- 29 Sep, 2015 40 commits
-
-
Dan Carpenter authored
There were several bugs here. 1) The done label was in the wrong place so we didn't copy any information out when there was no command given. 2) We were using PAGE_SIZE as the size of the buffer instead of "PAGE_SIZE - pos". 3) snprintf() returns the number of characters that would have been printed if there were enough space. If there was not enough space (and we had fixed the memory corruption bug #2) then it would result in an information leak when we do simple_read_from_buffer(). I've changed it to use scnprintf() instead. I also removed the initialization at the start of the function, because I thought it made the code a little more clear. Fixes: 5e6e3a92 ('wireless: mwifiex: initial commit for Marvell mwifiex driver') Signed-off-by:Dan Carpenter <dan.carpenter@oracle.com> Acked-by:
Amitkumar Karwar <akarwar@marvell.com> Signed-off-by:
Kalle Valo <kvalo@codeaurora.org>
-
Ondrej Zary authored
Add calls to netif_carrier_on and netif_carrier_off Signed-off-by:
Ondrej Zary <linux@rainbow-software.org> Signed-off-by:
-
Ondrej Zary authored
IW_AUTH_ALG_OPEN_SYSTEM is ambiguous in set_auth for WEP as wpa_supplicant uses it for both no encryption and WEP open system. Cache the last mode set (only of these two) and use it here. This allows wpa_supplicant to work with unencrypted APs. Signed-off-by:
-
Rafał Miłecki authored
Signed-off-by:
Rafał Miłecki <zajec5@gmail.com> Signed-off-by:
-
Rafał Miłecki authored
Signed-off-by:
-
Rafał Miłecki authored
It's used locally only. Signed-off-by:
-
Rafał Miłecki authored
It isn't used anywhere out of ssb code and we don't (plan to) build pcihost_wrapper.c as a separated module. Signed-off-by:
-
Hauke Mehrtens authored
of_default_bus_match_table was not exported earlier, so it could only be accessed by code compiled into the kernel. A new function of_platform_default_populate() was added which uses of_default_bus_match_table and this function is also exported. This way it is possible to create a bus with the content of of_default_bus_match_table and we can remove the hacks from bcma. Signed-off-by:
Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by:
-
Hauke Mehrtens authored
brcmfmac uses atomic_or() and other atomic_* functions, but does not include linux/atomic.h. This file gets included by some other header file so this normally does not cause problems. Signed-off-by:
Arend van Spriel <arend@broadcom.com> Signed-off-by:
-
Hante Meuleman authored
TDLS events are mapped back to primary interface but when p2p is in use then this fails because the check was incorrect by checking bsscfg number. Which can be different when a p2p device has been created. Reviewed-by:
Pieter-Paul Giesberts <pieterpg@broadcom.com> Signed-off-by:
Hante Meuleman <meuleman@broadcom.com> Signed-off-by:
-
Hante Meuleman authored
There is a workaround needed for p2p device setup which breaks tdls functionality. This patch fixes that by properly signalling fweh that p2p device setup is ongoing. Reviewed-by:
-
Hante Meuleman authored
This patch adds support for the BCM4365 and BCM4366 11ac Wave2 PCIE devices. Reviewed-by:
-
Hante Meuleman authored
When p2p device interface gets deleted by deinitialising discovery it will result in an event which removes the interface, but that is also done by delete p2p interface code. This results in race condition which sometimes results in lockup/crash. With this patch the delete device interface will wait for the event (with timeout) removing the possible race condition. Also on the stop device call from cfg80211 the deinitialisation of the discovery device should be avoided as it can result in a similar situation. Reviewed-by:
-
Hante Meuleman authored
For debugging purpose it is very handy to be able to disable features. It has happened a few times that new features turned out not always being properly detected for all devices/firmwares. Making it possible to disable the feature with a module parameter will make testing/debugging easier. Reviewed-by:
Franky (Zhenhui) Lin <frankyl@broadcom.com> Reviewed-by:
-
Hante Meuleman authored
Some devices with older firmwares are reporting new p2p device interface with the wrong type. Accept this type to get p2p working for these devices. Reviewed-by:
-
Hante Meuleman authored
In some situations it is possible that vif has been removed while cfg80211 invokes the p2p_stop_device handler. This will result in crash. Reviewed-by:
-
Hante Meuleman authored
When a p2p device gets deleted, the memory for the vif is not being released. This is solved by reorganizing the cleanup path and properly freeing the memory. Reviewed-by:
-
Hante Meuleman authored
Different wpa_supplicants have different behavior and expectations regarding the change_virtual_intf behavior. This patch implements a workaround for the different versions and possible brcmfmac configuration. Reviewed-by:
-
Hante Meuleman authored
p2pon module param is currently under define BRCMDBG. Though it is a needed option for older versions of the wpa_supplicant which do not support the P2P_DEVICE interface. Reviewed-by:
-
Hante Meuleman authored
The hash index stored in the flowrings is of type u16 but gets stored in u8. This can result in incorrect indexing and possibly result in crashes. This patch fixes the type. Reviewed-by:
-
Hante Meuleman authored
When the p2pon module param is used then p2p attach will initialize p2p device iface in the firmware, but it is doing that by checking data. It is cleaner to pass the p2pon information to p2p by API. This information is also needed for other patch. Reviewed-by:
-
Hante Meuleman authored
The if2bss allows for translation of ifidx to bssidx which has a 1:n relation. Therefor only the first (primary) netdev should be assigned in this array. This fixes the p2pon=1 module param usage. Reviewed-by:
-
Hante Meuleman authored
Implementation of tx-power (get and set) related functions are still assuming mW interface. This is wrong as functions use dbm (or mbm) nowadays. As a result a tx power configuration could result in wrong power configuration. Reviewed-by:
-
Hante Meuleman authored
This patch adds support fo the BRCM4350 2x2 11ac PCIE device. Reviewed-by:
-
Hante Meuleman authored
In some exception situations the ifp->vif was not properly assigned which could result in crash. Reviewed-by:
-
Zhaoyang Liu authored
This patch do the work to TX data with specific USB endpoint. At the same time, update data_sent flag according to multi port status. And is_port_ready API is added for BSS interface to check if current used usb data endpoint is available or not. Signed-off-by:
Zhaoyang Liu <liuzy@marvell.com> Signed-off-by:
Cathy Luo <cluo@marvell.com> Signed-off-by:
-
Zhaoyang Liu authored
This patch add support for usb multi endpoints resync. Once multi channel event is received from firmware, update usb_mc_setp flag to block TX data until setup is over. And available data endpoint will be attached to BSS interface. Signed-off-by:
-
Zhaoyang Liu authored
This patch add support for USB interface to TX data with different endpoint. And previous TX information are saved in new designed structure. Signed-off-by:
-
Zhaoyang Liu authored
This patch add multi channel event process for USB multi tx data endpoints. Driver receives firmware multi channel event only in case that new connection is setup or interface is disconnect. Different BSS interface need update used usb endpoint. Signed-off-by:
-
Zhaoyang Liu authored
This patch move data_sent flag update to usb.c file. >From now all data_sent update cases only happened in specific file: sdio.c, usb.c, pcie.c. Outside ot these files, it is only allowed to check the value of data_ent. Signed-off-by:
-
Zhaoyang Liu authored
This patch fix missing tx data_sent flag update for usb interface. Except USB interface, data_sent flag has been updated in specific file such as sdio.c and pcie.c. So only USB interface type need check when TX data completed. Signed-off-by:
-
Amitkumar Karwar authored
We have preallocated buffer pool for Tx data. During high data traffic, all buffers are submitted to USB and driver needs to wait until one of the buffers get available for next Tx packet. "data: -ENOSR is returned" errors is expected in this case. Let's lower the priority of this message. Signed-off-by:
-
Zhaoyang Liu authored
This patch fixes observed issues while updating counter for number of paused data packets in wmm queue when trying to delete packet or delete station entry. Signed-off-by:
-
Amitkumar Karwar authored
Preallocated PCIe buffer is being reused for all PCIe interface events. Physical address of the buffer is shared with firmware so that it can perform DMA on it. As event length is specified in the header, there should not be a problem if the buffer gets overwritten. We will save some cycles by avoiding memset everytime while submitting the buffer to firmware. Fixes: 2728cecd(mwifiex: corrections in PCIe event skb) Signed-off-by:
-
Aniket Nagarnaik authored
This NULL pointer dereference is observed during suspend resume stress test. All pending commands are cancelled when system goes into suspend state. There a corner case in which host may receive response for last scan command after this and try to trigger extra active scan for hidden SSIDs. The issue is fixed by adding a NULL check to skip that extra scan. Fixes: 2375fa2b (mwifiex: fix unable to connect hidden SSID..) Cc: <stable@vger.kernel.org> [v4.2+] Signed-off-by:
Aniket Nagarnaik <aniketn@marvell.com> Signed-off-by:
-
Aniket Nagarnaik authored
Current TDLS implementation always includes ht/vht information in tdls setup confirm frame which causes teardown by legacy peer station after TDLS handshake. We will inclue ht/vht capabilities in tdls setup confirm frame only if peer station supports it to fix this problem. Signed-off-by:
-
Xinming Hu authored
cfg80211_inform_bss would return null in some cases, such as memory allocation failure. This patch adds sanity check for this case, to avoid possible issues when above corner case is hit. Signed-off-by:
Xinming Hu <huxm@marvell.com> Signed-off-by:
-
Xinming Hu authored
64k Tx and Rx buffers are allocated during driver initialization for SDIO level data aggregations. When host is under memory pressure situation, kzalloc() request for 64k may fail. We will try allocating 32k buffers and disable our rx single port aggreagation feature in this situation. If the allocation still fails, we will disable our sdio multport aggregation feature as well. In this way, we will transmit and receive packets one by one, thus reduce the demand for big memory. Signed-off-by:
-
Xinming Hu authored
This patch includes below changes 1) Check if multiport aggregation feature flag while doing rx length validation check. 2) Reset buffer size variables when buffers are freed. Signed-off-by:
-
Zhaoyang Liu authored
This patch fixes the issue of delay time in A-MSDU tx packet. In generated new A-MSDU packet, the time stamp is initialized to zero. Choose the time of first MSDU packet as aggregated packet's time. Signed-off-by:
-
