Kalle Valo [Wed, 13 May 2020 16:10:08 +0000 (19:10 +0300)]
Merge tag 'iwlwifi-next-for-kalle-2020-05-08' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-next
Second set of iwlwifi patches intended for v5.8
* Support new FW APIs;
* Remove some old and unused features;
* HW configuration rework continues;
* Some queues rework by Johannes;
* Enable A-AMSDU in low latency;
* Some debugging fixes;
* Some other small fixes and clean-ups;
# gpg: Signature made Fri 08 May 2020 10:08:58 AM EEST using RSA key ID 1A3CC5FA
# gpg: Good signature from "Luciano Roth Coelho (Luca) <luca@coelho.fi>"
# gpg: aka "Luciano Roth Coelho (Intel) <luciano.coelho@intel.com>"
drivers/net/wireless/realtek/rtl818x/rtl8187/rtl8225.c:609:17: warning:
‘rtl8225z2_tx_power_ofdm’ defined but not used
static const u8 rtl8225z2_tx_power_ofdm[] = {
Acked-by: Hin-Tak Leung <htl10@users.sourceforge.net> Acked-by: Larry Finger <Larry.Finger@lwfinger.net> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: ChenTao <chentao107@huawei.com> Signed-off-by: Kalle Valo <kvalo@codeaurora.org> Link: https://lore.kernel.org/r/20200513011754.28432-1-chentao107@huawei.com
Ping-Ke Shih [Tue, 12 May 2020 10:26:20 +0000 (18:26 +0800)]
rtw88: fill zeros to words 0x06 and 0x07 of security cam entry
8723D adds some experimental features to word 0x06 of cam entry, so fill
zeros to initialize them to off state. For existing chips, these two words
are reserved and always zeros, so this change is harmless for them.
Ping-Ke Shih [Tue, 12 May 2020 10:26:19 +0000 (18:26 +0800)]
rtw88: 8723d: Add coex support
8723D is a Wifi+BT combo card. To make them work properly, we need coex
mechanism to avoid interference, such as TX simultaneously. Basically,
coex.c provide main algorithm to deal with many use cases, and this commit
adds some parameters and ops differ from other chips, because coex
hardware and WiFi generation are changed.
Ping-Ke Shih [Tue, 12 May 2020 10:26:18 +0000 (18:26 +0800)]
rtw88: 8723d: set ltecoex register address in chip_info
Since 8723D use different address of ltecoex register, this commit add a
new field in chip_info and fill proper address. Then, ltecoex_read_reg()
and ltecoex_reg_write() can use them to access ltecoex according to chip.
Ping-Ke Shih [Tue, 12 May 2020 10:26:17 +0000 (18:26 +0800)]
rtw88: 8723d: implement flush queue
Flush queue is used to check if queue is empty, before doing something
else. Since 8723D uses different registers and page number of
availabl/reserved occupy 8 bits instead of 16 bits, so use a 'wsize' field
to discriminate which rtw_read{8,16} is adopted.
Ping-Ke Shih [Tue, 12 May 2020 10:26:16 +0000 (18:26 +0800)]
rtw88: 8723d: Add shutdown callback to disable BT USB suspend
Without this patch, wifi card can't initialize properly due to BT in USB
suspend state. So, we disable BT USB suspend (wakeup) in shutdown callback
that is the moment before rebooting. To save BT USB power, we can't do this
in 'remove' callback.
Ping-Ke Shih [Tue, 12 May 2020 10:26:15 +0000 (18:26 +0800)]
rtw88: 8723d: Add power tracking
When chip's temperature is changed, RF characters are changed. To keep the
characters to be consistent, 8723d uses thermal meter to assist in
calibrating LCK, IQK, crystal and TX power.
A base thermal value is programmed in efuse, all calibration data in
MP process is based on this thermal value. So we calucate the delta of
thermal value between the base value, and use this delta to reference XTAL
and TX power offset tables to know how much we need to adjust.
For IQK and LCK, driver checks if delta of thermal value is over 8, then
they are triggered.
For crystal adjustment, when delta of thermal value is changed, we check
XTAL tables to get offset of XTAL value. If thermal value is larger than
base value, positive table (_p as suffix) is used. Otherwise, we use
negative table (_n as suffix). Then, we add offset to XTAL default value
programmed in efuse, and write sum value to register.
To compensate TX power, there are two hierarchical tables. First level use
delta of thermal value to access eight tables to yield delta of TX power
index. Then, plus base TX power index to get index of BB swing table
(second level tables) where register value is induced.
