- 27 Jul, 2010 40 commits
-
-
Joe Eykholt authored
This patch somewhat combines two fixes to remote port handing in libfc. The first problem was that rport work could be queued on a deleted and freed rport. This is handled by not resetting rdata->event ton NONE if the rdata is about to be deleted. However, that fix led to the second problem, described by Bhanu Gollapudi, as follows: > Here is the sequence of events. T1 is first LOGO receive thread, T2 is > fc_rport_work() scheduled by T1 and T3 is second LOGO receive thread and > T4 is fc_rport_work scheduled by T3. > > 1. (T1)Received 1st LOGO in state Ready > 2. (T1)Delete port & enter to RESTART state. > 3. (T1)schdule event_work, since event is RPORT_EV_NONE. > 4. (T1)set event = RPORT_EV_LOGO > 5. (T1)Enter RESTART state as disc_id is set. > 6. (T2)remember to PLOGI, and set event = RPORT_EV_NONE > 6. (T3)Received 2nd LOGO > 7. (T3)Delete Port & enter to RESTART state. > 8. (T3)schedule event_work, since event is RPORT_EV_NONE. > 9. (T3)Enter RESTART state as disc_id is set. > 9. (T3)set event = RPORT_EV_LOGO > 10.(T2)work restart, enter PLOGI state and issues PLOGI > 11.(T4)Since state is not RESTART anymore, restart is not set, and the > event is not reset to RPORT_EV_NONE. (current event is RPORT_EV_LOGO). > 12. Now, PLOGI succeeds and fc_rport_enter_ready() will not schedule > event_work, and hence the rport will never be created, eventually losing > the target after dev_loss_tmo. So, the problem here is that we were tracking the desire for the rport be restarted by state RESTART, which was otherwise equivalent to DELETE. A contributing factor is that we dropped the lock between steps 6 and 10 in thread T2, which allows the state to change, and we didn't completely re-evaluate then. This is hopefully corrected by the following minor redesign: Simplify the rport restart logic by making the decision to restart after deleting the transport rport. That decision is based on a new STARTED flag that indicates fc_rport_login() has been called and fc_rport_logoff() has not been called since then. This replaces the need for the RESTART state. Only restart if the rdata is still in DELETED state and only if it still has the STARTED flag set. Also now, since we clear the event code much later in the work thread, allow for the possibility that the rport may have become READY again via incoming PLOGI, and if so, queue another event to handle that. In the problem scenario, the second LOGO received will cause the LOGO event to occur again. Reported-by:
Bhanu Gollapudi <bprakash@broadcom.com> Signed-off-by:
Joe Eykholt <jeykholt@cisco.com> Signed-off-by:
Robert Love <robert.w.love@intel.com> Signed-off-by:
James Bottomley <James.Bottomley@suse.de>
-
Roel Kluin authored
In fnic_abort_cmd() and fnic_device_reset() assign `rport' earlier to make FNIC_SCSI_DBG() calls cleaner. In fnic_clean_pending_aborts() `rport' is not used. Signed-off-by:
Roel Kluin <roel.kluin@gmail.com> Acked-by:
Abhijeet Joglekar <abjoglek@cisco.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
Yi Zou authored
lport state is enum not bit mask. Signed-off-by:
Yi Zou <yi.zou@intel.com> Signed-off-by:
-
Bhanu Prakash Gollapudi authored
As per FC-BB-5 rev.2, section 7.8.7.1, strict ordering of FIP descriptors is required for ELS requests. Also, look for missing and duplicate critical descriptors. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
Clear virtual link for NPIV ports is now handled by resetting the matching vnport. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
As per FC-BB-5 rev 2, section 7.8.6.2, malformed FIP frame shall be discarded. Drop discovery adv, ELS and CLV's with duplicate critical descriptors. [Resending after incorporating Joe's review comments] Signed-off-by:
-
Joe Eykholt authored
Allow the D flag (indicating that keep-alives are not needed) to be updated dynamically from received FIP advertisements. Signed-off-by:
-
Joe Eykholt authored
keep alives are disabled due to fd_flags set and also stop updating keep alive values in that case. Update select fcf time only if fcf is not already selected or select time is not already determined from parse adv, and then have select time cleared only once after fcf is selected. Changed deadline check to time_after_eq() from time_after() since now next timeout will be on exact 2.5 times FKA followed by first advertisement. Signed-off-by:
Vasu Dev <vasu.dev@intel.com> Signed-off-by:
-
Joe Eykholt authored
