isci: direction check (2.6.39)
Subsystems (2)
RFC 1/1 isci: direction check (2.6.39)
James,
Here is a rundown of where the driver is at relative to a 2.6.39 merge candidate and the outstanding review comments.
=== Status ===
1/ From Christoph: "unify core/lldd objects and remove glue code". This has been done for the remote_device representation with the following commits.
isci: make a remote_node_context a proper member of a remote_device
isci: remove rnc->device back pointer
isci: unify remote_device data structures
isci: move remote_device handling out of the core
isci: cleanup remote device construction and comments
isci: kill smp_discover_response_protocols in favor of domain_device.dev_type
isci: kill smp_discover_response
isci: remove usage of sci_sas_address in scic_sds_remote_device
The full diffs from these last three commits are included below to illustrate the resulting glue code that has been deleted. Repeating this level of unification across the other objects is in progress.
For 2.6.39 we have stopped short of a full rename of all the objects while the other cleanups are pending. I am assuming that as long as the bulk cleanup is in place (diffstat clean i.e. no more drivers/scsi/isci/core/ and the object lifetimes are unified) that further fine grained cleanups can take place post merge? See the diffstat below to check the direction of this effort.
2/ From Christoph: "cleanup state machines". The scic_sds_controller state machine has been through the full gamut of cleanups and it is a matter of repeating the same template (from the commits below) across the other state machines.
8e131f53 isci: controller state machine cleanup step1 ba647271 isci: controller state machine cleanup step2 03355704 isci: controller state machine cleanup step3 13ca34a2 isci: remove base_controller abstraction 74b59773 isci: remove scic_controller state handlers
3/ From you: "no changes outside of drivers/scsi/isci/". I have reverted the tree's local copy of the pci_map_biosrom implementation that is pending in the x86 maintainer tree [1]. So a git diff --stat linus... shows only modifications in drivers/scsi/isci/ drivers/scsi/Makefile and drivers/scsi/Kconfig.
The x86-pending branch has the merge with tip.git and the patch to re-introduce its usage in the driver post 2.6.39.
4/ From the beginning: "Remove the intel_{sas|sata|scsi.h} headers". This is still in progress and is partially addressed by the unification work. We have found that there are some definitions that the driver needs that do not exist in include/scsi/. For those we will include a local definition with a note to move them to their proper location post merge.
5/ Miscellaneous, "remove timer list abstraction", "sci_base_object removal" and other minor cleanups are still in the works.
=== Issues that the review has not covered to date ===
1/ The driver's abort/error handling and task lifetime. This has likely been blocked behind review of the overall driver structure. At this point it is looking like a post 2.6.39 effort.
2/ A libsas gap around reset handling. We are currently seeing sas drives in a software raid array drop off because domain discovery after a device reset finds the device missing before it recovers from the reset. We have a fix that can be localized to the driver but that is likely a layering violation. From our determination libsas should be filtering broadcast change notifications while resets are pending (I'll send more details in a separate thread). For now this workaround is not included in the git tree.
3/ The status of the commits on the libsas-pending branch
libsas: flush initial device discovery before completing ->scan_finished()
libsas: fix/amend device gone notification in sas_deform_port()
libsas: check dev->gone before submitting sata i/oI'll plan to resend for 2.6.40.
The isci git tree has been updated with the following (since the last announcement [2])
git://git.kernel.org/pub/scm/linux/kernel/git/djbw/isci.git master
Bartosz Barcinski (2):
isci: sparse warnings cleanup
isci: audit usage of BUG_ON macro in isci driver
Christoph Hellwig (2):