BB swing table can't deal with all cases, if index of BB swing table is
over the size of the table. In this case, TX AGC is used to compensate the
remnant part. Assume 'upper' is the upper bound of BB swing table, and
'target' is the desired index. Then, we can illustrate them as
Ping-Ke Shih [Tue, 12 May 2020 10:26:14 +0000 (18:26 +0800)]
rtw88: 8723d: add IQ calibration
IQ calibration is used to calibrate RF characteristic to yield expected
performance. Basically, we do calibration twice and compare the similarity
to determine calibration is good or not, if not we do the third
calibration, and then compare with the results of first and second
calibration. If it still not similar, IQK is failed.
Before doing calibration, we need to backup registers that will be
modified in calibration procedure, and restore these registers after
calibration is done.
A calibration procedure can divided into four sub-procedures that are
S1-TX, S1-RX, S0-TX and S0-RX. Where, S1 and S0 represent to path A and B
respectively. Each sub-procedure configure proper registers, and then
rigger one-shot calibration and poll until completion. For RX calibration,
it needs to do twice one-shot calibration, first one is to yield parameter
used by second one.
The result of TX part is stored for TX power tracking that adjusts TX AGC
to output expected power.
Chung-Hsien Hsu [Tue, 12 May 2020 10:03:08 +0000 (05:03 -0500)]
brcmfmac: fix WPA/WPA2-PSK 4-way handshake offload and SAE offload failures
An incorrect value of use_fwsup is set for 4-way handshake offload for
WPA//WPA2-PSK, caused by commit 3b1e0a7bdfee ("brcmfmac: add support for
SAE authentication offload"). It results in missing bit
BRCMF_VIF_STATUS_EAP_SUCCESS set in brcmf_is_linkup() and causes the
failure. This patch correct the value for the case.
Also setting bit BRCMF_VIF_STATUS_EAP_SUCCESS for SAE offload case in
brcmf_is_linkup() to fix SAE offload failure.
Gustavo A. R. Silva [Thu, 7 May 2020 19:26:47 +0000 (14:26 -0500)]
rndis_wlan: Replace zero-length array with flexible-array
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
Gustavo A. R. Silva [Thu, 7 May 2020 19:19:26 +0000 (14:19 -0500)]
qtnfmac: Replace zero-length array with flexible-array
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Reviewed-by: Sergey Matyukevich <sergey.matyukevich.os@quantenna.com> Signed-off-by: Kalle Valo <kvalo@codeaurora.org> Link: https://lore.kernel.org/r/20200507191926.GA15970@embeddedor
Gustavo A. R. Silva [Thu, 7 May 2020 19:02:10 +0000 (14:02 -0500)]
prism54: Replace zero-length array with flexible-array
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
Gustavo A. R. Silva [Thu, 7 May 2020 18:59:14 +0000 (13:59 -0500)]
mwl8k: Replace zero-length array with flexible-array
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
Gustavo A. R. Silva [Thu, 7 May 2020 18:55:29 +0000 (13:55 -0500)]
iwlegacy: Replace zero-length array with flexible-array
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
Gustavo A. R. Silva [Thu, 7 May 2020 18:54:51 +0000 (13:54 -0500)]
ipw2x00: Replace zero-length array with flexible-array
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
Gustavo A. R. Silva [Tue, 5 May 2020 23:52:05 +0000 (18:52 -0500)]
rndis_wlan: Remove logically dead code
caps_buf is always of size sizeof(*caps) because
sizeof(caps->auth_encr_pair) * 16 is always zero. Notice
that when using zero-length arrays, sizeof evaluates to zero[1].
So, the code introduced by
commit 0308383f9591 ("rndis_wlan: get max_num_pmkids from device")
is logically dead, hence is never executed and can be removed. As a
consequence, the rest of the related code can be refactored a bit.
Notice that this code has been out there since March 2010.
Chen Zhou [Fri, 8 May 2020 01:32:49 +0000 (09:32 +0800)]
brcmfmac: make non-global functions static
Fix sparse warning:
drivers/net/wireless/broadcom/brcm80211/brcmfmac/p2p.c:2206:5:
warning: symbol 'brcmf_p2p_get_conn_idx' was not declared. Should it be static?
Soontak Lee [Wed, 6 May 2020 13:03:21 +0000 (08:03 -0500)]
brcmfmac: Use seq/seq_len and set iv_initialize when plumbing of rxiv in (GTK) keys
When plumbing rxiv for (GTK) keys, current code does not use seq/seq_len
when present nor set iv_initialized for iovar wsec_key. This could
result in missing broadcast traffic after GTK rekey. The fix is setting
iv_initialized and using seq/seq_len for iovar wsec_key.