[This patch has several improvements to the code in the fip timers. It hasn't been tested yet. I'm sending it out for review. Vasu, perhaps you can merge this with your patch and test it together.] The current code allows an advertisement to be used even if it has been 3 times the FCF keep-alive advertisement period (FKA) since one was received from that FCF. The spec. calls for 2.5 times FKA. Fix this and make sure we detect missed keep-alives promptly. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
Resubmitting after incorporating Joe's review comment. Unsolicited PRLO request is now handled by sending LS_ACC, and then relogin to the remote port if an N-port login session exists for that remote port. Note that this patch should be applied on top of Joe Eykholt's "Fix remote port restart problem" patch. Signed-off-by:
-
Joe Eykholt authored
Some old comments in fc_fcoe.h say TBD long after the standard has been passed by T11. Clean them up. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
As per FC-LS Rev 1.62 table 46, response codes are handled as follows: 1. If the Req executed is true, PRLI is accepted. 2. If Req executed is not set, if resp code is 5, PRLI is not retried and port is logged out. 3. If resp code is anything apart from 1 or 5, PRLI is retired upto max retry count. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
Retry upto max_rport_retry_count when a target responds with LS_RJT for a PRLI request. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
Host does not send discovery solicitation messages if Disc. Adv from FCF are dropped. It restarts sending solicitation only after receiving a Discovery Adv. from FCF. Fix is to restart solicitation immediately after CVL processing. Signed-off-by:
-
Bhanu Prakash Gollapudi authored
Avoid infinite loop while processing FIP ELS or discovery advertisement with non-critical descriptors. Signed-off-by:
-
Joe Eykholt authored
A check in fcoe_ctlr_send_keep_alive() returns if there's no port_id for the local port. This could miss a keep alive if we just did a host reset and have logged off and will log back in. Return only if we are doing the port keep alive, in which case we need to be logged in. Signed-off-by:
-
Wayne Boyer authored
A problem was found where the call to scsi_add_device() fails intermittently for an adapter. This is caused when __scsi_add_device() returns -ENODEV as a result of not calling scsi_probe_and_add_lun() since the call to scsi_host_scan_allowed() fails. scsi_host_scan_allowed() fails because the adapter state is set to SHOST_RECOVERY instead of SHOST_RUNNING. The state of the adapter is being set to SHOST_RECOVERY by scsi_eh_scmd_add() during error handling. This problem is avoided by moving the setting of the allow_restart flag to later in the device initialization sequence. This prevents further error handling if we get a NOT_READY response from a TUR command by causing scsi_check_sense() to return SUCCESS. Therefore, scsi_eh_scmd_add() will not run and the adapter state will remain as SHOST_RUNNING. Signed-off-by:
Wayne Boyer <wayneb@linux.vnet.ibm.com> Acked-by:
Brian King <brking@linux.vnet.ibm.com> Signed-off-by:
-
Dan Carpenter authored
"phba" is always null here so we can't dereference it. Signed-off-by:
Dan Carpenter <error27@gmail.com> Reviewed-by:
Mike Christie <michaelc@cs.wisc.edu> Signed-off-by:
-
Dan Carpenter authored
I added a kfree(pwrb_arr) in front of the return. Signed-off-by:
-
Wayne Boyer authored
Compiling the driver will fail on 32 bit powerpc and other architectures where writeq is not defined. This patch adds a definition for writeq. Signed-off-by:
-
Mike Christie authored
This patch fixes 4 bugs in the connection connect/disconnect cleanup path. 1. If beiscsi_open_conn fails beiscsi_free_ep was always being called, and if beiscsi_open_conn failed because beiscsi_get_cid failed then we would free an unallocated cid. 2. If beiscsi_ep_connect failed due to a beiscsi_open_conn failure it was leaking iscsi_endpoints. 3. beiscsi_ep_disconnect was leaking iscsi_endpoints. beiscsi_ep_disconnect should free the iscsi_endpoint. We cannot do it in beiscsi_conn_stop because that is only called for iscsi connection cleanup. If beiscsi_ep_connect returns success, but then the poll function fails or the connect times out then beiscsi_ep_disconnect will be called to clean up the ep. The conn_stop callout will not be called in that path. 4. beiscsi_conn_stop was freeing the iscsi_endpoint then accessing it a couple lines later. Signed-off-by:
-
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
