- 13 May, 2020 40 commits
-
-
Houssem KADI authored
Missing empty line after variable declaration Signed-off-by:
Houssem KADI <kadi.houssem.eddine@gmail.com> Link: https://lore.kernel.org/r/20200509180849.GA30426@houssem-MS-7808Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Malcolm Priestley authored
Now mac89211 is doing frame timing in rxtx these vendor adjustments need to be removed. Signed-off-by:
Malcolm Priestley <tvboxspy@gmail.com> Link: https://lore.kernel.org/r/034e445c-b245-52c4-c855-431b9783bcff@gmail.comSigned-off-by:
-
Malcolm Priestley authored
rsvtime is the time needed in firmware to process the received frame time in firmware so they can be the same as vnt_get_rts_duration or vnt_get_cts_duration where appropriate. The rts_rrv_time are now all the same timing in vnt_rxtx_rts. So vnt_get_rtscts_rsvtime_le and and vnt_get_frame_time are no longer required. Signed-off-by:
-
Malcolm Priestley authored
split vnt_get_rtscts_duration_le into vnt_get_rts_duration and vnt_get_cts_duration. The duration's are all the same in vnt_rxtx_rts_g_head. Signed-off-by:
-
Malcolm Priestley authored
use the mac80211 ieee80211_ctstoself_duration for CTS to self frames. Signed-off-by:
-
Malcolm Priestley authored
ieee80211_generic_frame_duration is the mac80211 equivalent to vnt_get_rsvtime use this to get our frame time. There is a change where there is rrv_time_a and rrv_time_b the frame duration is always the same so both are equal. Signed-off-by:
-
Malcolm Priestley authored
use the mac80211 ieee80211_rts_duration for RTS frames. Signed-off-by:
-
Colin Ian King authored
Currently channel is being sanity checked after it has been used as an index into some arrays. Fix this by moving the sanity check of channel before the arrays are indexed with it. Addresses-Coverity: ("Negative array index read") Fixes: 59ed0480 ("Staging: most: replace pr_*() functions by dev_*()") Signed-off-by:Colin Ian King <colin.king@canonical.com> Link: https://lore.kernel.org/r/20200507150652.52238-1-colin.king@canonical.comSigned-off-by:
-
Christian Gromm authored
This patch adds the implementation of the PM functions resume and suspend. Signed-off-by:
Christian Gromm <christian.gromm@microchip.com> Link: https://lore.kernel.org/r/1588680892-9413-1-git-send-email-christian.gromm@microchip.comSigned-off-by:
-
Jérôme Pouiller authored
Update the TODO list associated to the wfx driver with the last progresses. Signed-off-by:
Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20200512150414.267198-18-Jerome.Pouiller@silabs.comSigned-off-by:
-
Jérôme Pouiller authored
The field 'channel_number' from the structs hif_ind_rx and hif_req_start is a __le32. Sparse complains this field is not always correctly accessed: drivers/staging/wfx/data_rx.c:95:55: warning: incorrect type in argument 1 (different base types) drivers/staging/wfx/data_rx.c:95:55: expected int chan drivers/staging/wfx/data_rx.c:95:55: got restricted __le16 const [usertype] channel_number However, the value of channel_number cannot be greater than 14 (this device only support 2.4Ghz band). So, we only have to access to the least significant byte. It is finally easier to declare it as an array of bytes and only access to the first one. Signed-off-by: -
Jérôme Pouiller authored
The field 'num_tx_confs' from the struct hif_cnf_multi_transmit is a __le32. Sparse complains this field is not always correctly accessed: drivers/staging/wfx/hif_rx.c:82:9: warning: restricted __le32 degrades to integer drivers/staging/wfx/hif_rx.c:87:29: warning: restricted __le32 degrades to integer However, the value of num_tx_confs cannot be greater than 15. So, we only have to access to the least significant byte. It is finally easier to declare it as an array of bytes and only access to the first one. Signed-off-by: -
Jérôme Pouiller authored
The field 'status' appears in most of structs returned by the hardware. This field is encoded as little endian. Sparse complains this field is not always correctly accessed: drivers/staging/wfx/data_rx.c:53:16: warning: restricted __le32 degrades to integer drivers/staging/wfx/data_rx.c:84:16: warning: restricted __le32 degrades to integer drivers/staging/wfx/data_tx.c:526:24: warning: restricted __le32 degrades to integer drivers/staging/wfx/data_tx.c:569:23: warning: restricted __le32 degrades to integer drivers/staging/wfx/hif_rx.c:128:33: warning: restricted __le32 degrades to integer drivers/staging/wfx/./traces.h:401:1: warning: restricted __le32 degrades to integer drivers/staging/wfx/./traces.h:401:1: warning: restricted __le32 degrades to integer In most of cases, this field is only compared with HIF_STATUS values. Finally, it is more convenient to solve the problem by defining the HIF_STATUS values directly in little endian. It is also the right time to make some clean up in the HIF_STATUS names. Signed-off-by: -