Ryohei Kondo [Wed, 6 May 2020 13:03:20 +0000 (08:03 -0500)]
brcmfmac: use actframe_abort to cancel ongoing action frame
The driver sends an action frame down and waits for dwell time to be
completed or aborted before sending out the next action frame.
Driver issues "scan abort" to cancel the current time slot, but this
doesn't have any effect because, we are not using scan engine for
sending action frame.
Fix is to use "actframe_abort" to cancels the current action frame.
Pramod Prakash [Tue, 5 May 2020 06:51:27 +0000 (01:51 -0500)]
brcmfmac: fix 802.1d priority to ac mapping for pcie dongles
802.1d defines 0,3 for BE and 1,2 for BK. In pcie dongles, 0 & 3 are
mapped to 0 and 1,2 are mapped to 1. This change corrects this mapping,
so that BE & BK are given access precedence accordingly by pcie dongles.
Saravanan Shanmugham [Tue, 5 May 2020 06:51:26 +0000 (01:51 -0500)]
brcmfmac: map 802.1d priority to precedence level based on AP WMM params
In WLAN, priority among various access categories of traffic is
always set by the AP using WMM parameters and this may not always
follow the standard 802.1d priority.
In this change, priority is adjusted based on the AP WMM params
received as part of the Assoc Response and the same is later used
to map the priority of all incoming traffic.
In a specific scenario where EDCA parameters are configured to be same
for all ACs, use the default FW priority definition to avoid queuing
packets of all ACs to the same priority queue.
This change fixes the following 802.11 certification tests:
* 11n - 5.2.31 ACM Bit Conformance test
* 11n - 5.2.32 AC Parameter Modification test
* 11ac - 5.2.33 TXOP Limit test
Signed-off-by: Saravanan Shanmugham <saravanan.shanmugham@cypress.com> Signed-off-by: Justin Li <justin.li@cypress.com> Signed-off-by: Madhan Mohan R <madhanmohan.r@cypress.com> Signed-off-by: Chi-hsien Lin <chi-hsien.lin@cypress.com> Signed-off-by: Kalle Valo <kvalo@codeaurora.org> Link: https://lore.kernel.org/r/1588661487-21884-2-git-send-email-chi-hsien.lin@cypress.com
iwlwifi: dbg_ini: differentiate ax210 hw with same hw type
There are several "flavors" of HW that have the same HW type, but
can be told apart after reading a certain perph register. This
is easy to do in runtime, but more complicated to do when looking
at the logs offline.
To make it easier to tell apart these "flavors" when looking at
the dumped dbg info, add these bits to the HW type, allowing
simple differentiation.
Johannes Berg [Sat, 25 Apr 2020 10:04:58 +0000 (13:04 +0300)]
iwlwifi: pcie: skip fragmented receive buffers
We don't really expect fragmented RBs, and don't seem to be seeing
them in practice since that would've caused a crash. Nevertheless,
we should be expecting the hardware to send them.
Parse the flag indicating a fragmented buffer, but then discard it
and any fragments thereof, at least for now. We need to do more
work in the higher layers to properly deal with this, since we may
not get "normal" firmware notifications that are fragmented, only
RX, and then we need to put it back together and add the necessary
API to report a chain of things to the higher layers, this doesn't
fit into the struct iwl_rx_cmd_buffer today.
Johannes Berg [Sat, 25 Apr 2020 10:04:56 +0000 (13:04 +0300)]
iwlwifi: mvm: don't transmit on unallocated queue
We can currently end up transmitting on an unallocated queue, if
the allocation fails. Stop doing that, by simply not transmitting.
We don't have any better strategy here, unfortunately, but the
previous commits make that much less likely.
Johannes Berg [Sat, 25 Apr 2020 10:04:54 +0000 (13:04 +0300)]
iwlwifi: use longer queues for 256-BA
When we have 256 block-ack support, we may need to be very fast
to provide a lot of frames to the hardware to transmit, but that
cannot be guaranteed. Use a longer queue size to have more time,
and the next possible queue size is 1024 since it must be a power
of two.
Johannes Berg [Sat, 25 Apr 2020 10:04:53 +0000 (13:04 +0300)]
iwlwifi: pcie: gen2: use DMA pool for byte-count tables
Since the recent patch in this area, we no longer allocate 64k
for a single queue, but only 1k, which still means a full page.
Use a DMA pool to reduce this further, since we will have a lot
of queues in a typical system that can share pages.
Johannes Berg [Sat, 25 Apr 2020 10:04:52 +0000 (13:04 +0300)]
iwlwifi: pcie: remove some dead code
We can never get into this code with a gen2/3 device, and therefore
don't need to allocate the byte count tables in a single contiguous
DMA region. Just WARN and bail out if something is misconfigured.