- Fixed accounting of allocated SCSI buffers when post sgl fails. - Restrict scsi buffer allocation based on LUN count (sdev_cnt). - Create __lpfc_sli_free_rpi that doesn't take out the hbalock. - Modify lpfc_sli_free_rpi to call __lpfc_sli_free_rpi. - Call __lpfc_sli_free_rpi in lpfc_cleanup_pending_mbox. - Do not swap the strings returned in mailbox commands and do not swap byte aligned data in VPD. Signed-off-by:
-
James Smart authored
- Prevent unregistring of unused FCF when FLOGI is pending. - Prevent point to point discovery on a FCoE HBA. - Fixed FCF discovery failure after swapping FCoE port by switching over to fast failover method when no FCF matches in-use FCF. Signed-off-by:
-
James Smart authored
- Allow enabling MSI-X intterupts with fewer vectors than requested by looking at the return value from pci_enable_msix. - Implemented driver PCI AER error handling routines for supporting AER error recovering on SLI4 devices. - Remove redundant SLI_ACTIVE checks Signed-off-by:
-
Anil Ravindranath authored
1. MSI-X interrupt support 2. Driver changes to support new maxRAID controller FW version. The changes are mainly done to handle async notification changes done in newer controller FW version. 3. Added state change notifications to notify applications of controller states. Signed-off-by:
Anil Ravindranath <anil_ravindranath@pmc-sierra.com> Signed-off-by:
-
James Smart authored
Signed-off-by:
-
James Smart authored
Signed-off-by:
-
James Smart authored
- Change the Max receive size on CIN FCFs to 0x800 - (From linux community) Check boundary before checking for NULL. - Update last completion time for completed I/O to prevent heartbeat. - Add Balius PCI Device IDs Signed-off-by:
-
James Smart authored
- Fix hba_queue_depth to reflect actual available XRIs - Add support for new SLER specific firmware status codes. - Free SCSI buffer when iotag allocation fails. Signed-off-by:
-
James Smart authored
- Add poll or wait flag parameter to hba_init_link and hba_down_link. - (From Linux Community) Make return with ENXIO negative. - Remove unused INB code from driver. - Prevent block_magmt_io from returning until mailbox is inactive. Signed-off-by:
-
James Smart authored
- Retry PLOGI up to 48 times when LS_RJT reason is "Unable to supply requested data." - When dev loss timeout occures do not change state if there is an outstanding REG_LOGIN. - Add logic to ignore REG_LOGIN completion if discovery is restarted while waiting for REG_LOGIN. - Only change state on REG_LOGIN completion if still in state waiting for REG_LOGIN completion. - Only send ADISCs to FCP-2 Targets (not Initiators). Signed-off-by:
-
Wayne Boyer authored
A change in the hardware design of the chip for the new adapters changes the default endianness of MMIO operations. This patch adds a register definition which when written to with a predefined value will change the endianness back to what the driver expects. This patch also fixes two problems found during testing. First, the first reserved field in the ipr_hostrcb64_fabirc_desc structure only reserved one byte. The correct amount to reserve is 2 bytes. Second, the reserved field of the ipr_hostrcb64_error structure only reserved 2 bytes. The correct amount to reserve is 16 bytes. Signed-off-by:
-
Julia Lawall authored
The function adpt_i2o_post_wait is called from several places, in some of which, such as adpt_abort, a lock may be held. The functions adpt_i2o_reparse_lct and adpt_i2o_lct_get are called from several places, including adpt_rescan where a lock may be held. The semantic patch that makes this change is as follows: (http://coccinelle.lip6.fr/) // <smpl> @gfp exists@ identifier fn; position p; @@ fn(...) { ... when != spin_unlock_irqrestore when any GFP_KERNEL@p ... when any } @locked@ identifier gfp.fn; @@ spin_lock_irqsave(...) ... when != spin_unlock_irqrestore fn(...) @depends on locked@ position gfp.p; @@ - GFP_KERNEL@p + GFP_ATOMIC // </smpl> Signed-off-by:
Julia Lawall <julia@diku.dk> Signed-off-by:
-
Giridhar Malavali authored
Signed-off-by:
Giridhar Malavali <giridhar.malavali@qlogic.com> Signed-off-by:
-
Andrew Vasquez authored
[SCSI] qla2xxx: For ISP 23xx, select user specified login timeout value if greater than minuimum value(4 secs). Signed-off-by:
-
Giridhar Malavali authored
Signed-off-by:
-
Giridhar Malavali authored
Signed-off-by:
-
Arun Easi authored
Signed-off-by:
-
Lalit Chandivade authored
Driver was not unprotecting correctly, use correct bits to unprotect the flash on ISP 82xx. Signed-off-by:
-
