Jubin John [Fri, 18 Dec 2015 00:24:14 +0000 (19:24 -0500)]
staging/rdma/hfi1: Fix module parameter spelling
Fix the spelling of user_credit_return_threshold, it was incorrectly
spelled as user_credit_return_theshold causing two module parameters,
one with typo, to be shown in modinfo
Reviewed-by: Ira Weiny <ira.weiny@intel.com> Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging/rdma/hfi1: Change num_rcv_contexts to num_user_contexts and its meaning
num_rcv_contexts sets the number of user contexts, both receive and send.
Renaming it to num_user_contexts makes sense to reflect its true meaning.
When num_rcv_contexts is 0, the default behavior is the number of CPU
cores instead of 0 contexts. This commit changes the variable
num_rcv_contexts to num_user_contexts, and it also makes any negative
value for this variable default to the number of CPU cores, so if
num_user_contexts is set >= 0, the value will number of contexts.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Sebastian Sanchez <sebastian.sanchez@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging/rdma/hfi1: Fix for module parameter hdrq_entsize when it's 0
If driver is loaded with parameter hdrq_entsize=0, then
there's a NULL dereference when the driver gets unloaded.
This causes a kernel Oops and prevents the module from
being unloaded. This patch fixes this issue by making sure
-EINVAL gets returned when hdrq_entsize=0.
Easwar Hariharan [Thu, 10 Dec 2015 16:13:38 +0000 (11:13 -0500)]
staging/rdma/hfi1: Fix a possible null pointer dereference
A code inspection pointed out that kmalloc_array may return NULL and
memset doesn't check the input pointer for NULL, resulting in a possible
NULL dereference. This patch fixes this.
Erik E. Kahn [Thu, 10 Dec 2015 14:59:40 +0000 (09:59 -0500)]
staging/rdma/hfi1: HFI now sends OPA Traps instead of IBTA
send_trap() was still using old ib_smp instead of opa_smp
for formatting and sending traps.
Reviewed-by: Arthur Kepner <arthur.kepner@intel.com> Reviewed-by: Ira Weiny <ira.weiny@intel.com> Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Erik E. Kahn <erik.kahn@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Yash Shah [Tue, 8 Dec 2015 10:24:17 +0000 (10:24 +0000)]
Staging: rdma:Delete unnecessary NULL check before calling function "kmem_cache_destroy"
The kmem_cache_destroy() function tests whether its argument is NULL
and then returns immediately.
Thus the NULL check before calling this function is not needed.
This issue was detected by using the Coccinelle software.
Signed-off-by: Yash Shah <yshah1@visteon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging/rdma/hfi1: fix pio progress routine race with allocator
The allocation code assumes that the shadow ring cannot
be overrun because the credits will limit the allocation.
Unfortuately, the progress mechanism in sc_release_update() updates
the free count prior to processing the shadow ring, allowing the
shadow ring to be overrun by an allocation.
Reviewed-by: Mark Debbage <mark.debbage@intel.com> Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Mitko Haralanov [Tue, 8 Dec 2015 22:10:12 +0000 (17:10 -0500)]
staging/rdma/hfi1: Detect SDMA transmission error early
It is possible for an SDMA transmission error to happen
during the processing of an user SDMA transfer. In that
case it is better to detect it early and abort any further
attempts to send more packets.
Mitko Haralanov [Tue, 8 Dec 2015 22:10:13 +0000 (17:10 -0500)]
staging/rdma/hfi1: Add page lock limit check for SDMA requests
The driver pins pages on behalf of user processes in two
separate instances - when the process has submitted a
SDMA transfer and when the process programs an expected
receive buffer.
When pinning pages, the driver is required to observe the
locked page limit set by the system administrator and refuse
to lock more pages than allowed. Such a check was done for
expected receives but was missing from the SDMA transfer
code path.
This commit adds the missing check for SDMA transfers. As of
this commit, user SDMA or expected receive requests will be
rejected if the number of pages required to be pinned will
exceed the set limit.
Due to the fact that the driver needs to take the MM semaphore
in order to update the locked page count (which can sleep), this
cannot be done by the callback function as it [the callback] is
executed in interrupt context. Therefore, it is necessary to put
all the completed SDMA tx requests onto a separate list (txcmp) and
offload the actual clean-up and unpinning work to a workqueue.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com> Reviewed-by: Ira Weiny <ira.weiny@intel.com> Signed-off-by: Mitko Haralanov <mitko.haralanov@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Mitko Haralanov [Tue, 8 Dec 2015 22:10:09 +0000 (17:10 -0500)]
staging/rdma/hfi1: Convert to use get_user_pages_fast
Convert hfi1_get_user_pages() to use get_user_pages_fast(),
which is much fatster. The mm semaphore is still taken to
update the pinned page count but is for a much shorter
amount of time.
There is no need to cleck if the packet queue is allocated
when cleaning up a user context. The hfi1_user_sdma_free_queues()
function already does all the required checks.
