first instance. The essence of the change is to split the monolithic error handler into two components: one which handles a queue of ata commands for processing and the other which handles the back end of readying a port. This allows the upper error handler fine grained control in calling libsas functions (and making sure they only get called for ATA commands whose lower errors have been fixed up). Cc: Tejun Heo <tj@kernel.org> Cc: Jeff Garzik <jeff@garzik.org> Signed-off-by:James Bottomley <James.Bottomley@suse.de>
-
James Bottomley authored
The SCSI host eh_cmd_q should be protected by the host lock (not the port lock). This probably doesn't matter that much at the moment, since we try to serialise the add and eh pieces, but it might matter in future for more convenient error handling. Plus this switches libata to the standard eh pattern where you lock, remove from the cmd queue to a local list and unlock and then operate on the local list. Cc: Tejun Heo <tj@kernel.org> Cc: Jeff Garzik <jeff@garzik.org> Signed-off-by: -
James Bottomley authored
For historical reasons, libsas uses the scsi host lock as the ata port lock, and libata always uses the ata host. For the old eh, this was largely irrelevant since the two locks were never mixed inside the code. However, the new eh has a case where it nests acquisition of the host lock inside the port lock (this does look rather deadlock prone). Obviously this would be an instant deadlock if the port lock were the host lock, so switch the libsas paths to use the ata host lock as well. Cc: Tejun Heo <tj@kernel.org> Cc: Jeff Garzik <jeff@garzik.org> Signed-off-by: -
James Bottomley authored
The function ata_sas_port_init() has always really done its own thing. However, as a precursor to moving to the libata new eh, it has to be properly using the standard libata scan paths. This means separating the current libata scan paths into pieces which can be shared with libsas and pieces which cant (really just the async call and the host scan). Cc: Tejun Heo <tj@kernel.org> Cc: Jeff Garzik <jeff@garzik.org> Signed-off-by:
-
- 12 Feb, 2011 36 commits
-
-
Moger, Babu authored
Adding MODULE_VERSION for scsi_dh_rdac. This will be helpful sometimes to get the code level without looking at the code. Signed-off-by:
Babu Moger <babu.moger@lsi.com> Signed-off-by:
-
Hillf Danton authored
instead of doing sizeof(struct X) it's better to do sizeof(*v) where v is the variable pointing to struct X. Signed-off-by:
Hillf Danton <dhillf@gmail.com> Signed-off-by:
-
Moger, Babu authored
During one of our testing, we noticed that mode select command sent from the host did not have the lun_table updated. Problem is root caused to the way lun table is updated. Lun table update was done after the call to blk_rq_map_kern is made. This was causing problem because kernel uses bounce buffer(bio_copy_kern) if the address is not aligned. The command buffer updated after the call(blk_rq_map_kern) was not going on the wire. Moved the code to update the lun_table before the call to fix the problem. Signed-off-by:
Somasundaram Krishnasamy <Somasundaram.Krishnasamy@lsi.com> Signed-off-by:
Yanling Qi <Yanling.Qi@lsi.com> Signed-off-by:
-
Kashyap, Desai authored
Bus reset is not required for SAS Controller. It is valid for mptspi and mptfc, but for mptsas it is not required. It is an extra work for Error handling escallation for mptsas. Removing bus reset from error handling will eventually speedup Error handling for SAS controller. Signed-off-by:
Kashyap Desai <kashyap.desai@lsi.com> Signed-off-by:
-
Kashyap, Desai authored
SAS1.0 Controller was not able to detect SAS2.0 Expanders due to Link RATE detection was limited to 1.5 Gbps and 3.0 Gbps for SAS1 controllers. Added detection for 6.0 Gbps link. Now, user can mix-up 6.0 Gpbs links with SAS1.0 controller. e.g SAS1.0 HBA <----> SAS2.0 Expander <------> SAS2.0 Expander <--------> SAS1.0 Drive. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
To facilitate LLDDs to reuse the code, skb queue related functions are moved to libfcoe, so that both fcoe and bnx2fc drivers can use them. The common structures fcoe_port, fcoe_percpu_s are moved to libfcoe. fcoe_port will now have an opaque pointer that points to corresponding driver's interface structure. Also, fcoe_start_io and fcoe_fc_crc are moved to libfcoe. As part of this change, fixed fcoe_start_io to return ENOMEM if skb_clone fails. Signed-off-by:
Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by:
Robert Love <robert.w.love@intel.com> Signed-off-by:
-
Tejun Heo authored
fcoe uses the system_wq to destroy ports and the work items need to be flushed before the driver is unloaded. As the work items free the containing data structure, they can't be flushed directly. The workqueue should be flushed instead. Also, the destruction works can be chained - ie. destruction of a port may lead to destruction of another port where the work item for the former queues the work for the latter. Currently, the depth of chain can be at most two and fcoe_exit() makes sure everything is complete by calling flush_scheduled_work() twice. With commit c8efcc25 (workqueue: allow chained queueing during destruction), destroy_workqueue() can take care of chained works on workqueue destruction. Add and use fcoe_wq instead. Simply destroying fcoe_wq on driver unload takes care of flushing. Signed-off-by:
Tejun Heo <tj@kernel.org> Signed-off-by:
-
Bhanu Prakash Gollapudi authored
This patch enables LLD to listen to rport events and perform LLD specific operations based on the rport event. This patch also stores sp_features and spp_type in rdata for further reference by LLD. Signed-off-by:
-
Kiran Patil authored
Problem: From initaitor machine, when queried role of target (other end of connection), it is "initiator", hence SCSI-ml doesn't send any LUN Inquiry commands. Fix: If there is a registered target for FC_TYPE_FCP, extend lport's params (capability) to be target as well, By default lport params are INITIATOR only. Having this fix, caused initiator to send SCSI LUN inquiry command to target. Signed-off-by:
Kiran Patil <kiran.patil@intel.com> Signed-off-by:
-
Yi Zou authored
Remove the existing sysfs entry points of the fcoe.ko module parameters that are used to create/destroy/enable/disable an FCoE instance, rather, use the newly added fcoe transport code to attach itself as an FCoE transport provider when fcoe.ko gets loaded. There is no functionality change on the logic of fcoe interacts with upper libfc and lower netdev. The fcoe transport only acts as thin layer to provide a unified interface for all fcoe transport providers so all FCoE instances on any network interfaces from all vendors can be managed through the same Open-FCoE.org's user space tool package, which also has full DCB support. Signed-off-by:
Yi Zou <yi.zou@intel.com> Signed-off-by:
-
Yi Zou authored
Prepare the fcoe to convert it to use the newly added fcoe transport, making it as the default fcoe transport provider for libfcoe. This patch is to rename some of the variables to avoid any confusing names later as now there are several transports in the same file. Signed-off-by:
-
Yi Zou authored
Now we can include the fcoe_transport.c to the build of the kernel libfcoe module. Move the module information to fcoe_transport, and it will have all the module parameters later for the create/destroy/enable/disable of an FCoE instance. Signed-off-by:
-
Yi Zou authored
The existing libfcoe.c is mostly for FIP support, rename it to reflect that fact and so we can add fcoe_transport.c to the make file to include both into the libfcoe kernel module. [ Minor modifications by Robert Love converting a few "__attribute__((packed))" modifiers to "__packed" to remove new checkpatch.pl WARNINGS ] Signed-off-by:
-
Yi Zou authored
Add the new fcoe_transport.c file that implements basic fcoe transport interface. Eventually, the sysfs entries to create/destroy/enable/disable an FCoE instance will be coming to the fcoe transport layer, who does a look-up to find the corresponding transport provide and pass the corresponding action over to the identified provider. The fcoe.ko will become the default fcoe transport provider that can support FCoE on any given netdev interfaces, as the Open-FCoE.org's default software FCoE HBA solution. Any vendor specific FCoE HBA driver that is built on top of Open-FCoE's kernel stack of libfc & libfcoe as well as the user land tool of fcoe-utils can easily plug-in and start running FCoE on their network interfaces. The fcoe.ko will be converted to act as the default provider if no vendor specific transport provider is found, as it is always added to the very end of the list of attached transports. Signed-off-by:
-
Yi Zou authored
