- 25 Jan, 2022 40 commits
-
-
Kees Cook authored
When building with -Warray-bounds, the following warning is emitted: In file included from ./include/linux/string.h:253, from ./arch/x86/include/asm/page_32.h:22, from ./arch/x86/include/asm/page.h:14, from ./arch/x86/include/asm/thread_info.h:12, from ./include/linux/thread_info.h:60, from ./arch/x86/include/asm/preempt.h:7, from ./include/linux/preempt.h:78, from ./include/linux/rcupdate.h:27, from ./include/linux/rculist.h:11, from ./include/linux/sched/signal.h:5, from ./drivers/staging/rtl8723bs/include/drv_types.h:17, from drivers/staging/rtl8723bs/core/rtw_recv.c:7: In function 'memcpy', inlined from 'wlanhdr_to_ethhdr' at drivers/staging/rtl8723bs/core/rtw_recv.c:1554:2: ./include/linux/fortify-string.h:41:33: warning: '__builtin_memcpy' offset [0, 5] is out of the bounds [0, 0] [-Warray-bounds] 41 | #define __underlying_memcpy __builtin_memcpy | ^ This is because the compiler sees it is possible for "ptr" to be a NULL value, and concludes that it has zero size and attempts to copy to it would overflow. Instead, remove the get_recvframe_data() entirely, as it's not possible for this to ever be NULL. Additionally add missing NULL checks after recvframe_pull() (which are present in the rtl8712 driver). Cc: Larry Finger <Larry.Finger@lwfinger.net> Cc: Phillip Potter <phil@philpotter.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Michael Straube <straube.linux@gmail.com> Cc: Fabio Aiuto <fabioaiuto83@gmail.com> Cc: linux-staging@lists.linux.dev Reviewed-by:Hans de Goede <hdegoede@redhat.com> Signed-off-by:
Kees Cook <keescook@chromium.org> Link: https://lore.kernel.org/r/20220118193327.2822099-2-keescook@chromium.orgSigned-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Jérôme Pouiller authored
Since the WF200 VID/PID are not reliable, it's recommended to declare it in the DT. Until now, if the device was not declared, the driver just printed a warning and continue. But, the risk of a collision is too high, the driver now returns an error. Signed-off-by:
Jérôme Pouiller <jerome.pouiller@silabs.com> Link: https://lore.kernel.org/r/20220113085524.1110708-32-Jerome.Pouiller@silabs.comSigned-off-by:
-
Jérôme Pouiller authored
"config-file" is too broad. Replace it by "silabs,antenna-config-file" which is more explicit. The attribute "config-file" is probably not widely used. This patch obviously breaks setups that use this attribute. Signed-off-by:
-
Jérôme Pouiller authored
Values "silabs,wfx-sdio" and "silabs,wfx-spi" are deprecated for a while now. We take advantage of getting out of the staging tree to drop them and start from a blank sheet. Signed-off-by:
-
Jérôme Pouiller authored
There is currently, a mismatch between the location of the firmware in linux-firmware and the path written in the driver. We take this opportunity to relocate the WF200 firmware in wfx/ instead of silabs/. Signed-off-by:
-
Jérôme Pouiller authored
"WF200" only designates the chip. To make a WiFi board, the chip must be associated with an antenna. The antenna configuration is located in separate files (aka PDS files in Silabs wording). Currently, user has to write in his DT something like: compatible = "silabs,wf200"; config = "brd4001a.pds"; It is far better to embed a list of known boards (chip + antenna) in the driver. So the user just have to declare: compatible = "silabs,brd4001a"; This patch add the configurations for the evaluation boards sold by Silabs. To provide a full plug-and-play experience, the associated PDS files[1] will be available in linux-firmware. This patch does not break compatibility with existing setups. Signed-off-by: -
Jérôme Pouiller authored
ps_timeout should be in nl80211, not in debugfs. Let's remove it until the driver is accepted. Signed-off-by:
-
Jérôme Pouiller authored
It is not necessary to prefix error logs with the function name when an error message is unique in the code. Note this patch still prefixes the message 'received event for non-existent vif' with the function name since it is used several times. Signed-off-by:
-
Jérôme Pouiller authored
It seems that BUILD_BUG_ON_MSG() is a bit more popular. Signed-off-by:
-
Jérôme Pouiller authored
Prefer fully named labels to handle errors instead of err0, err1, ... Signed-off-by:
-
Jérôme Pouiller authored
Some structs members were not properly aligned. Signed-off-by:
-
Jérôme Pouiller authored