isci: simplify dma coherent allocation
isci: remove scic_controller state handlers
Dan Williams (16):
isci: replace sci_sas_link_rate with sas_linkrate
isci: fix oem parameter header definition
isci: validate oem parameters early, and fallback
isci: rely on irq core for intx multiplexing, and silence screaming intx
isci: make a remote_node_context a proper member of a remote_device
isci: remove rnc->device back pointer
isci: unify remote_device data structures
isci: move remote_device handling out of the core
isci: cleanup remote device construction and comments
isci: kill smp_discover_response_protocols in favor of domain_device.dev_type
isci: kill smp_discover_response
isci: remove usage of sci_sas_address in scic_sds_remote_device
isci: remove scic_sds_remote_device_get_port_index
isci: allow fallback to option-rom if efi variable retrieval fails
isci: defer pci_map_biosrom
isci: fix CONFIG_EFI=n compile error
Dave Jiang (3):
isci: replace this_* and the_* variables with more meaningful names
isci: removing non-working ATAPI code
isci: Remove excessive log noise with expander hot-unplug
Maciej Trela (3):
isci: remove base_remote_device abstraction
isci: remove base_port abstraction
isci: remove base_phy abstraction
drivers/scsi/isci/Makefile | 14 +-
drivers/scsi/isci/core/intel_sas.h | 211 ---
drivers/scsi/isci/core/sati_device.h | 156 --
drivers/scsi/isci/core/sati_translator_sequence.h | 304 ----
drivers/scsi/isci/core/sati_types.h | 145 --
.../isci/core/sci_base_memory_descriptor_list.c | 159 --
.../isci/core/sci_base_memory_descriptor_list.h | 153 --
drivers/scsi/isci/core/sci_base_phy.h | 200 ---
drivers/scsi/isci/core/sci_base_port.h | 187 ---
drivers/scsi/isci/core/sci_base_remote_device.h | 274 ----
.../scsi/isci/core/sci_memory_descriptor_list.h | 168 --
drivers/scsi/isci/core/sci_pool.h | 60 +-
drivers/scsi/isci/core/scic_config_parameters.h | 2 +
drivers/scsi/isci/core/scic_controller.h | 8 +-
drivers/scsi/isci/core/scic_phy.h | 3 +-
drivers/scsi/isci/core/scic_remote_device.h | 294 ----
drivers/scsi/isci/core/scic_sds_controller.c | 1629 ++++++-------------
drivers/scsi/isci/core/scic_sds_controller.h | 175 +--
drivers/scsi/isci/core/scic_sds_phy.c | 720 ++++----
drivers/scsi/isci/core/scic_sds_phy.h | 94 +-
drivers/scsi/isci/core/scic_sds_port.c | 764 ++++-----
drivers/scsi/isci/core/scic_sds_port.h | 231 ++--
.../isci/core/scic_sds_port_configuration_agent.c | 2 +-
drivers/scsi/isci/core/scic_sds_remote_device.c | 1714 --------------------
drivers/scsi/isci/core/scic_sds_remote_device.h | 548 -------
drivers/scsi/isci/core/scic_sds_request.c | 251 ++--
drivers/scsi/isci/core/scic_sds_request.h | 40 +-
drivers/scsi/isci/core/scic_sds_smp_request.c | 132 +-
drivers/scsi/isci/core/scic_sds_smp_request.h | 3 +-
drivers/scsi/isci/core/scic_sds_ssp_request.c | 30 +-
.../scsi/isci/core/scic_sds_stp_packet_request.c | 806 ---------
.../scsi/isci/core/scic_sds_stp_packet_request.h | 39 -
drivers/scsi/isci/core/scic_sds_stp_request.c | 387 +++---
drivers/scsi/isci/core/scic_sds_stp_request.h | 8 +-
.../isci/core/scic_sds_unsolicited_frame_control.c | 66 +-
.../isci/core/scic_sds_unsolicited_frame_control.h | 22 +-
drivers/scsi/isci/host.c | 127 +--
drivers/scsi/isci/host.h | 21 +-
drivers/scsi/isci/init.c | 42 +-
drivers/scsi/isci/port.c | 2 +-
drivers/scsi/isci/port.h | 1 +
drivers/scsi/isci/probe_roms.c | 31 +-
drivers/scsi/isci/probe_roms.h | 2 +-
drivers/scsi/isci/remote_device.c | 1655 +++++++++++++++++--
drivers/scsi/isci/remote_device.h | 688 ++++++++-
...remote_node_context.c => remote_node_context.c} | 479 +++---
...remote_node_context.h => remote_node_context.h} | 31 +-
...sds_remote_node_table.c => remote_node_table.c} | 4 +-
...sds_remote_node_table.h => remote_node_table.h} | 0
drivers/scsi/isci/request.c | 33 +-
drivers/scsi/isci/request.h | 5 +-
drivers/scsi/isci/sata.c | 3 +-
.../scsi/isci/{core => }/scu_remote_node_context.h | 0
...sds_smp_remote_device.c => smp_remote_device.c} | 120 +-
...sds_stp_remote_device.c => stp_remote_device.c} | 377 ++---
drivers/scsi/isci/task.c | 77 +-
56 files changed, 4601 insertions(+), 9096 deletions(-)