Johannes Berg [Sat, 25 Apr 2020 10:04:51 +0000 (13:04 +0300)]
iwlwifi: dbg: mark a variable __maybe_unused
If CONFIG_IWLWIFI_DEBUGFS is not set, the variable is assigned
but not checked, resulting in a compiler warning. Suppress it,
we need the variable for the debugfs-enabled case.
Johannes Berg [Sat, 25 Apr 2020 10:04:50 +0000 (13:04 +0300)]
iwlwifi: mvm: attempt to allocate smaller queues
We currently attempt to allocate queues that are 512 entries long,
but that requires 32 KiB memory, which may not be available, at
least not contiguously. If we fail to allocate, attempt to use a
smaller queue all the way down to 16 entries (which fit into a
single page).
Johannes Berg [Sat, 25 Apr 2020 10:04:49 +0000 (13:04 +0300)]
iwlwifi: pcie: allocate much smaller byte-count table
The hardware needs a byte-count table with the size of each frame
on the queue to build A-MPDUs, but:
* newer generation no longer have the duplicated space at the end,
they can deal with the wrap properly - and we don't even fill
the dup anyway
* we have a maximum queue size of 512 right now and don't use the
theoretical hardware maximum of 65536.
Together, this reduces the byte count table DMA allocation from
64KiB (65536*2 + 64*2 rounded up) to 1 KiB (though that might be
rounded up to a full 4 KiB page by the allocator, not sure it can
share the allocations.)
iwlwifi: pcie: convert all AX101 devices to the device tables
Convert all Qu/Hr1 devices to the new device tables, by modifying the
corresponding structures, adding a new name and generalizing the
device recognition.
iwlwifi: pcie: remove mangling for iwl_ax101_cfg_qu_hr
All devices that use iwl_ax101_cfg_qu_hr are recognized via the device
info table, so the cfg will never be iwl_ax101_cfg_qu_hr. Remove the
code that converts this into QuZ and Qu-C, since it's not needed
anymore.
iwlwifi: mvm: add IML/ROM information to the assertion dumps
Dump the IML/ROM error code and data, which are read from some
registers, when printing an assertion dump. This makes it easier to
debug IML/ROM errors.
iwlwifi: mvm: initialize iwl_dev_tx_power_cmd to zero
If the REDUCE_TX_POWER_CMD version is v4 or v5, we are not
initializing some values before sending to the FW, which causes SAR
not to work properly. Solve this by initializing the struct in the
declaration.
Shahar S Matityahu [Fri, 24 Apr 2020 16:46:57 +0000 (19:46 +0300)]
iwlwifi: dbg: set debug descriptor to NULL outside of iwl_fw_free_dump_desc
To avoid static analysis warning and to make the flow more readable, set
the debug descriptor to NULL outside iwl_fw_free_dump_desc and only in
the required places.
iwlwifi: mvm: set properly station flags in STA_HE_CTXT_CMD
For ACK_ENABLED and 32BIT_BA_BITMAP flags check the station capabilities
rather than bss_conf.ack_enabled and bss_conf.multi_sta_back_32bit.
These fields are stations capabilities and should not be in bss_conf.
Also note that the bss_conf flags are set in station mode only.
In the next patch I will remove ack_enabled and multi_sta_back_32bit
from the bss_conf structure.
iwlwifi: avoid debug max amsdu config overwriting itself
If we set amsdu_len one after another the second one overwrites
the orig_amsdu_len so allow only moving from debug to non debug state.
Also the TLC update check was wrong: it was checking that also the orig
is smaller then the new updated size, which is not the case in debug
amsdu mode.
This should be controlled by the firmware debugging mechanism
and not by a module parameter. This has always been true.
Remove it and assume it is set.
This module parameter should not be mangled by users.
This relates to a very old driver and I doubt people can
really check the antenna coupling in a way that would make
the BT Coexistence work better with a real value.
Drop it.
Johannes Berg [Fri, 24 Apr 2020 15:48:07 +0000 (18:48 +0300)]
iwlwifi: fw api: fix PHY data 2/3 position
In AX210 devices, the PHY data wasn't actually reported, but now
that it's going to be reported it turns out that the position is
supposed to be the other way around, fix that.
