- 25 Feb, 2011 7 commits
-
-
Christof Schmitt authored
Move the kmem_cache for allocating the qtcb to zfcp_fsf.c and rename it accordingly. Signed-off-by:
Christof Schmitt <christof.schmitt@de.ibm.com> Signed-off-by:
Steffen Maier <maier@linux.vnet.ibm.com> Signed-off-by:
James Bottomley <James.Bottomley@suse.de>
-
Christof Schmitt authored
Switch the allocation of the GPN_FT request data to the FC kmem_cache and remove the zfcp_gpn kmem_cache. Signed-off-by:
-
Christof Schmitt authored
Allocate the data for the GID_PN request through the new FC kmem_cache. While updating the GID_PN code, also introduce a helper function for initializing the CT header for FC nameserver requests. Remove the "paranoia" check as well, the GID_PN request data does not suddenly change. Signed-off-by:
-
Christof Schmitt authored
A data buffer that is passed to the hardware must not cross a page boundary. zfcp uses a series of kmem_caches to align the data to not cross a page boundary. Introduce a new kmem_cache for the FC requests sent from the zfcp driver and use it for the ELS ADISC data. The goal is to migrate to the FC kmem_cache in later patches and remove the request specific kmem_caches. Signed-off-by:
-
Christof Schmitt authored
zfcp requires a mempool for the status read data blocks to resubmit the "status read" requests at any time. Each status read data block has the size of a page (4096 bytes) and needs to be placed in one page. Instead of having a kmem_cache for allocating page sized chunks, use mempool_create_page_pool to create a mempool returning pages and remove the zfcp kmem_cache. Signed-off-by:
-
Christof Schmitt authored
Signed-off-by:
-
Tobias Klauser authored
IS_ERR() already implies unlikely(), so it can be omitted here. Signed-off-by:
Tobias Klauser <tklauser@distanz.ch> Signed-off-by:
-
- 24 Feb, 2011 17 commits
-
-
Mike Christie authored
The ep_disconnect function could be freeing the ep while beiscsi_conn_get_param is running. This has the driver use the get ep param callback instead of the get conn param to fix this. Signed-off-by:
Mike Christie <michaelc@cs.wisc.edu> Signed-off-by:
-
Mike Christie authored
This pactch has iser export the address and port of the endpoint. Signed-off-by:
-
Mike Christie authored
bnx2i has some checks to try and make sure the ep is not destroyed while the addr/port is getting read. However, if after this check: if (!(bnx2i_conn && bnx2i_conn->ep && bnx2i_conn->ep->hba)) goto out; bnx2i_conn->ep is cleared by ep_disconnect then we will oops. This patches fixes the problem by having the driver use the get_ep_param callback instead of get_conn_param. Signed-off-by:
-
Mike Christie authored
This has cxgbi use the iscsi_conn_get_addr_param helper and the get ep callback. Signed-off-by:
-
Mike Christie authored
For drivers using the ep callbacks the addr and port are attached to the endpoint instead of the conn. This adds a callout to the iscsi_transport to get ep values. It also adds locking around the get param call to make sure that ep_disconnect does not free the LLD's ep interconnect structs from under us (the ep has a refcount so it will not go away but the LLD may have structs from other subsystems that are not allocated in the ep so we need to protect them from getting freed). Signed-off-by:
-
Mike Christie authored
This has iscsi_tcp use the iscsi_conn_get_addr_param libiscsi function. It also drops the use of the libiscsi session portal buffers, so they can be removed in the next patches. Instead of copying the values at bind time we get them during get() time. If we are not connected userspace will now get -ENOTCONN, so it knows that connection is disconnected instead of a possible stale value. Signed-off-by:
-
Mike Christie authored
This adds a helper to convert a addr struct to a string. This will be used by the drivers in the next patches. Signed-off-by:
-
Mike Christie authored
The active variable on the iscsi_cls_conn is not used so this patch removes it. Signed-off-by:
-
Mike Christie authored
When iscsid restarts it does not know the connection's endpoint, so it is getting leaked. This fixes the problem by having the iscsi class force a disconnect before a new connection is bound. Signed-off-by:
-
Mike Christie authored
cxgb3i and cxgb4i support TEXT PDU offloading, so set the bits to enable it. Signed-off-by:
-
Eddie Wai authored
Signed-off-by:
Eddie Wai <eddie.wai@broadcom.com> Signed-off-by:
-
Eddie Wai authored
This is part of an effort to support send target discovery via the iSCSI offload path. Signed-off-by:
-
Eddie Wai authored
For scenarios where the shost is not being passed to bnx2i for the iSCSI offload connection request, the code would consult the routing table to select the CNIC device. This code path will erroneously error out if the corresponding L2 interface's MTU has been setup to > 1500. Signed-off-by:
-
Eddie Wai authored
Moved all PCI_DEVICE_ID_NX2_57712(E) definitions to pci_ids.h Signed-off-by:
Michael Chan <mchan@broadcom.com> Acked-by:
Eilon Greenstein <eilong@broadcom.com> Signed-off-by:
-
Eddie Wai authored
The hba will now be unregistered and freed when iSCSI offload is not supported by the NIC. Signed-off-by:
-
Eddie Wai authored
Fixed a bug where the 64-bit LUN field for nopouts were 32-bit swapped. This only pertains to 5771X devices. Signed-off-by:
-
Eddie Wai authored