delete mode 100644 drivers/scsi/isci/core/sati_device.h
delete mode 100644 drivers/scsi/isci/core/sati_translator_sequence.h
delete mode 100644 drivers/scsi/isci/core/sati_types.h
delete mode 100644 drivers/scsi/isci/core/sci_base_memory_descriptor_list.c
delete mode 100644 drivers/scsi/isci/core/sci_base_memory_descriptor_list.h
delete mode 100644 drivers/scsi/isci/core/sci_base_phy.h
delete mode 100644 drivers/scsi/isci/core/sci_base_port.h
delete mode 100644 drivers/scsi/isci/core/sci_base_remote_device.h
delete mode 100644 drivers/scsi/isci/core/sci_memory_descriptor_list.h
delete mode 100644 drivers/scsi/isci/core/scic_remote_device.h
delete mode 100644 drivers/scsi/isci/core/scic_sds_remote_device.c
delete mode 100644 drivers/scsi/isci/core/scic_sds_remote_device.h
delete mode 100644 drivers/scsi/isci/core/scic_sds_stp_packet_request.c
rename drivers/scsi/isci/{core/scic_sds_remote_node_context.c => remote_node_context.c} (71%)
rename drivers/scsi/isci/{core/scic_sds_remote_node_context.h => remote_node_context.h} (92%)
rename drivers/scsi/isci/{core/scic_sds_remote_node_table.c => remote_node_table.c} (99%)
rename drivers/scsi/isci/{core/scic_sds_remote_node_table.h => remote_node_table.h} (100%)
rename drivers/scsi/isci/{core => }/scu_remote_node_context.h (100%)
rename drivers/scsi/isci/{core/scic_sds_smp_remote_device.c => smp_remote_device.c} (69%)
rename drivers/scsi/isci/{core/scic_sds_stp_remote_device.c => stp_remote_device.c} (62%)commit 2db162f22e4176baa0a9ae20932178ddb5550050
Author: Christoph Hellwig <[email protected]> Date: Sat Apr 2 08:15:04 2011 -0400
isci: simplify dma coherent allocation
Remove the insane infrastructure for preallocating coheren DMA regions, and just allocate the memory where needed. This also gets rid of the aligment adjustments given that Documentation/DMA-API-HOWTO.txt sais:
"The cpu return address and the DMA bus master address are both
guaranteed to be aligned to the smallest PAGE_SIZE order which is greater than or equal to the requested size. This invariant exists (for example) to guarantee that if you allocate a chunk which is smaller than or equal to 64 kilobytes, the extent of the buffer you receive will not cross a 64K boundary."
commit 74b597736d5792690b6953c51d72e3c7d29c2313
Author: Christoph Hellwig <[email protected]> Date: Sat Apr 2 08:15:20 2011 -0400
isci: remove scic_controller state handlers
Remove the state handler indirections for the scic_controller, and replace them with procedural calls that check for the correct state first.
commit 2c4302dce071ff3b1e5e1f4e168cd66a97d4c7ff
Author: Maciej Trela <[email protected]> Date: Tue Apr 12 17:28:35 2011 -0700
isci: remove base_remote_device abstraction
Merge struct sci_base_remote_device into scic_sds_remote_device. As for now sci_base_remote_device was accessed indirectly using scic_sds_remote_device->parent field. Both machine state handlers are also merged together.
commit 2ac7e0116df9aa5e8cf2b834c096c2f11c82c10d
Author: Maciej Trela <[email protected]> Date: Tue Apr 12 17:28:37 2011 -0700
isci: remove base_port abstraction
Merge struct sci_base_port into scic_sds_port. Until now sci_base_port was referenced indirectly with scic_sds_port->parent field.
'sci_base_port' state machine handlers were also incorporated into scic_sds_port handlers.
commit f32e3000c55b826e5268afe978a31f5eac50f2e3
Author: Maciej Trela <[email protected]> Date: Tue Apr 12 17:28:39 2011 -0700
isci: remove base_phy abstraction
Merge struct sci_base_phy into scic_sds_phy. Until now sci_base_phy was referenced using scic_sds_phy->parent field.
'sci_base_phy' state machine handlers were also merged into scic_sds_phy state handlers.
commit d71288553c9da16524bd3cbe8bb388678e2c8302
Author: Dan Williams <[email protected]> Date: Thu Apr 14 18:27:49 2011 -0700
isci: replace sci_sas_link_rate with sas_linkrate
Drop duplicated enum definition.
commit 6a0340c1bca8f396531653b43bc55b38bd58b491
Author: Bartosz Barcinski <[email protected]> Date: Tue Apr 12 17:28:41 2011 -0700
isci: sparse warnings cleanup
Clean warnings and errors reported by sparse tool.
