- 16 Oct, 2018 32 commits
-
-
Bart Van Assche authored
The purpose of sg_alloc_table() is to allocate and initialize an sg-list. Use that function instead of open-coding it. This patch will make it easier to share code for caching sg-list allocations between the SCSI and NVMe target cores. Signed-off-by:
Bart Van Assche <bvanassche@acm.org> Cc: Nicholas Bellinger <nab@linux-iscsi.org> Cc: Mike Christie <mchristi@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Hannes Reinecke <hare@suse.de> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com>
-
Bart Van Assche authored
Instead of duplicating the SECTOR_SHIFT definition from <linux/blkdev.h>, use it. This patch does not change any functionality. Reviewed-by:
Christoph Hellwig <hch@lst.de> Cc: Nicholas Bellinger <nab@linux-iscsi.org> Cc: Mike Christie <mchristi@redhat.com> Cc: Hannes Reinecke <hare@suse.de> Signed-off-by:
-
Bart Van Assche authored
Commit 057085e5 ("target: Fix race for SCF_COMPARE_AND_WRITE_POST checking") removed the code that checks the SCF_COMPARE_AND_WRITE_POST flag. Hence also remove the flag itself. Cc: Nicholas Bellinger <nab@linux-iscsi.org> Cc: Mike Christie <mchristi@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Hannes Reinecke <hare@suse.de> Signed-off-by:
-
Bart Van Assche authored
A value is assigned to the xcopy_op member of struct xcopy_pt_cmd but that value is never used. Hence remove the xcopy_op member. Reviewed-by:
-
Bart Van Assche authored
Change one occurrence of "aleady" into "already" and one occurrence of "is" into "if". Reviewed-by:
-
David Disseldorp authored
sess_err_stats are currently filled on NOP ping timeout, but not Data-Out timeout. Stash details of Data-Out timeouts using a ISCSI_SESS_ERR_CXN_TIMEOUT value for last_sess_failure_type. Signed-off-by:
David Disseldorp <ddiss@suse.de> Signed-off-by:
-
David Disseldorp authored
Replace existing nested code blocks with helper function calls. Signed-off-by:
-
David Disseldorp authored
Events resulting in connection outages like this should be logged as errors. Include the I_T Nexus in the message to aid path identification. Signed-off-by:
-
David Disseldorp authored
Data-Out timeouts resulting in connection outages should be logged as errors. Include the I_T Nexus in the message to aid path identification. Signed-off-by:
-
David Disseldorp authored
Move the ISCSI_IQN_LEN definition up, so that it can be used in more places instead of a hardcoded value. Signed-off-by:
-
Xiang Chen authored
Update registers as follows: - Default value of AIP timer is 1ms, and it is easy for some expanders to cause IO error. Change the value to max value 65ms to avoid IO error for those expanders. - A CQ completion will be reported by HW when 4 CQs have occurred or the aging timer expires, whichever happens first. Sor serial IO scenario, it will still wait 8us for every IO before it is reported. So in the situation, the performance is poor. So to improve it, change the limit time to the least value. For other scenario, it does little affect to the performance. Signed-off-by:
Xiang Chen <chenxiang66@hisilicon.com> Signed-off-by:
John Garry <john.garry@huawei.com> Signed-off-by:
-
Xiang Chen authored
Currently we use the IPTT defined in LLDD to identify IOs. Actually for IOs which are from the block layer, they have tags to identify them. So for those IOs, use tag of the block layer directly, and for IOs which is not from the block layer (such as internal IOs from libsas/LLDD), reserve 96 IPTTs for them. Signed-off-by:
-
Xiang Chen authored
The interrupts of ent72 and ent74 are not processed by PCIe AER handling, so we need to unmask the interrupts and process them first in the driver. Signed-off-by:
-
Xiang Chen authored
If an SSP/SMP IO times out, it may be actually in reality be simultaneously processing completion of the slot in slot_complete_vx_hw(). Then if the slot is freed in slot_complete_vx_hw() (this IPTT is freed and it may be re-used by other slot), and we may abort the wrong slot in hisi_sas_abort_task(). So to solve the issue, free the slot after the check of SAS_TASK_STATE_ABORTED in slot_complete_vx_hw(). Signed-off-by:
-
Xiang Chen authored