add the fcoe_transport struct to the common libfcoe.h header so all fcoe transport provides can use it to attach itself as an fcoe transport. This is the header part, and the next patch will be the transport code itself. Signed-off-by:
-
Yi Zou authored
libfcoe kernel module debug macros will used by the fcoe transport code as well when later it gets added. Signed-off-by:
-
Kiran Patil authored
Problem: In case of exchange responder case, EMA selection was defaulted to the last EMA from EMA list (lport.ema_list). If exchange ID is selected from offload pool and not setup DDP, resulting into incorrect selection of EMA, and eventually dropping the packet because unable to find exchange. Fix: Enhanced the exchange ID selection (depending upon request type and exchange responder) Made necessary enhancement in EMA selection algorithm. Signed-off-by:
-
Joe Eykholt authored
Target modules using lport->tt.seq_assign() get a hold on the exchange but have no way of releasing it. Add that. Signed-off-by:
Joe Eykholt <jeykholt@cisco.com> Signed-off-by:
-
Robert Love authored
This patch removes the use of the Scsi_Host's host_lock within fc_queuecommand. It also removes the DEF_SCSI_QCMD usage so that libfc has fully moved on to the new queuecommand interface. Signed-off-by:
Ross Brattain <ross.b.brattain@intel.com> Reviewed-by:
Nicholas A. Bellinger <nab@linux-iscsi.org> Signed-off-by:
-
Joe Eykholt authored
When sending an outgoing PRLI as an initiator, get the parameters from registered providers so that they all get a chance to decide on roles. The passive provider is called last, and could override the initiator role. Signed-off-by:
-
Joe Eykholt authored
When an SCST provider is registered, it needs to know what local ports are available for configuration as targets. Add a notifier chain that is invoked when any local port that is added or deleted. Maintain a global list of local ports and add an interator function that calls a given function for every existing local port. This is used when first loading a provider. Signed-off-by:
-
Joe Eykholt authored
The target provider needs a per-instance lookup table or other way to lookup sessions quickly without going through a linear list or serializing too much. Add a simple void * array indexed by FC-4 type to the fc_lport. Signed-off-by:
-
Joe Eykholt authored
Add a method for setting handler for incoming exchange. For multi-sequence exchanges, this allows the target driver to add a response handler for handling subsequent sequences, and exchange manager resets. The new function is called fc_seq_set_resp(). Signed-off-by:
-
Joe Eykholt authored
Allow FC-4 provider modules to hook into libfc, mostly for targets. This should allow any FC-4 module to handle PRLI requests and maintain process-association states. Each provider registers its ops with libfc and then will be called for any incoming PRLI for that FC-4 type on any instance. The provider can decide whether to handle that particular instance using any method it likes, such as ACLs or other configuration information. A count is kept of the number of successful PRLIs from the remote port. Providers are called back with an implicit PRLO when the remote port is about to be deleted or has been reset. fc_lport_recv_req() now sends incoming FC-4 requests to FC-4 providers, and there is a built-in provider always registered for handling incoming ELS requests. The call to provider recv() routines uses rcu_read_lock() so that providers aren't removed during the call. That lock is very cheap and shouldn't affect any performance on ELS requests. Providers can rely on the RCU lock to protect a session lookup as well. Signed-off-by:
-
Randy Dunlap authored
Fix sparse warning for non-ANSI function declaration. Declare workqueue structs as static. Signed-off-by:
Randy Dunlap <randy.dunlap@oracle.com> Cc: Robert Love <robert.w.love@intel.com> Signed-off-by:
-
Vasu Dev authored
Allowing FCoE LOGO followed by CVL in this case prevents FIP login back to the FCF and then keeps lport offline, only FIP LOGO and CLV needs to be processed while in FIP mode, therefore this patch drops FCoE LOGO in FIP mode. Added fcoe_filter_frames() to filter out above mentioned LOGO in fcoe rx path along with other existing filtering in code for bad CRC frames. Adding separate fcoe_filter_frames function helped with better code indentations and if needed then same will allow adding more filters at one place in future. This LOGO drop is added after FCP frames passed up to avoid any additional checks on fast path for this. Signed-off-by:
Vasu Dev <vasu.dev@intel.com> Signed-off-by:
-
Robert Love authored
vports are not grabbing module references but are releasing them. This causes the module reference count to decrement too many times and it wraps around past 0. The solution is to do a module_put() in fcoe_interface_release() so that the reference is only released when the fcoe_interface is released. There is a one-to-one relationship between the N_Port and the fcoe_interface, so the module reference will only be dropped when the N_Port is destroyed To create symetry in the code this patch moves the try_module_get() call into fcoe_interface_create(). This means that only the N_Port will grab a reference to the module when its corresponding fcoe_interface is created. This patch also makes it so that the fcoe_interface_create() routine encodes any error codes in the fcoe_interface pointer returned. This way its caller, fcoe_create(), can return an accurate error code. Signed-off-by:
-
Venkata Siva Vijayendra Bhamidipati authored
Signed-off-by:
Venkata Siva Vijayendra Bhamidipati <vbhamidi@cisco.com> Signed-off-by:
-
Venkata Siva Vijayendra Bhamidipati authored
Fix memory leak arising due to incorrect freeing of allocated memory for vnic stats when unregistering a vnic. Signed-off-by:
Abhijeet Joglekar <abjoglek@cisco.com> Signed-off-by:
-
Dan Carpenter authored
If we goto out, then it tries to call kfree_skb() on an ERR_PTR which will oops. Just return directly. Signed-off-by:
Dan Carpenter <error27@gmail.com> Signed-off-by:
-
Hillf Danton authored
This patch makes it so that we only have one call to fc_rport_error. This patch does not completely consolidate return statements, there is still one return used when not calling fc_rport_error, but alternative solutions made the code more confusing. [ Patch modified by Robert Love ] [ Patch title and commit message edited by Robert Love to make it more relevant ] Signed-off-by:
-
Hillf Danton authored
Here ticks_left is added to record the result of wait_for_completion_timeout(). [ Patch title and description edited by Robert Love to make it more descriptive ] Signed-off-by:
-
Yi Zou authored
The fsp's xfer_ddp is used as indication of the exchange id for the DDPed I/O. We should always initialize it as FC_XID_UNKNOWN for a newly allocated fsp, otherwise the fsp allocated in fc_fcp, i.e., not from queuecommand like LUN RESET that is not doing DDP may still think DDP is setup for it since xid 0 is valid and goes on to call fc_fcp_ddp_done() in fc_fcp_resp() from fc_tm_done(). So, set xfer_ddp as FC_XID_UNKNOWN in fc_fcp_pkt_alloc() now. Also removes the setting of fsp->lp as it's already done when fsp is allocated. Signed-off-by:
-
Hannes Reinecke authored
Classify severity of I/O errors for target, nexus, and transport errors. Signed-off-by:
Mike Snitzer <snitzer@redhat.com> Signed-off-by:
Hannes Reinecke <hare@suse.de> Acked-by:
Jens Axboe <jaxboe@fusionio.com> Signed-off-by:
-
Hannes Reinecke authored
DM now has more information about the nature of the underlying storage failure. Path failure is avoided if a request failed due to a target error. Instead the target error is immediately passed up the stack. Discard requests that fail due to non-target errors may now be retried. Errors restricted to the path will be retried or returned if no paths are available, irregarding the no_path_retry setting. Signed-off-by:
Alasdair G Kergon <agk@redhat.com> Signed-off-by:
-
Hannes Reinecke authored
Instead of just passing 'EIO' for any I/O error we should be notifying the upper layers with more details about the cause of this error. Update the possible I/O errors to: - ENOLINK: Link failure between host and target - EIO: Retryable I/O error - EREMOTEIO: Non-retryable I/O error - EBADE: I/O error restricted to the I_T_L nexus 'Retryable' in this context means that an I/O error _might_ be restricted to the I_T_L nexus (vulgo: path), so retrying on another nexus / path might succeed. 'Non-retryable' in general refers to a target failure, so this error will always be generated regardless of the I_T_L nexus it was send on. I/O errors restricted to the I_T_L nexus might be retried on another nexus / path, but they should _not_ be queued if no paths are available. Signed-off-by:
-