request.c:430:50: warning: mixing different enum types
remote_device.c:534:39: warning: symbol 'flags' shadows an earlier one task.c:495:44: warning: mixing different enum types scic_sds_controller.c:2155:24: warning: mixing different enum types scic_sds_controller.c:2272:36: warning: mixing different enum types scic_sds_controller.c:2911:38: warning: incorrect type in initializer (different address spaces) scic_sds_controller.c:2913:25: warning: incorrect type in argument 2 (different address spaces) scic_sds_request.c:875:34: warning: cast removes address space of expression scic_sds_request.c:876:123: warning: incorrect type in argument 2 (different address spaces) scic_sds_port.c:585:51: warning: incorrect type in assignment (different address spaces) scic_sds_port.c:712:9: warning: incorrect type in argument 2 (different address spaces) scic_sds_port.c:1770:25: warning: incorrect type in argument 2 (different address spaces)
commit d16ee15014ecf411282cb0e777f0e132135ab67d
Author: Bartosz Barcinski <[email protected]> Date: Tue Apr 12 17:28:43 2011 -0700
isci: audit usage of BUG_ON macro in isci driver
Removes unnecessary usage of BUG_ON macro, excluding core directory. In some cases macro is unnecesary, check is done in caller function. In other cases macro is replaced by if construction with appropriate warning.
commit 91203928719e898f967441a39c7816a80e0fd661
Author: Dan Williams <[email protected]> Date: Tue Apr 19 15:29:25 2011 -0700
isci: fix oem parameter header definition
The element_length is 2 bytes.
commit a49dab1e977b6918918e8817123cac7ee64e6fa0
Author: Dan Williams <[email protected]> Date: Wed Apr 20 12:57:08 2011 -0700
isci: validate oem parameters early, and fallback
If the platform specifies invalid parameters warn the user and fallback to internal defaults rather than fail the driver load altogether.
commit 5aa5c74b003d9755b8c9ed95f5a624dba67b49e7
Author: Dave Jiang <[email protected]> Date: Thu Apr 21 05:34:49 2011 +0000
isci: replace this_* and the_* variables with more meaningful names
Removed any instances of the_* and this_* to variable names that are more meaningful and tell us what they actually are.
commit 88f5fad12e291afd891b6e3e446478a0cd3fd435
Author: Dan Williams <[email protected]> Date: Tue Apr 19 12:32:51 2011 -0700
isci: rely on irq core for intx multiplexing, and silence screaming intx
Remove the extra logic to poll each controller for interrupts, that's the core's job for shared interrupts.
While testing noticed that a number of interrupts fire while waiting for the completion tasklet to run, so added an irq-ack.
commit 21eabe033a27f1a1ef871a21e3d5ca2a47b3fb63
Author: Dan Williams <[email protected]> Date: Tue Apr 19 13:48:49 2011 -0700
isci: make a remote_node_context a proper member of a remote_device
A rnc object has the same lifetime as its associated remote_device. It might get re-initialized, but a remote device always has an rnc member. Preparation for unifying scic_sds_remote_device and isci_remote_device
commit 03f85d1289ffe62c46286418d79abfa7813fc72f
Author: Dan Williams <[email protected]> Date: Tue Apr 19 18:35:58 2011 -0700
isci: remove rnc->device back pointer
Now that they are one in the same object remove the back pointer reference in favor of container_of.
commit c5a3ba2881f3d8dd8d5bb8e25ba98923b8a9b92b
Author: Dan Williams <[email protected]> Date: Thu Apr 21 18:14:45 2011 -0700
isci: unify remote_device data structures
Make it explicit that isci_remote_device and scic_sds_remote_device are one in the same object.
commit 3276e8fbe5eab91ef46448c6bf8d7919841f1e03
Author: Dan Williams <[email protected]> Date: Fri Apr 22 19:18:03 2011 -0700
isci: move remote_device handling out of the core
Now that the core/lldd remote_device data structures are nominally unified merge the corresponding sources into the top-level directory. Also move the remote_node_context infrastructure which has no analog at the lldd level.
commit 76da8017a5cfa696caa659e7bd28cab05c4b3747
Author: Dan Williams <[email protected]> Date: Mon Apr 25 11:48:29 2011 -0700
isci: cleanup remote device construction and comments
The construction routines scic_remote_device_[de]a_construct both reference the need to call scic_remote_device_construct first. Delete that comment and just have them call it explicitly, also:
* move the comments from header to source
* delete dead references to scic_[de]a_remote_device_add_phy
commit eb66c0a4ba77c362669ea9f48b88759121eb7201
Author: Dan Williams <[email protected]> Date: Thu Apr 21 18:44:45 2011 -0700
isci: kill smp_discover_response_protocols in favor of domain_device.dev_type
This is step 1 of removing the contortions to:
1/ unparse expander phy data into a smp discover frame
2/ open-code-parse the smp discover fram into a domain_device.dev_type equivalent
libsas has already spent cycles determining the dev_type, so now that scic_sds_remote_device is unified with isci_remote_device we can directly reference dev_type.