If SMP/internal IO times out, we will possibly free the task immediately. However if the IO actually completes at the same time, the IO completion may refer to task which has been freed. So to solve the issue, flush the tasklet to finish IO completion before free'ing slot/task. Signed-off-by:
-
Luo Jiaxing authored
In evaluating hisi_hba, the sas_port may be NULL, so for safety relocate the the check to value possible NULL deference. Signed-off-by:
Luo Jiaxing <luojiaxing@huawei.com> Signed-off-by:
-
Luo Jiaxing authored
At directly attached situation, if the user modifies the sysfs interface of maximum_linkrate and minimum_linkrate to renegotiate the linkrate between SAS controller and target, the value of both files mentioned above should have change to user setting after renegotiate is over, but it remains unchanged. To fix this bug, maximum_linkrate and minimum_linkrate will be directly fed back to relevant sas_phy structure. Signed-off-by:
-
Nicholas Bellinger authored
With the addition of commit 00d909a1 ("scsi: target: Make the session shutdown code also wait for commands that are being aborted") in v4.19-rc, it incorrectly assumes no signals will be pending for task_struct executing the normal session shutdown and I/O quiesce code-path. For example, iscsi-target and iser-target issue SIGINT to all kthreads as part of session shutdown. This has been the behaviour since day one. As-is when signals are pending with se_cmds active in se_sess->sess_cmd_list, wait_event_interruptible_lock_irq_timeout() returns a negative number and immediately kills the machine because of the do while (ret <= 0) loop that was added in commit 00d909a1 to spin while backend I/O is taking any amount of extended time (say 30 seconds) to complete. Here's what it looks like in action with debug plus delayed backend I/O completion: [ 4951.909951] se_sess: 000000003e7e08fa before target_wait_for_sess_cmds [ 4951.914600] target_wait_for_sess_cmds: signal_pending: 1 [ 4951.918015] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 0 [ 4951.921639] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 1 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 2 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 3 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 4 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 5 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 6 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 7 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 8 [ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 9 ... followed by the usual RCU CPU stalls and deadlock. There was never a case pre commit 00d909a1 where wait_for_complete(&se_cmd->cmd_wait_comp) was able to be interrupted, so to address this for v4.19+ moving forward go ahead and use wait_event_lock_irq_timeout() instead so new code works with all fabric drivers. Also for commit 00d909a1, fix a minor regression in target_release_cmd_kref() to only wake_up the new se_sess->cmd_list_wq only when shutdown has actually been triggered via se_sess->sess_tearing_down. Fixes: 00d909a1 ("scsi: target: Make the session shutdown code also wait for commands that are being aborted") Cc: <stable@vger.kernel.org> # v4.19+ Cc: Bart Van Assche <bvanassche@acm.org> Cc: Mike Christie <mchristi@redhat.com> Cc: Hannes Reinecke <hare@suse.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Sagi Grimberg <sagi@grimberg.me> Cc: Bryant G. Ly <bryantly@linux.vnet.ibm.com> Tested-by:
Nicholas Bellinger <nab@linux-iscsi.org> Signed-off-by:
Bryant G. Ly <bly@catalogicsoftware.com> Signed-off-by:
-
Nicholas Bellinger authored
Short of reverting commit 00d909a1 ("scsi: target: Make the session shutdown code also wait for commands that are being aborted") for v4.19, target-core needs a wait_event_t macro can be executed using TASK_UNINTERRUPTIBLE to function correctly with existing fabric drivers that expect to run with signals pending during session shutdown and active se_cmd I/O quiesce. The most notable is iscsi-target/iser-target, while ibmvscsi_tgt invokes session shutdown logic from userspace via configfs attribute that could also potentially have signals pending. So go ahead and introduce wait_event_lock_irq_timeout() to achieve this, and update + rename __wait_event_lock_irq_timeout() to make it accept 'state' as a parameter. Fixes: 00d909a1 ("scsi: target: Make the session shutdown code also wait for commands that are being aborted") Cc: <stable@vger.kernel.org> # v4.19+ Cc: Bart Van Assche <bvanassche@acm.org> Cc: Mike Christie <mchristi@redhat.com> Cc: Hannes Reinecke <hare@suse.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Sagi Grimberg <sagi@grimberg.me> Cc: Bryant G. Ly <bryantly@linux.vnet.ibm.com> Cc: Peter Zijlstra (Intel) <peterz@infradead.org> Tested-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Reviewed-by:
-
Hannes Reinecke authored
When we fail to analyse the payload of a PRLI response we should reset the state machine to retry the PRLI; eventually we will be getting a proper frame. Not doing so will result in a stuck state machine and the port never to be presented to the systsm. Suggested-by:
Chad Dupuis <chad.dupuis@cavium.com> Signed-off-by:
Hannes Reinecke <hare@suse.com> Reviewed-by:
Arun Easi <arun.easi@cavium.com> Signed-off-by:
-
Thomas Abraham authored
We should not assume the payload of a PRLI or PLOGI respons is always present. Signed-off-by:
Thomas Abraham <tabraham@suse.com> Reviewed-by:
-
Hannes Reinecke authored
When an RSCN gets delayed (or not being sent at all), the transport class will detect an error, EH kicks in, and eventually will be setting the device to offline. If we receive an RSCN after that, the device will stay in 'offline'. This patch allows for an 'offline' to 'blocked' transition, thereby allowing the device to become active again. Signed-off-by:
-
Christoph Hellwig authored
The dma_addr_t member is unused ever since we switched the SCSI layer to send down single-segement command using a scatterlist as well many years ago. Signed-off-by:
Finn Thain <fthain@telegraphics.com.au> Signed-off-by:
-
Christoph Hellwig authored
Except for the mac_esp driver, which uses PIO or pseudo DMA, all drivers share the same dma mapping calls. Move the dma mapping into the core code using the scsi_dma_map / scsi_dma_unmap helpers, with a special identify mapping variant triggered off a new ESP_FLAG_NO_DMA_MAP flag for mac_esp. Signed-off-by:
-
Christoph Hellwig authored
We can simplify use esp->dev now. Signed-off-by:
-
Christoph Hellwig authored
esp->dev is a void pointer that points either to a struct device, or a struct platform_device. As we can easily get from the device to the platform_device if needed change it to always point to a struct device and properly type the pointer to avoid errors. Signed-off-by:
-
Christoph Hellwig authored
esp_sbus_map_command_block is called straight from the probe routine without any locks held, so we can safely use GFP_KERNEL here. Signed-off-by:
-
Christoph Hellwig authored
Remove usage of the legacy PCI DMA API. To make this easier we also store a struct device instead of pci_dev in the dev field of struct esp. Signed-off-by:
-
Venkat Gopalakrishnan authored
Per Qcom's UFS host controller HW design, the UFS Tx lane1 clock could be muxed with Tx lane0 clock, hence keep Tx lane1 clock optional by ignoring it if it is not provided in device tree. This change also performs some cleanup to lanes per direction checks when enable/disable lane clocks just for symmetry. Signed-off-by:
Venkat Gopalakrishnan <venkatg@codeaurora.org> Signed-off-by:
Subhash Jadavani <subhashj@codeaurora.org> Signed-off-by:
Can Guo <cang@codeaurora.org> Reviewed-by:
Vivek Gautam <vivek.gautam@codeaurora.org> Reviewed-by:
Douglas Anderson <dianders@chromium.org> Tested-by:
-
Arnd Bergmann authored
We get a warning from 'make headers_check' about a newly introduced usage of integer types in the scsi/scsi_bsg_ufs.h uapi header: usr/include/scsi/scsi_bsg_ufs.h:18: found __[us]{8,16,32,64} type without #include <linux/types.h> Aside from the missing linux/types.h inclusion, I also noticed that it uses the wrong types: 'u32' is not available at all in user space, and 'uint32_t' depends on the inclusion of a standard header that we should not include from kernel headers. Change the all to __u32 and similar types here. I also note the usage of '__be32' and '__be16' that seems unfortunate for a user space API. I wonder if it would be better to define the interface in terms of a CPU-endian structure and convert it in kernel space. Fixes: e77044c5 ("scsi: ufs-bsg: Add support for uic commands in ufs_bsg_request()") Fixes: df032bf2 ("scsi: ufs: Add a bsg endpoint that supports UPIUs") Signed-off-by:Arnd Bergmann <arnd@arndb.de> Reviewed-by:
Avri Altman <avri.altman@wdc.com> Signed-off-by:
-
Christoph Hellwig authored
dma_alloc_coherent allocates memory that can be used by the cpu and the device at the same time, calls to pci_dma_sync_* are not required, and in fact actively harmful on some architectures like arm. Signed-off-by:
Johannes Thumshirn <jthumshirn@suse.de> Signed-off-by:
-
Christoph Hellwig authored
The dma_map_sg / dma_unmap_sg APIs called from scsi_dma_map / scsi_dma_unmap already transfer memory ownership to the device or cpu respectively. Adding additional calls to pci_dma_sync_sg_* will in fact lead to data corruption if we end up using swiotlb for some reason. Also remove the now pointless megaraid_mbox_sync_scb function. Signed-off-by:
-
- 11 Oct, 2018 8 commits
-
-
Colin Ian King authored
There are extraneous parantheses that are causing clang to produce a warning so remove these. Clean up 3 clang warnings: equality comparison with extraneous parentheses [-Wparentheses-equality] Signed-off-by:
Colin Ian King <colin.king@canonical.com> Acked-by:
Ching Huang <ching2048@areca.com.tw> Signed-off-by:
-
Avri Altman authored
Make ufshcd_send_uic_cmd() public for that. Signed-off-by:
Bart Van Assche <Bart.VanAssche@wdc.com> Signed-off-by:
-
Avri Altman authored
Do that for the currently supported UPIUs: query, nop out, and task management. We do not support UPIU of type scsi command yet, while we are using the job's request and reply pointers to hold the payload. We will look into it in later patches. We might need to elaborate the raw upiu api for that. We also still not supporting uic commands: For first phase, we plan to use the existing api, and send only uic commands that are already supported. Anyway, all that will come in the next patch. Signed-off-by:
-
Avri Altman authored
The UFS host software uses a combination of a host register set and Transfer Request Descriptors in system memory to communicate with host controller hardware. In its mmio space, a separate places are assigned to UTP Transfer Request Descriptor ("utrd") list, and to UTP Task Management Request Descriptor ("utmrd") list. The provided API supports utrd-typed requests: nop out and device management commands. It also supports utmrd-type requests: task management requests. Other UPIU types are not supported for now. We utilize the already existing code for tag and task work queues. That is, all utrd-typed UPIUs are "disguised" as device management commands. Similarly, the utmrd-typed UPUIs uses the task management infrastructure. It is up to the caller to fill the upiu request properly, as it will be copied without any further input validations. Signed-off-by: -
Avri Altman authored
Use the structure size in pointer arithmetic instead of an opaque 32 bytes for the over-allocation of descriptors. Signed-off-by:
-
Avri Altman authored
For now, just provide an API to allocate and remove ufs-bsg node. We will use this framework to manage ufs devices by sending UPIU transactions. For the time being, implements an empty bsg_request() - will add some more functionality in coming patches. Nonetheless, we reveal here the protocol we are planning to use: UFS Transport Protocol Transactions. UFS transactions consist of packets called UFS Protocol Information Units (UPIU). There are UPIU’s defined for UFS SCSI commands, responses, data in and data out, task management, utility functions, vendor functions, transaction synchronization and control, and more. By using UPIUs, we get access to the most fine-grained internals of this protocol, and able to communicate with the device in ways, that are sometimes beyond the capacity of the ufs driver. Moreover and as a result, our core structure - ufs_bsg_node has a pretty lean structure: using upiu transactions that contains the outmost detailed info, so we don't really need complex constructs to support it. Signed-off-by:
-
Avri Altman authored
in preparation to send UPIU requests via bsg. Signed-off-by:
-
Christoph Hellwig authored
Add a helper that takes a utp_task_req_desc and issues it, which will be useful for UFS bsg support. Rewrite ufshcd_issue_tm_cmd0x to use this new helper. Signed-off-by:
-