Andrea Lowe [Tue, 1 Dec 2015 20:38:26 +0000 (15:38 -0500)]
staging/rdma/hfi1: Adding counter resolutions for DataPortCounters
Changing the 32-bit reserved field in opa_port_data_counters_msg
to the new 'resolution' field. PMA will use resolutions to right-
shift values for LocalLinkIntegrity and LinkErrorRecovery when
computing the ErrorCounterSummary for a DataPortCounters request.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Andrea Lowe <andrea.l.lowe@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Harish Chegondi [Tue, 1 Dec 2015 20:38:24 +0000 (15:38 -0500)]
staging/rdma/hfi1: Consider VL15 MTU also when calculating the maximum VL MTU
Currently, only MTUs of VLs 0-7 are checked when calculating the maximum VL
MTU which is used to set the port MTU capability in DCC_CFG_PORT_CONFIG CSR
This can cause a port MTU capability to be set to 0 if MTUs of VLs 0-7 is 0
This would affect the VL15 traffic.
Reviewed-by: Dean Luick <dean.luick@intel.com> Reviewed-by: Arthur Kepner <arthur.kepner@intel.com> Signed-off-by: Harish Chegondi <harish.chegondi@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ira Weiny [Sat, 21 Nov 2015 00:43:47 +0000 (19:43 -0500)]
staging/rdma/hfi1: Eliminate WARN_ON when VL is invalid
sdma_select_engine_vl only needs to protect itself from an invalid VL.
Something higher up the stack should be warning the user when they try
to use an SL which maps to an invalid VL.
Reviewed-by: Dean Luick <dean.luick@intel.com> Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Reviewed-by: Kaike Wan <kaike.wan@intel.com> Signed-off-by: Ira Weiny <iweiny@gmail.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Jubin John [Fri, 20 Nov 2015 23:13:08 +0000 (18:13 -0500)]
staging/rdma/hfi1: Fix error in hfi1 driver build
hfi1 driver build fails with the following error:
In function ‘handle_receive_interrupt’:
error: implicit declaration of function ‘skip_rcv_packet’ [-Werror=implicit-function-declaration]
last = skip_rcv_packet(&packet, thread);
^
This is due to the inclusion of the skip_rcv_packet() in the
CONFIG_PRESCAN_RXQ ifdef block. This function is independent of
CONFIG_PRESCAN_RXQ and should be outside this block.
Fixes: 82c2611daaf0 ("staging/rdma/hfi1: Handle packets with invalid RHF on context 0") Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Jubin John [Tue, 17 Nov 2015 02:59:27 +0000 (21:59 -0500)]
staging/rdma/hfi1: Add space between concatenated string elements
Space between concantenated string elements is more human
readable and fixes the checkpatch issue:
CHECK: Concatenated strings should use spaces between elements
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ira Weiny [Tue, 17 Nov 2015 02:59:26 +0000 (21:59 -0500)]
staging/rdma/hfi1: Remove rcv bubbles code
Rcv bubbles were improperly calculated for HFIs, fix that here.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Reviewed-by: Arthur Kepner <arthur.kepner@intel.com> Signed-off-by: Ira Weiny <iweiny@gmail.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dean Luick [Tue, 17 Nov 2015 02:59:24 +0000 (21:59 -0500)]
staging/rdma/hfi1: Fix downgrade race
A link downgrade can race with link up. Avoid the race
in two ways. First, by having the downgrade application logic
take the link state mutex for all of its checking. Second, by
waiting for the link to move out of the going up state.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com> Signed-off-by: Dean Luick <dean.luick@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dmitry Torokhov [Tue, 15 Dec 2015 01:34:08 +0000 (17:34 -0800)]
android: unconditionally remove callbacks in sync_fence_free()
Using fence->status to determine whether or not there are callbacks
remaining on the sync_fence is racy since fence->status may have been
decremented to 0 on another CPU before fence_check_cb_func() has
completed. By unconditionally calling fence_remove_callback() for each
fence in the sync_fence, we guarantee that each callback has either
completed (since fence_remove_callback() grabs the fence lock) or been
removed.
staging/android/sync: Support sync points created from dma-fences
Debug output assumes all sync points are built on top of Android sync points
and when we start creating them from dma-fences will NULL ptr deref unless
taught about this.
Glen Lee [Mon, 21 Dec 2015 05:18:46 +0000 (14:18 +0900)]
staging: wilc1000: remove used functions
This patch remove unused functions add_handler_in_list and
remove_handler_in_list, and it's related global variable wfidrv_list and codes.
label fail_timer_2 and it's codes are removed since label is not used anymore.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:45 +0000 (14:18 +0900)]
staging: wilc1000: change join_req_drv type and it's name
To use wilc_get_vif_idx instead of the last get_id_from_handler, join_req_drv
needs to be changed it's type with wilc_vif and name as well.
As a result, get_id_from_handler is not used anymore, so remove it.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:44 +0000 (14:18 +0900)]
staging: wilc1000: wilc_set_wfi_drv_handler: pass vif index
Pass index of vif instead of hif_drv. wilc_get_vif_idx is used to get correct
index of vif.