This might also change multi-level expander detection as we previously only looked at dev_type == EDGE_DEV and we did not consider the FANOUT_DEV case.
drivers/scsi/isci/core/scic_sds_controller.h +11 -9
@@ -74,9 +74,9 @@#include "scic_sds_port.h" #include "scic_sds_phy.h" #include "remote_node_table.h" +#include "remote_device.h" #include "scu_registers.h" #include "scu_constants.h" -#include "scu_remote_node_context.h" #include "scu_task_context.h" #include "scu_unsolicited_frame.h" #include "scic_sds_unsolicited_frame_control.h"
@@ -498,14 +498,16 @@ enum scic_sds_controller_states {#define scic_sds_io_sequence_increment (value) \ ((value) = (((value) + 1) & 0x000F)) -#define scic_sds_remote_device_node_count (device) \ - (\ - (\ - (device)->target_protocols.u.bits.attached_stp_target \ - && ((device)->is_direct_attached != true) \ - ) \ - ? SCU_STP_REMOTE_NODE_COUNT : SCU_SSP_REMOTE_NODE_COUNT \ - ) +/* expander attached sata devices require 3 rnc slots */ +static inline int scic_sds_remote_device_node_count (struct scic_sds_remote_device *sci_dev) +{ +struct domain_device *dev =sci_dev_to_domain (sci_dev); + +if ((dev->dev_type ==SATA_DEV || (dev->tproto &SAS_PROTOCOL_STP )) && + !sci_dev->is_direct_attached) +return SCU_STP_REMOTE_NODE_COUNT ; +return SCU_SSP_REMOTE_NODE_COUNT ; +}/** * scic_sds_controller_set_invalid_phy() -
drivers/scsi/isci/core/scic_sds_port.c +1 -1
@@ -434,7 +434,7 @@ void scic_sds_port_get_attached_sas_address(* multi-protocol support. * */ -void scic_sds_port_get_attached_protocols ( +static void scic_sds_port_get_attached_protocols (struct scic_sds_port *sci_port,struct sci_sas_identify_address_frame_protocols *protocols) {
drivers/scsi/isci/core/scic_sds_port.h +0 -4
@@ -447,8 +447,4 @@ void scic_sds_port_get_attached_sas_address(struct scic_sds_port *sci_port, struct sci_sas_address *sas_address); -void scic_sds_port_get_attached_protocols ( -struct scic_sds_port *sci_port, -struct sci_sas_identify_address_frame_protocols *protocols); -#endif /* _SCIC_SDS_PORT_H_ */
drivers/scsi/isci/core/scic_sds_request.c +9 -15
@@ -1758,8 +1758,8 @@ enum sci_status scic_io_request_construct(struct scic_sds_controller *scic,struct scic_sds_request *sci_req, struct scic_sds_request **new_scic_io_request_handle) { + struct domain_device *dev = sci_dev_to_domain(sci_dev);enum sci_status status =SCI_SUCCESS ; -struct smp_discover_response_protocols device_protocol;/* Build the common part of the request */ scic_sds_general_request_construct(scic, sci_dev, io_tag,
@@ -1768,19 +1768,16 @@ enum sci_status scic_io_request_construct(struct scic_sds_controller *scic,if (sci_dev->rnc.remote_node_index ==SCIC_SDS_REMOTE_NODE_CONTEXT_INVALID_INDEX )return SCI_FAILURE_INVALID_REMOTE_DEVICE ; -scic_remote_device_get_protocols (sci_dev, &device_protocol); - -if (device_protocol.u.bits.attached_ssp_target) { +if (dev->dev_type ==SAS_END_DEV ) {scic_sds_ssp_io_request_assign_buffers (sci_req); - }else if (device_protocol.u.bits.attached_stp_target) { + }else if (dev->dev_type ==SATA_DEV || (dev->tproto &SAS_PROTOCOL_STP )) {scic_sds_stp_request_assign_buffers (sci_req);memset (sci_req->command_buffer,0 ,sizeof (struct sata_fis_reg_h2d )); - }else if (device_protocol.u.bits.attached_smp_target) { + }else if (dev_is_expander (dev)) {scic_sds_smp_request_assign_buffers (sci_req);memset (sci_req->command_buffer,0 ,sizeof (struct smp_request )); - }else { + }else status =SCI_FAILURE_UNSUPPORTED_PROTOCOL ; - }if (status ==SCI_SUCCESS ) { memset(sci_req->task_context_buffer, 0,
@@ -1798,17 +1795,15 @@ enum sci_status scic_task_request_construct(struct scic_sds_controller *scic,struct scic_sds_request *sci_req, struct scic_sds_request **new_sci_req) { + struct domain_device *dev = sci_dev_to_domain(sci_dev);enum sci_status status =SCI_SUCCESS ; -struct smp_discover_response_protocols device_protocol;/* Build the common part of the request */ scic_sds_general_request_construct (scic, sci_dev, io_tag, user_io_request_object, sci_req); -scic_remote_device_get_protocols (sci_dev, &device_protocol); - -if (device_protocol.u.bits.attached_ssp_target) { +if (dev->dev_type ==SAS_END_DEV ) {scic_sds_ssp_task_request_assign_buffers (sci_req); sci_req->has_started_substate_machine = true;
