- Jun 06, 2017
-
-
Luca Coelho authored
When we removed dev_cmd_headroom, the check for dev_cmd_ptr == NULL became unnecessary, since we just return dev_cmd_ptr anyway. Cleanup the function to avoid useless code. Signed-off-by:Luca Coelho <luciano.coelho@intel.com>
-
Luca Coelho authored
Bump the maximum accepted firmware API number for devices in the 8000, 9000 and A000 families. Signed-off-by: -
Mordechai Goodstein authored
The include in the deleted file are included in the fw-api.h file. Which caused a cycle include in the dependencies. Signed-off-by:
Mordechai Goodstein <mordechay.goodstein@intel.com> Signed-off-by:
-
Johannes Berg authored
The firmware no longer uses this command group, so remove it from the driver as well. Signed-off-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
-
Johannes Berg authored
Link from the TX_CMD enum value to the TX command structs. Signed-off-by:
-
Johannes Berg authored
Signed-off-by:
-
Johannes Berg authored
Document which structures are used with the command for the A000 hardware flavour. Signed-off-by:
-
Johannes Berg authored
In order to more clearly document which parts of this file are firmware API and which are something else, split the firmware API into a separate file to include here. Signed-off-by:
-
Johannes Berg authored
The kernel-doc here is on an enum, so don't tag it as struct but correctly as enum instead, preventing an error. Signed-off-by:
-
Johannes Berg authored
Clear the firmware running bit before flushing the FW (error) dump work, because otherwise debugfs isn't blocked (previous patch) and can cause a new work to be scheduled, which will then run after we actually shut down the device, wreaking havoc. Clearing it ensures that debugfs can't interfere anymore, and we can safely cancel or flush the work struct. Signed-off-by:
-
Johannes Berg authored
Convert ucode_loaded to a status bit called FIRMWARE_RUNNING. This will make it easier to clear this earlier, to avoid any spurious accesses while shutting down, for example through debugfs. Signed-off-by:
-
Johannes Berg authored
This will help refactor this later. Signed-off-by:
-
Johannes Berg authored
There's no need to refer to system_wq directly, use the provided wrapper schedule_delayed_work(). Made with the following spatch: @@ expression E,F; @@ -queue_delayed_work(system_wq, E, F); +schedule_delayed_work(E, F); Signed-off-by: -
Johannes Berg authored
Protect various debugfs files that need to communicate with the firmware from being used when the firmware isn't running. Some will just reject getting written to, while others that store some state will simply store it and not apply it immediately. Signed-off-by:
-
Liad Kaufman authored
Apart from the current list of PRPH that can't be collected in A000 HW, the rest of the debug dump data the driver collects is valid, so there is no need to disable collection only because of this. Disable PRPH collecting in A000 HW, and allow collecting the rest of the debug data. Signed-off-by:
Liad Kaufman <liad.kaufman@intel.com> Signed-off-by:
-
Liad Kaufman authored
FH in A000 HW are placed in a different location, and need to be read as prph, rather than direct. Support A000 dumping as well as legacy. Signed-off-by:
-
Sara Sharon authored
Add two new device families to differentiate them from 8000. Signed-off-by:
Sara Sharon <sara.sharon@intel.com> Signed-off-by:
-
Liad Kaufman authored
In case an assert happens on init flow, the current driver powers down the NIC, except if iwlmvm modparam init_dbg=1, and only on very specific flows. Extend this capability to cover most failure cases by keeping track of what init configurations have been completed. This way, we can allow NOT powering down the NIC, while making sure that when the driver is removed we don't try to free resources that haven't been allocated. (This can result in a kernel panic.) Signed-off-by:
-
Beni Lev authored
Due to NAN FW API change, add TLV in order to distiguish between the 2 API versions Signed-off-by:
Beni Lev <beni.lev@intel.com> Signed-off-by:
-
Sara Sharon authored
Use transport configuration to determine DBGC support instead of relying on device family. Signed-off-by:
-
Sara Sharon authored
We don't have any 8000 B-step right now, and there is no firmware loading code for them anyway. Further more, 9000 B-step devices will hit those code paths. Remove code that was introduced only for 8000 B-step. Signed-off-by:
-
Sara Sharon authored
NVM code is tightly coupled with 8000 family, while it really refers to extended NVM format introduced back then. Separate it to a configuration dependent boolean, and rename defines accordingly. Signed-off-by:
-
Sara Sharon authored
This is essentially the same code as gen1, except that it uses gen2 functions and SW checksum is not included. Signed-off-by:
-
Johannes Berg authored
Document the structures used in RX and link them to the command ID. Signed-off-by:
-
Sara Sharon authored
Enable offload assist for AMSDU when the AMSDU present flag is set. Fixes: a830baba9c2e ("iwlwifi: mvm: support new TX API") Signed-off-by:
-
- Jun 02, 2017
-
-
Sara Sharon authored
Now that we have 512 queues, add a wait for single TX queue to gen2. This replaces gen1 wait_tx_queues_empty, which was limited to 32 queues. Signed-off-by:
-
Johannes Berg authored
Fix many kernel-doc warnings. In one case, this required adding a new enum value to be able to document things properly. Signed-off-by:
-
Johannes Berg authored
This value is unused, and there's no reason we'd ever use it. Just remove it. Signed-off-by:
-
Johannes Berg authored
Using %enum instead of &enum (and in one case, %struct) results in the wrong parsing. Fix that so that if documentation is generated, the result is clickable links. Signed-off-by:
-
Johannes Berg authored
Signed-off-by:
-
Haim Dreyfuss authored
Integrated chip may have different HW and RF steps. Currently, the driver supports only different HW steps. Add logic to support different RF steps enables combining different HW and RF steps. Signed-off-by:
Haim Dreyfuss <haim.dreyfuss@intel.com> Signed-off-by:
-
Liad Kaufman authored
In gen2, page dumping needs to be done in the trans layer, as it is the one with access to the paging pointers. Signed-off-by:
-
Liad Kaufman authored
In gen2, page dumping should be done in transport layer, since the addresses needed for the paging mechanism reside there. Signed-off-by:
-
Johannes Berg authored
Add documentation to a lot of command IDs that links to the appropriate structure(s) used with those IDs. In one case, actually add and use a new struct for that purpose. Signed-off-by:
-
Johannes Berg authored
Some values should be in enums so documentation can refer to them, some values should be named for the same reason. Signed-off-by:
-
Johannes Berg authored
The comments/size of the different binding commands get lost in documentation, so introduce two different command structs that can be used there. Signed-off-by:
-
Sara Sharon authored
Three configurations will share device ID 2720, and will be differentiated by RF ID. Signed-off-by:
-
Johannes Berg authored
Link various fields to the documentation of the enums that define their values. Signed-off-by:
-
Johannes Berg authored
This mostly fixes missing tags/struct names, but also some other things. Lots of issues remain though. Signed-off-by:
-
Sara Sharon authored
The function flushed only agg queue and no more traffic is queued on it. However, it waits for all queues to empty, which is not necessary and may take more time. Signed-off-by:
-