In the handler function handle_set_wfi_drv_handler, use vif instead of hif_drv,
and use hif_drv_handler->handler instead of hif_drv when deinitialize wilc
device.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:43 +0000 (14:18 +0900)]
staging: wilc1000: use vif index to communicate with wilc device
We now have vif index in all functions related with host interface thread.
wilc_get_vif_idx and wilc_get_vif_from_idx are added to get id and vif
respectively.
Relace get_id_from_handler with wilc_get_vif_idx and get_handler_from_id with
wilc_get_vif_from_idx. Remove unused function get_handler_from_id as well.
We get vif where wilc_get_vif_from_idx is called, so pass vif to msg.vif.
There are two get_id_from_handler left. They will be removed in later patch.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:41 +0000 (14:18 +0900)]
staging: wilc1000: take vif instead of drv in hostIFthread
In the first patch, we sent vif to hostIFthread. we can use vif instead of
drv in the all functions which handle the commands from cfg operations.
Change first argument host_if_drv with wilc_vif and use hif_drv of wilc_vif.
Pass vif to the functions as well.
In case of timer callback functions, set vif to the data and use vif instead
of hif_drv.
Lastly, initialize u32RcvdAssocRespInfoLen since changing hif_drv with vif
causes one uninitialied build warning.
Now we have vif that currently being used so we can use interface index of
wilc_vif to send to wilc device.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:40 +0000 (14:18 +0900)]
staging: wilc1000: remove argument hif_drv
In previous patch we add new argument vif which has hif_drv in it's member.
Therefore, no need to pass hif_drv in those functions. Remove argument
struct host_if_drv and use hif_drv of vif.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:39 +0000 (14:18 +0900)]
staging: wilc1000: pass vif to hostIFthread
We will pass vif, which is currently being used as net_device, instead of
hif_dev. This is the first step to use index of vif to pass to the driver.
Add new argument vif to all the functions that send message to hostIFthread and
set vif to msg.vif. As a result, hostIfthread will get vif.
In later patch, we will remove drv of host_if_msg and use vif instead of it.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:38 +0000 (14:18 +0900)]
staging: wilc1000: remove duplicate netdev
There are two net_device pointer which is the same because two structures
are merged into wilc_vif in previous patch. Remove wilc_netdev and change
with ndev.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:37 +0000 (14:18 +0900)]
staging: wilc1000: change vif to pointer to refence real private data
vif of struct has it's own memory which is not necessary because we have
allocated vif from netdev_priv.
Change vif to pointer type and assign vif which is netdev private data.
Change it's operator on related codes as well.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:36 +0000 (14:18 +0900)]
staging: wilc1000: move perInterface_wlan_t to wilc_vif
perInterface_wlan_t and wilc_vif are all about interface control informations.
We will combine those two structures and maintain as one network interface
control information.
Move all the members of perInterface_wlan_t to wilc_vif and remove the
structure. Rename perInterace_wlan_t to wilc_vif and rename variable name nic
to vif which is proper name for it.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:33 +0000 (14:18 +0900)]
staging: wilc1000: linux_wlan_spi.c: move all the codes to wilc_spi.c
This patch moves all the codes in linux_wlan_spi.c to wilc_spi.c to make
one spi module. Make wilc_spi_tx, wilc_spi_rx and wilc_spi_tx_rx static
functions. Remove function declaration in linux_wlan_spi.h, which is unnedded
now. No modification has been made inside the codes.
linux_wlan_spi.[ch] will be remove in the next patch.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:31 +0000 (14:18 +0900)]
staging: wilc1000: rename spi function names
There are several similar function names, such as wilc_spi_write and
_wilc_spi_write. It is likely to be confused after merging linux_wlan_spi.c and
wilc_spi.c, so rename following functions properly.
Rename wilc_spi_write to wilc_spi_tx, wilc_spi_read to wilc_spi_rx,
wilc_spi_write_read to wilc_spi_tx_rx, _wilc_spi_write to wilc_spi_write,
_wilc_spi_read to wilc_spi_read.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Glen Lee [Mon, 21 Dec 2015 05:18:29 +0000 (14:18 +0900)]
staging: wilc1000: linux_wlan_sdio.c: move all the codes to wilc_sdio.c
To Combine linux_wlan_sdio.c and wilc_sdio.c as one file, move all the codes
in linux_wlan_sdio.c to wilc_sdio.c, and make functions static only.
No Modification has not been made except static, just moved them.
Function declaration in linux_wlan_sdio.h is needless, so just remove them.
linux_wlan_sdio.[ch] will be deleted in the next patch.
Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes the error reported by smatch.
- Handle_ListenStateExpired() error: potential null dereference 'wid.val'
If kmalloc failed, referenced to a NULL pointer.
Signed-off-by: Leo Kim <leo.kim@atmel.com> Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Leo Kim [Mon, 21 Dec 2015 05:18:23 +0000 (14:18 +0900)]
staging: wilc1000: replace explicit NULL comparisons with !
This patch replace explicit NULL comparison with ! operator to simplify code.
Reported by checkpatch.pl for Comparison to NULL could be written !XXX" or "XXX".
Signed-off-by: Leo Kim <leo.kim@atmel.com> Signed-off-by: Glen Lee <glen.lee@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / mirror_ubuntu-eoan-kernel.git / log