@@ -1820,11 +1815,10 @@ enum sci_status scic_task_request_construct(struct scic_sds_controller *scic, scic_sds_io_request_started_task_mgmt_substate_table,SCIC_SDS_IO_REQUEST_STARTED_TASK_MGMT_SUBSTATE_AWAIT_TC_COMPLETION ); - }else if (device_protocol.u.bits.attached_stp_target) { + }else if (dev->dev_type ==SATA_DEV || (dev->tproto &SAS_PROTOCOL_STP ))scic_sds_stp_request_assign_buffers (sci_req); - }else { +else status =SCI_FAILURE_UNSUPPORTED_PROTOCOL ; - }if (status ==SCI_SUCCESS ) { sci_req->is_task_management_request = true;
drivers/scsi/isci/remote_device.c +47 -63
184 lines not on this page · raw message
drivers/scsi/isci/remote_device.h +12 -22
@@ -100,11 +100,6 @@ struct scic_sds_remote_device {enum sas_linkrate connection_rate;/** -* This field contains the allowed target protocols for this remote device. -*/ -struct smp_discover_response_protocols target_protocols; - -/** * This field contains the device SAS address. */ struct sci_sas_address device_address;
@@ -264,23 +259,6 @@ enum sci_status scic_remote_device_reset_complete(enum sas_linkrate scic_remote_device_get_connection_rate (struct scic_sds_remote_device *remote_device); -/** -* scic_remote_device_get_protocols() - This method will indicate which -* protocols are supported by this remote device. -* @remote_device: This parameter specifies the device for which to return the -* protocol. -* @protocols: This parameter specifies the output values, from the remote -* device object, which indicate the protocols supported by the supplied -* remote_device. -* -* The type of protocols supported by this device. The values are returned as -* part of a bit mask in order to allow for multi-protocol support. -*/ -void scic_remote_device_get_protocols ( -struct scic_sds_remote_device *remote_device, -struct smp_discover_response_protocols *protocols); - -#if !defined(DISABLE_ATAPI )/** * scic_remote_device_is_atapi() -
@@ -477,6 +455,18 @@ static inline struct scic_sds_remote_device *rnc_to_dev(struct scic_sds_remote_nreturn sci_dev; } +static inline struct domain_device *sci_dev_to_domain (struct scic_sds_remote_device *sci_dev) +{ +struct isci_remote_device *idev =container_of (sci_dev,typeof (*idev), sci); + +return idev->domain_dev; +} + +static inline bool dev_is_expander (struct domain_device *dev) +{ +return dev->dev_type ==EDGE_DEV || dev->dev_type ==FANOUT_DEV ; +} +typedef enum sci_status (*scic_sds_remote_device_request_handler_t )(struct scic_sds_remote_device *device,struct scic_sds_request *request);
drivers/scsi/isci/remote_node_context.c +22 -41
123 lines not on this page · raw message
drivers/scsi/isci/request.c +11 -8
@@ -175,7 +175,6 @@ static enum sci_status isci_io_request_build(struct isci_request *request, struct isci_remote_device *isci_device) { - struct smp_discover_response_protocols dev_protocols; enum sci_status status = SCI_SUCCESS;struct sas_task *task =isci_request_access_task (request);struct scic_sds_remote_device *sci_device = &isci_device->sci;
@@ -228,15 +227,19 @@ static enum sci_status isci_io_request_build(sci_object_set_association (request->sci_request_handle, request); -/* Determine protocol and call the appropriate basic constructor */ -scic_remote_device_get_protocols (sci_device, &dev_protocols); -if (dev_protocols.u.bits.attached_ssp_target) +switch (task->task_proto) { +case SAS_PROTOCOL_SMP : + status =isci_smp_request_build (request); +break ; +case SAS_PROTOCOL_SSP : status =isci_request_ssp_request_construct (request); -else if (dev_protocols.u.bits.attached_stp_target) +break ; +case SAS_PROTOCOL_SATA : +case SAS_PROTOCOL_STP : +case SAS_PROTOCOL_SATA |SAS_PROTOCOL_STP : status =isci_request_stp_request_construct (request); -else if (dev_protocols.u.bits.attached_smp_target) - status =isci_smp_request_build (request); -else { +break ; +default :dev_warn (&isci_host->pdev->dev,"%s: unknown protocol\n" , __func__);return SCI_FAILURE ;