Allow CNIC to go through the proper cleanup procedure for an endpoint which failed to connect. Proper cleanup is necessary for the chip to reset back to the initial state for the offloaded endpoint. Signed-off-by:
-
- 18 Feb, 2011 13 commits
-
-
James Smart authored
Signed-off-by:
Alex Iannicelli <alex.iannicelli@emulex.com> Signed-off-by:
James Smart <james.smart@emulex.com> Signed-off-by:
-
James Smart authored
- Add the driver debugfs framework for supporting debugfs read and write operations, and iDiag command structure. - Add read and write to SLI4 device PCI config space registers. - Add the driver support of debugfs PCI config space register bits set/clear methods to the provided bitmask. - Add iDiag driver support for SLI4 device queue diagnostic. Signed-off-by:
-
James Smart authored
- Make link speed not supported by port message an error message. - Add support for new SLI failure codes add sysfs parameter to reflect the security setting and current state. - Add all lpfc module parameters to the /sys/modules/lpfc/parameters directory. [jejb: fix up compile failure] Signed-off-by:
-
James Smart authored
FC Discovery changes - Treat received PLOGI while logged in as a relogin (unregister and reregister). - Added a timer to delay Nport discovery when clean bit is cleared and Fabric portname/nodename/FCID is changed. - Invalidate Port's DID when receiving PLOGI from p2p port with CONFIG_PORT mailbox command. Signed-off-by:
-
James Smart authored
RRQ Implementation fixes - Added checks to prevent a call to findnode_did in clr_active_rrq - Added the del_sync_timer call for the rrq_tmr to the stop_hba_timers routine. - Added a check in __lpfc_set_active_rrq for the driver unloading to prevent adding an rrq when the driver is being removed. - Add code to scsi_iocb_cmpl to check for the remote stop and add the rrq. - Added the same check to els retry. - Added code to compare the source did in the els rrq to the vports did and chose the right exchange ID. - Initialize the start_cmd pointer to indicate when we have looped through all of the scsi buffers. - Remove the need for the lock around the clearing of the active bit in the rrq. - Added code to clean the els and fcp xri aborted list and remove the all of the RRQs for a deleted vport. Signed-off-by:
-
James Smart authored
Critical Errors: - Correctly handle non-zero return lpfc_workq_post_event and return ENOMEM - Save the irq level when locking the host_lock in lpfc_findnode_did Bug Fixes: - Adjust payload_length and request_length for sli4_config mailbox commands. - Add the freed sgl/XRI to the tail of the list rather than to the head. - Set the FC_VPORT_NEEDS_INIT_VPI on vport deletes and check it before issuing a fdisc on an els retry. - Only call lpfc_hba_init_link() if phba->cfg_suppress_link_up is LPFC_INITIALIZE_LINK. - Add support for SLI-4 Performance Hints Signed-off-by:
-
Dan Carpenter authored
'!' has higher precedence than '&'. CFGTBL_ChangeReq is 0x1 so the original code is equivelent to if (!doorbell_value) {... Signed-off-by:Dan Carpenter <error27@gmail.com> Acked-by:
Stephen M. Cameron <scameron@beardog.cce.hp.com> Signed-off-by:
-
Stephen M. Cameron authored
We can get completions left over from before the attempted reset which will interfere with the kdump. Better to just not make the attempt in that case. Signed-off-by:
-
Stephen M. Cameron authored
Controller will transfer only 32-bits on completion if it knows we are only using 32-bit tags. Also, some newer controllers apparently (and erroneously) require that we only use 32-bit tags, and that we inform the controller of this. Signed-off-by:
-
Stephen M. Cameron authored
Signed-off-by:
-
Stephen M. Cameron authored
It's not enough to simple avoid putting the board into performant mode, as we have to set up the interrupts differently, etc. When I originally tested this module parameter, I tested it incorrectly without realizing it, and the driver was running in performant mode the whole time unbeknownst to me. Signed-off-by:
-
Stephen M. Cameron authored
Driver's internal queues should be FIFO, not LIFO. This is a port of an almost identical patch from cciss by Jens Axboe. Signed-off-by:
-
Kleber Sacilotto de Souza authored
In a multiple configuration change scenario a remove notification can be followed by an immediate add notification for the same device, which will cause the device to be removed but never added back. This patch fixes the problem by ensuring that in such situations the device will be added back. Signed-off-by:
Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com> Acked-by:
Brian King <brking@linux.vnet.ibm.com> Signed-off-by:
-
- 13 Feb, 2011 3 commits
-
-
James Bottomley authored
The conversion is quite complex given that the libata new error handler has to be hooked into the current libsas timeout and error handling. The way this is done is to process all the failed commands via libsas first, but if they have no underlying sas task (and they're on a sata device) assume they are destined for the libata error handler and send them accordingly. Finally, activate the port recovery of the libata error handler for each port known to the host. This is somewhat suboptimal, since that port may not need recovering, but given the current architecture of the libata error handler, it's the only way; and the spurious activation is harmless. Signed-off-by: -
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 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:
-
