- 03 Oct, 2011 40 commits
-
-
Roland Vossen authored
The fullmac dongle communicates with the driver using certain codes. These codes were named 'ioctls' in the code, but this term is confusing since it is used in kernel<->driver context. The term 'ioctl' has been replaced with 'dcmd' for 'dongle command'. Reported-by:
Johannes Berg <johannes@sipsolutions.net> Reviewed-by:
Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by:
Arend van Spriel <arend@broadcom.com> Reviewed-by:
Franky (Zhenhui) Lin <frankyl@broadcom.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@suse.de>
-
Roland Vossen authored
Parameter was always called with the value 'true'. Reviewed-by:
-
Roland Vossen authored
Function is not called anymore. Reviewed-by:
-
Roland Vossen authored
No element in array locale_bn[] has the 'no mimo' flag bit set. Reviewed-by:
-
Roland Vossen authored
Usage of spinlock depended on module parameters that are going to be removed in a subsequent patch. Parameter wlc->threads_only is always 'true'. Reviewed-by:
-
Arend van Spriel authored
Some functions in the source file had a lot of indentation levels forcing statements to be spread over multiple lines impairing the readability of the code. Reported-by:
Joe Perches <joe@perches.com> Reviewed-by:
Roland Vossen <rvossen@broadcom.com> Signed-off-by:
-
Roland Vossen authored
By introducing 1 new function. Reviewed-by:
-
Roland Vossen authored
By introducing 1 new function. Reviewed-by:
-
Roland Vossen authored
By introducing two new functions. Reviewed-by:
-
Arend van Spriel authored
The driver only supports one network interface to be associated with it. Only for the first call it does actually something. For subsequent calls it does return -ENODEV error code. Reported-by:
-
Roland Vossen authored
Code cleanup. Makes it clear when an address is mapped IO space. Reviewed-by:
-
Roland Vossen authored
Code cleanup. Makes it clear when an address is mapped IO space. Reviewed-by:
-
Roland Vossen authored
Code cleanup. Makes it clear when an address is mapped IO space. Reviewed-by:
-
Roland Vossen authored
By declaring certain variables static. One unused array induced a warning, array was deleted. Reviewed-by:
-
Roland Vossen authored
So it is clear to the reader what memory is IO mapped Reviewed-by:
-
Roland Vossen authored
Code cleanup. Code was unnecessarily complex. Reviewed-by:
-
Roland Vossen authored
A remnant of the past. Reviewed-by:
-
Roland Vossen authored
Reviewed-by:
-
Roland Vossen authored
Global variables are undesirable unless they are read only. Reported-by:
-
Roland Vossen authored
Code cleanup resulting in less sparse warnings. Reported-by:
-
Arend van Spriel authored
All information needed for this function is available in the d11rxhdr structure. This is the last place where the brcms_d11rxhdr structure so it can be removed. Reviewed-by:
-
Arend van Spriel authored
The rssi computation was done upon getting the frame from the queue, but the value is needed when filling in the receive status data for mac80211. The call to wlc_phy_rssi_compute() has been deferred. Reviewed-by:
-
Arend van Spriel authored
Structures interfacing with the device have a specific endianess and structures exchanged between host and device have been annotated so sparse checking can be done. The Makefile has been modified to add the __CHECK_ENDIAN__ flag. Reported-by:
-
Arend van Spriel authored
The received frame from the device is in little-endian order and converted in the receive path. However, the phy code was doing the rssi calculation with packet in LE order. This has been changed to do the le_to_cpu conversion before doing the rssi computation and further receive processing. Reviewed-by:
-
Arend van Spriel authored
The function wlc_phy_rssi_compute() now uses d11rxhdr struct as purely input parameter and returns the computed rssi value. Reviewed-by:
-
Arend van Spriel authored
The function wlc_phy_rssi_compute_nphy() was called with pointer to brcms_d11rxhdr structure in which it filled in the received power per antenna. However, these are not used further in the driver so it only needs the d11rxhdr structure as input for rssi calculation. Reviewed-by:
-
Roland Vossen authored
Placed variable on the stack instead and deleted unused functions. Softmac was tested to function properly with multiple adapters in one system. Reported-by:
-
Roland Vossen authored
Reported-by:
-
Roland Vossen authored
Global variables are undesirable unless they are read only. Removed by instead using an already defined Station Control Block variable in a per-device structure. Reported-by:
-
Roland Vossen authored
Global variables are undesirable unless they are read only. Variables are now maintained in a device specific structure. Reported-by:
-
Arend van Spriel authored
Several callback were implemented without executing any further function calls into the driver. Review feedback indicated that these could be removed. Reported-by:
-
Arend van Spriel authored
When the device dies the driver could extract cpu registers on the device to analyze the trap handling on the dongle. As the firmware with this driver is stable this code does not belong in the brcmfmac driver. Reported-by:
-
Arend van Spriel authored
For chip initialisation of the wme parameters a table is used, but it was not marked as constant. Reviewed-by:
-
Arend van Spriel authored
The function is not used in the driver and has been removed. Reviewed-by:
-
Arend van Spriel authored
In the transmit path the field seq_ctrl is filled in, but the fragment number was not properly determined. Reviewed-by:
-
Arend van Spriel authored
The field ht_cap was typed ieee80211_ht_cap from ieee80211.h. This contained little endian annotated field cap_info resulting in sparse endian warnings. It turned out the driver was setting the field, but it was actually never used. Therefore it has been removed. Reported-by:
-
Roland Vossen authored
WAR16165 is only used on older PCI chips, the driver does not support these chips. Reviewed-by:
-
Roland Vossen authored
Moved global vars into a per-device structure. Reviewed-by:
-
Roland Vossen authored
Reviewed-by:
-
Roland Vossen authored
After initialization ('attach'), this struct member always pointed at the same memory as wlc->bsscfg. Reviewed-by:
-