drivers/scsi/isci/task.c +5 -15
@@ -273,15 +273,14 @@ static enum sci_status isci_task_request_build(enum sci_status status =SCI_FAILURE ;struct isci_request *request =NULL ;struct isci_remote_device *isci_device; -/* struct sci_sas_identify_address_frame_protocols dev_protocols; */ -struct smp_discover_response_protocols dev_protocols; - +struct domain_device *dev;dev_dbg (&isci_host->pdev->dev,"%s: isci_tmf = %p\n" , __func__, isci_tmf); isci_device = isci_tmf->device; sci_device = &isci_device->sci; + dev = isci_device->domain_dev;/* do common allocation and init of request object. */ status = isci_request_alloc_tmf(
@@ -319,16 +318,8 @@ static enum sci_status isci_task_request_build(request ); -scic_remote_device_get_protocols ( - sci_device, - &dev_protocols - ); - -/* let the core do it's protocol -* specific construction. -*/ -if (dev_protocols.u.bits.attached_ssp_target) { - +/* XXX convert to get this from task->tproto like other drivers */ +if (dev->dev_type ==SAS_END_DEV ) { isci_tmf->proto =SAS_PROTOCOL_SSP ; status =scic_task_request_construct_ssp ( request->sci_request_handle
@@ -337,8 +328,7 @@ static enum sci_status isci_task_request_build( goto errout; } -if (dev_protocols.u.bits.attached_stp_target) { - +if (dev->dev_type ==SATA_DEV || (dev->tproto &SAS_PROTOCOL_STP )) { isci_tmf->proto =SAS_PROTOCOL_SATA ; status =isci_sata_management_task_request_build (request);
commit 247e6242acaf1a82631d6fbc6b1f21730b506a2b
Author: Dan Williams <[email protected]> Date: Mon Apr 25 14:29:29 2011 -0700
isci: kill smp_discover_response
An lldd need never look at the contents of an smp_discover_response frame. Kill the remaining locations where isci is looking at it:
1/ covering for expanders that do not set the stp_attached bit (already
handled by sas_ex_discover_end_dev)
2/ an overkill method to notifiy the rest of the driver about remote_device
sas addresses
drivers/scsi/isci/core/intel_sas.h +0 -197
209 lines not on this page · raw message
drivers/scsi/isci/core/scic_sds_smp_request.c +0 -20
@@ -339,26 +339,6 @@ static enum sci_status scic_sds_smp_request_await_response_frame_handler( smp_response_buffer,sizeof (union smp_response_body ) /sizeof (u32) ); -if (rsp_hdr->function ==SMP_FUNCTION_DISCOVER ) { -struct smp_response *smp_resp; - - smp_resp = (struct smp_response *)user_smp_buffer; - -/* -* Some expanders only report an attached SATA device, and -* not an STP target. Since the core depends on the STP -* target attribute to correctly build I/O, set the bit now -* if necessary. */ -if (smp_resp->response.discover.protocols.u.bits.attached_sata_device - && !smp_resp->response.discover.protocols.u.bits.attached_stp_target) { - smp_resp->response.discover.protocols.u.bits.attached_stp_target =1 ; - -dev_dbg (scic_to_dev (sci_req->owning_controller), -"%s: scic_sds_smp_request_await_response_frame_handler(0x%p) Found SATA dev, setting STP bit.\n" , - __func__, sci_req); - } - } -/* * Don't need to copy to user space. User instead will refer to * core request's response buffer. */
drivers/scsi/isci/remote_device.c +20 -118
182 lines not on this page · raw message
drivers/scsi/isci/remote_node_context.c +3 -2
@@ -124,8 +124,9 @@ static void scic_sds_remote_node_context_construct_buffer( rnc->ssp.logical_port_index =scic_sds_remote_device_get_port_index (sci_dev); - rnc->ssp.remote_sas_address_hi =SCIC_SWAP_DWORD (sci_dev->device_address.high); - rnc->ssp.remote_sas_address_lo =SCIC_SWAP_DWORD (sci_dev->device_address.low); +/* address is always big endian, destination is always little */ + rnc->ssp.remote_sas_address_hi =swab32 (sci_dev->device_address.high); + rnc->ssp.remote_sas_address_lo =swab32 (sci_dev->device_address.low); rnc->ssp.nexus_loss_timer_enable = true; rnc->ssp.check_bit = false;
commit c8cec8e26f0d7c2228b2a14f451e9b187458cbb8
Author: Dan Williams <[email protected]> Date: Tue Apr 26 09:41:52 2011 -0700
isci: remove usage of sci_sas_address in scic_sds_remote_device
The sas address can be retrieved from the domain device and then converted to the always little-endian format in the remote node context.