Until now, this driver was written in 80 columns style. However, since all the functions are prefixed with "wfx_", this constraint is no more respected in the last patches. From the perspective of kernel Coding Style, it is not a problem since it is now allowed to write code on 100 columns. This patch just unify the code to use 100 columns in every comments. Signed-off-by:
-
Jérôme Pouiller authored
Until now, this driver was written in 80 columns style. However, since all the functions are prefixed with "wfx_", this constraint is no more respected in the last patches. From the perspective of kernel Coding Style, it is not a problem since it is now allowed to write code on 100 columns. This patch just unify the code to use 100 columns. Signed-off-by:
-
Jérôme Pouiller authored
All the types related to a driver should use the same prefix. Signed-off-by:
-
Jérôme Pouiller authored
All the types related to a driver should use the same prefix. Signed-off-by:
-
Jérôme Pouiller authored
tx_policy_is_equal() was the only function from data_tx.c without the prefix wfx_. Signed-off-by:
-
Jérôme Pouiller authored
All the functions related to a driver should use the same prefix. Signed-off-by:
-
Jérôme Pouiller authored
All the functions related to a driver should use the same prefix. Signed-off-by:
-
Jérôme Pouiller authored
All the functions related to a driver should use the same prefix. Signed-off-by:
-
Jérôme Pouiller authored
The prefix 'ieee80211' is reserved for mac80211. It should not been used. Signed-off-by:
-
Jérôme Pouiller authored
The prefix 'ieee80211' is reserved for mac80211. It should not been used. Signed-off-by:
-
Jérôme Pouiller authored
The hardware fills struct hif_ind_startup with little endian values. So, declare it with little endian fields. It is now a bit more verbose to access to fields of struct hif_ind_startup, but it is less confusing. Signed-off-by:
-
Jérôme Pouiller authored
Magic values are not recommended. Signed-off-by:
-
Jérôme Pouiller authored
Magic values are not recommended. Signed-off-by:
-
Jérôme Pouiller authored
It "IS_ALIGNED(ptr, 4)" is more explicit than "ptr & 3". Signed-off-by:
-
Jérôme Pouiller authored
In the old days, this file was shared with other projects. Obviously, this "#ifdef __KERNEL__" have no reasons to be in the kernel. Signed-off-by:
-
Jérôme Pouiller authored
Braces are not necessary for single statement blocks Signed-off-by:
-
Jérôme Pouiller authored
I have got questions about this line from several reviewers. A comment is definitively welcome. Signed-off-by:
-
Jérôme Pouiller authored
Using DMA with stack allocated buffers is not supported, whatever the value of CONFIG_VMAP_STACK. Signed-off-by:
-
Jérôme Pouiller authored
Each headers files should include every types it needs to compile (ie. "gcc $CFLAGS -xc file.h" should compile) Signed-off-by:
-
Jérôme Pouiller authored
Apache-2.0 is not allowed in the kernel. Signed-off-by:
-
Jérôme Pouiller authored
License was missing in Kconfig. Makefile license was GPL-2.0 while rest of the driver use GPL-2.0-only. Signed-off-by:
-
Martin Kaiser authored
ODM_ConfigBBWithHeaderFile is a "multiplexer" for ODM_ReadAndConfig_... functions. It's called only from phy_BB8188E_Config_ParaFile. We can remove ODM_ConfigBBWithHeaderFile and call the ODM_ReadAndConfig_... functions directly. ODM_ReadAndConfig_PHY_REG_PG_8188E does not return an error status, there's no need for a check. Signed-off-by:
Martin Kaiser <martin@kaiser.cx> Link: https://lore.kernel.org/r/20220110124638.6909-10-martin@kaiser.cxSigned-off-by:
-
Martin Kaiser authored
Remove some unused defines for cck, ofdm and mcs rates. Signed-off-by:
-
Martin Kaiser authored
The ODM_ConfigMACWithHeaderFile function is only one line. Merge it into PHY_MACConfig8188E, which is its only caller. Signed-off-by:
-
Martin Kaiser authored
Most of the ODM_ReadAndConfig_... functions return an error status. Update their callers to process these return values. Signed-off-by:
-
Martin Kaiser authored
The READ_AND_CONFIG macro builds a function name from chip type and table name. Remove the macro and use the resulting function names directly. This makes the code easier to read. Signed-off-by:
-
Martin Kaiser authored
The only caller of ODM_ConfigRFWithHeaderFile sets rfpath to RF_PATH_A. Remove the parameter and the check for RF_PATH_A inside the function. Signed-off-by:
-
Martin Kaiser authored
ODM_ConfigRFWithHeaderFile does not use its content parameter. Remove it. Signed-off-by:
-
Martin Kaiser authored
The eRFPath variable in phy_RF6052_Config_ParaFile is always 0. Remove the variable and use 0 directly. (enum rf_radio_path)eRFPath can be replaced with RF_PATH_A, which is defined as 0. Signed-off-by:
-