Jérôme Pouiller authored
Sparse complains about the accesses to the field 'len' from struct hif_msg: drivers/staging/wfx/bh.c:88:32: warning: restricted __le16 degrades to integer drivers/staging/wfx/bh.c:88:32: warning: restricted __le16 degrades to integer drivers/staging/wfx/bh.c:93:32: warning: restricted __le16 degrades to integer drivers/staging/wfx/bh.c:93:32: warning: cast to restricted __le16 drivers/staging/wfx/bh.c:93:32: warning: restricted __le16 degrades to integer drivers/staging/wfx/bh.c:121:25: warning: incorrect type in argument 2 (different base types) drivers/staging/wfx/bh.c:121:25: expected unsigned int len drivers/staging/wfx/bh.c:121:25: got restricted __le16 [usertype] len drivers/staging/wfx/hif_rx.c:27:22: warning: restricted __le16 degrades to integer drivers/staging/wfx/hif_rx.c:347:39: warning: incorrect type in argument 7 (different base types) drivers/staging/wfx/hif_rx.c:347:39: expected unsigned int [usertype] len drivers/staging/wfx/hif_rx.c:347:39: got restricted __le16 const [usertype] len drivers/staging/wfx/hif_rx.c:365:39: warning: incorrect type in argument 7 (different base types) drivers/staging/wfx/hif_rx.c:365:39: expected unsigned int [usertype] len drivers/staging/wfx/hif_rx.c:365:39: got restricted __le16 const [usertype] len drivers/staging/wfx/./traces.h:195:1: warning: incorrect type in assignment (different base types) drivers/staging/wfx/./traces.h:195:1: expected int msg_len drivers/staging/wfx/./traces.h:195:1: got restricted __le16 const [usertype] len drivers/staging/wfx/./traces.h:195:1: warning: incorrect type in assignment (different base types) drivers/staging/wfx/./traces.h:195:1: expected int msg_len drivers/staging/wfx/./traces.h:195:1: got restricted __le16 const [usertype] len drivers/staging/wfx/debug.c:319:20: warning: restricted __le16 degrades to integer drivers/staging/wfx/secure_link.c:85:27: warning: restricted __le16 degrades to integer drivers/staging/wfx/secure_link.c:85:27: warning: restricted __le16 degrades to integer Indeed, the attribute len is little-endian. We have to take to the endianness when we access it. Signed-off-by: -
Jérôme Pouiller authored
The struct hif_ind_startup is received from the hardware. So it is declared as little endian. However, it is also stored in the main driver structure and used on different places in the driver. Sparse complains about that: drivers/staging/wfx/data_tx.c:388:43: warning: restricted __le16 degrades to integer drivers/staging/wfx/bh.c:199:9: warning: restricted __le16 degrades to integer drivers/staging/wfx/bh.c:221:62: warning: restricted __le16 degrades to integer In order to make Sparse happy and to keep access from the driver easy, this patch declare hif_ind_startup with native endianness. On reception of this struct, this patch takes care to do byte-swap and keep Sparse happy. Signed-off-by: -
Jérôme Pouiller authored
The field packet_id is not interpreted by the device. It is only used as identifier for the device answer. So it is not necessary to declare it little endian. It fixes some warnings raised by Sparse without complexifying the code. Signed-off-by:
-
Jérôme Pouiller authored
The attribute indication_type is little-endian. We have to take to the endianness when we access it. Signed-off-by:
-
Jérôme Pouiller authored
The attribute event_id is little-endian. We have to take to the endianness when we access it. Signed-off-by:
-
Jérôme Pouiller authored
The attribute ps_mode_error is little-endian. We have to take to the endianness when we access it. Signed-off-by:
-
Jérôme Pouiller authored
The structs hif_{req,cnf}_read_mib contain only little endian values. Thus, it is necessary to fix byte ordering before to use them. Especially, sparse detected wrong accesses to fields mib_id and length. Signed-off-by: -
Jérôme Pouiller authored
The struct hif_cnf_tx contains only little endian values. Thus, it is necessary to fix byte ordering before to use them. Especially, sparse detected wrong access to fields media_delay and tx_queue_delay. Signed-off-by:
-
Jérôme Pouiller authored
The struct hif_rx_stats contains only little endian values. Thus, it is necessary to fix byte ordering before to use them. Signed-off-by:
-
Jérôme Pouiller authored
The field wakeup_period_max from struct hif_mib_beacon_wake_up_period is a u8. So, assigning it a __le16 produces a nasty bug on big-endian architectures. Signed-off-by:
-
Jérôme Pouiller authored
Sparse detects that le16_to_cpup() expects a __le16 * as argument. Change the cast operator to be compliant with sparse. Signed-off-by:
-
Jérôme Pouiller authored
le32_to_cpu(*x) can be advantageously converted in le32_to_cpup(x). Signed-off-by:
-
Jérôme Pouiller authored
Sparse detected that le32_to_cpu should be used instead of cpu_to_le32. Signed-off-by:
-
Jérôme Pouiller authored
The kernel coding style promotes the use of kernel types (u8, u16, u32, etc...) instead of the C99 ones. Signed-off-by:
-
Jérôme Pouiller authored
The kernel coding style expects no space after cast operator. This patch make the wfx driver compliant with this rule. Signed-off-by:
-
Jérôme Pouiller authored
Some function prototypes were not correctly aligned and/or exceed 80 columns. In some other cases, the prototypes were written on more lines than necessary. Signed-off-by:
-
Jérôme Pouiller authored
In order to keep the compilation times reasonable, we try to only include the necessary headers (especially header included from other headers). This patch clean up unnecessary headers inclusions. Signed-off-by:
-
Jérôme Pouiller authored
When possible, we prefer to use the macro ARRAY_SIZE rather than hard coding the number of elements. Signed-off-by:
-
Jérôme Pouiller authored
Sparse tool noticed that wfx_enable_beacon() is never used outside of sta.c. Therefore, it can be declared static. Signed-off-by:
-
Jérôme Pouiller authored
The function get_firmware() is only used from fwio.c. It can be declared static. Signed-off-by:
-
Jérôme Pouiller authored
When the chip starts in SDIO mode, the external IRQ (aka Out-Of-Band IRQ) cannot be used before to configure it. Therefore, the first exchanges with the chip have to be done without the OOB IRQ. This patch allow to poll the data until the OOB IRQ is correctly setup. In order to keep the code simpler, this patch also poll data even if OOB IRQ is not used. Signed-off-by:
-
Jérôme Pouiller authored
It is possible to check if an IRQ is ending by polling the control register. This function must used with care: if an IRQ fires while the host reads control register, the IRQ can be lost. However, it could be useful in some cases. Signed-off-by:
-
Jérôme Pouiller authored
Currently, the SPI implementation use a workqueue to acknowledge IRQ while the SDIO-OOB implementation use a threaded IRQ. The threaded also offers the advantage to allow level triggered IRQs. Uniformize the code and use threaded IRQ in both case. Therefore, prefer level triggered IRQs if the user does not specify it in the DT. Signed-off-by:
-
Jérôme Pouiller authored
When used over SDIO bus, device is able to use an external line to signal IRQs (also called Out-Of-Band IRQ). The current code have several problems: 1. The ISR cannot directly acknowledge IRQ since access to the bus is not atomic. This patch use a threaded IRQ to solve that issue. 2. On certain platforms, it is necessary to keep SDIO interruption enabled (with register SDIO_CCCR_IENx) (this part has inspired from the brcmfmac driver). Signed-off-by: -
Jérôme Pouiller authored
Currently, the ISR check if bus->core is not NULL. But, it is a useless check. bus->core is initialiased before to request IRQ and it is not assigned to NULL when it is released. Signed-off-by:
-
Jérôme Pouiller authored
In case of error in wfx_probe(), wdev->hw is freed. Since an error occurred, wfx_free_common() is called, then wdev->hw is freed again. Signed-off-by:
Michał Mirosław <mirq-linux@rere.qmqm.pl> Fixes: 4033714d ("staging: wfx: fix init/remove vs IRQ race") Link: https://lore.kernel.org/r/20200505123757.39506-4-Jerome.Pouiller@silabs.comSigned-off-by:
-
Jérôme Pouiller authored
The device take a few hundreds of milliseconds to start. However, the current code wait up to 10 second for the chip. We can safely reduce this value to 1 second. Thanks to that change, it is no more necessary to use an interruptible timeout. Signed-off-by:
-