drivers/scsi/isci/remote_device.c +0 -3
@@ -1501,8 +1501,6 @@ static enum sci_status scic_remote_device_da_construct(struct scic_sds_port *sci sci_dev->rnc.remote_node_index = remote_node_index; -scic_sds_port_get_attached_sas_address (sci_port, &sci_dev->device_address); -if (dev->dev_type ==SAS_END_DEV ) sci_dev->has_ready_substate_machine = false;else if (dev->dev_type ==SATA_DEV || (dev->tproto &SAS_PROTOCOL_STP )) {
@@ -1553,7 +1551,6 @@ static enum sci_status scic_remote_device_ea_construct(struct scic_sds_port *scienum sci_status status;scic_remote_device_construct (sci_port, sci_dev); -memcpy (&sci_dev->device_address, dev->sas_addr,SAS_ADDR_SIZE ); status =scic_sds_controller_allocate_remote_node_context ( scic, sci_dev, &sci_dev->rnc.remote_node_index);
drivers/scsi/isci/remote_device.h +0 -5
@@ -100,11 +100,6 @@ struct scic_sds_remote_device {enum sas_linkrate connection_rate;/** -* This field contains the device SAS address. -*/ -struct sci_sas_address device_address; - -/** * This filed is assinged the value of true if the device is directly * attached to the port. */
drivers/scsi/isci/remote_node_context.c +5 -3
@@ -110,6 +110,7 @@ static void scic_sds_remote_node_context_construct_buffer(struct domain_device *dev =sci_dev_to_domain (sci_dev);union scu_remote_node_context *rnc;struct scic_sds_controller *scic; +__le64 sas_addr; scic =scic_sds_remote_device_get_controller (sci_dev);
@@ -124,9 +125,10 @@ static void scic_sds_remote_node_context_construct_buffer( rnc->ssp.logical_port_index =scic_sds_remote_device_get_port_index (sci_dev); -/* address is always big endian, destination is always little */ - rnc->ssp.remote_sas_address_hi =swab32 (sci_dev->device_address.high); - rnc->ssp.remote_sas_address_lo =swab32 (sci_dev->device_address.low); +/* sas address is __be64, context ram format is __le64 */ + sas_addr =cpu_to_le64 (SAS_ADDR (dev->sas_addr)); + rnc->ssp.remote_sas_address_hi =upper_32_bits (sas_addr); + rnc->ssp.remote_sas_address_lo =lower_32_bits (sas_addr); rnc->ssp.nexus_loss_timer_enable = true; rnc->ssp.check_bit = false;
commit b3d5627b8574abbb6e550e1360bfdd65c2aa2e34
Author: Dan Williams <[email protected]> Date: Tue Apr 26 11:44:06 2011 -0700
isci: remove scic_sds_remote_device_get_port_index
Longer to type than the open-coded equivalent.
commit bbb7f7c517f67d3054c7105be1b4f6d2fadb3dec
Author: Dave Jiang <[email protected]> Date: Tue Apr 26 12:31:37 2011 -0700
isci: removing non-working ATAPI code
Removing not used / bit-rotten ATAPI code. This needs to go back and debugged at a later date.
commit 235bb93bd2647ec5ea5aeda88498886db495329f
Author: Dan Williams <[email protected]> Date: Tue Apr 26 13:19:53 2011 -0700
isci: allow fallback to option-rom if efi variable retrieval fails
If the scu efi driver is disabled but the option-rom is enabled (during an efi boot) allow the code to fallback to scanning legacy option-rom space for the parameters.
commit a89d81cb6bd3670ba55d32da7accaba34212bb08
Author: Dan Williams <[email protected]> Date: Tue Apr 26 13:31:02 2011 -0700
isci: defer pci_map_biosrom
The implementation was accepted into tip/x86/platform, but not 2.6.39. Add a dummy local definition to enable a self-contained a 2.6.39 merge.
commit ea04ff78fdbdc089e733569a93affcddd4aa2f83
Author: Dave Jiang <[email protected]> Date: Thu Apr 21 05:36:23 2011 +0000
isci: Remove excessive log noise with expander hot-unplug
We are logging excessive output when hot unplug from expander. Moving that to debug.
commit e1577d24adf3500e36f789526113caac3dd9e8e0
Author: Dan Williams <[email protected]> Date: Tue Apr 26 16:58:24 2011 -0700
isci: fix CONFIG_EFI=n compile error
When efi variable services are disabled the "efi" symbol is unavailable.
Todo: sync up with 2.6.39 infrastructure for drivers to consume efi variable services.
[1]: http://git.kernel.org/?p=linux/kernel/git/tip/linux-2.6-tip.git;a=shortlog;h=refs/heads/x86/platform
[2]: http://marc.info/?l=linux-scsi&m=130170650423175&w=2