Eric Dumazet [Thu, 7 May 2020 17:05:39 +0000 (10:05 -0700)]
net: relax SO_TXTIME CAP_NET_ADMIN check
Now sch_fq has horizon feature, we want to allow QUIC/UDP applications
to use EDT model so that pacing can be offloaded to the kernel (sch_fq)
or the NIC.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Willem de Bruijn <willemb@google.com> Acked-by: Willem de Bruijn <willemb@google.com> Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Thu, 7 May 2020 16:32:22 +0000 (09:32 -0700)]
bonding: propagate transmit status
Currently, bonding always returns NETDEV_TX_OK to its caller.
It is worth trying to be more accurate : TCP for instance
can have different recovery strategies if it can have more
precise status, if packet was dropped by slave qdisc.
This is especially important when host is under stress.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Jay Vosburgh <j.vosburgh@gmail.com> Cc: Veaceslav Falico <vfalico@gmail.com> Cc: Andy Gospodarek <andy@greyhouse.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Colin Ian King [Thu, 7 May 2020 14:34:30 +0000 (15:34 +0100)]
net: phy: fix less than zero comparison with unsigned variable val
The unsigned variable val is being checked for an error by checking
if it is less than zero. This can never occur because val is unsigned.
Fix this by making val a plain int.
Addresses-Coverity: ("Unsigned compared against zero") Fixes: bdbdac7649fa ("ethtool: provide UAPI for PHY master/slave configuration.") Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
YueHaibing [Thu, 7 May 2020 14:24:06 +0000 (16:24 +0200)]
net/smc: remove set but not used variables 'del_llc, del_llc_resp'
Fixes gcc '-Wunused-but-set-variable' warning:
net/smc/smc_llc.c: In function 'smc_llc_cli_conf_link':
net/smc/smc_llc.c:753:31: warning:
variable 'del_llc' set but not used [-Wunused-but-set-variable]
struct smc_llc_msg_del_link *del_llc;
^
net/smc/smc_llc.c: In function 'smc_llc_process_srv_delete_link':
net/smc/smc_llc.c:1311:33: warning:
variable 'del_llc_resp' set but not used [-Wunused-but-set-variable]
struct smc_llc_msg_del_link *del_llc_resp;
^
Signed-off-by: YueHaibing <yuehaibing@huawei.com> Signed-off-by: Karsten Graul <kgraul@linux.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jacob Keller [Thu, 7 May 2020 00:58:27 +0000 (17:58 -0700)]
net: remove newlines in NL_SET_ERR_MSG_MOD
The NL_SET_ERR_MSG_MOD macro is used to report a string describing an
error message to userspace via the netlink extended ACK structure. It
should not have a trailing newline.
Add a cocci script which catches cases where the newline marker is
present. Using this script, fix the handful of cases which accidentally
included a trailing new line.
I couldn't figure out a way to get a patch mode working, so this script
only implements context, report, and org.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Andy Whitcroft <apw@canonical.com> Cc: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
net: ethernet: ti: am65x-cpts: follow up dt bindings update
This series is follow update for TI A65x/J721E Common platform time sync (CPTS)
driver [1] to implement DT bindings review comments from
Rob Herring <robh@kernel.org> [2].
- "reg" and "compatible" properties are made required for CPTS DT nodes which
also required to change K3 CPSW driver to use of_platform_device_create()
instead of of_platform_populate() for proper CPTS and MDIO initialization
- minor DT bindings format changes
- K3 CPTS example added to K3 MCU CPSW bindings
Grygorii Strashko [Wed, 6 May 2020 18:14:00 +0000 (21:14 +0300)]
dt-binding: net: ti: am65x-cpts: make reg and compatible required
This patch follows K3 CPTS review comments from Rob Herring
<robh@kernel.org>.
- "reg" and "compatible" properties are required now
- minor format changes
- K3 CPTS example added to K3 MCU CPSW bindings
Cc: Rob Herring <robh@kernel.org> Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Grygorii Strashko [Wed, 6 May 2020 18:13:59 +0000 (21:13 +0300)]
net: ethernet: ti: am65-cpsw-nuss: use of_platform_device_create() for mdio
The MCU CPSW expected to populate only MDIO device, but follow up patches
will add "compatible" property to the MCU CPSW CPTS node which will cause
creation of CPTS device and MCU CPSW init failure. Hence, switch to use
of_platform_device_create() instead of of_platform_populate() for MDIO
device population.
Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ioana Ciornei [Wed, 6 May 2020 17:47:17 +0000 (20:47 +0300)]
dpaa2-eth: create a function to flush the XDP fds
Create an independent function that takes a particular frame queue and
an array of frame descriptors and tries to enqueue them until it hits
the maximum number fo retries. The same function will be used in the
next patch also on the XDP_TX path.
Also, create the dpaa2_eth_xdp_fds structure to incorporate the array of
FDs as well as the number of FDs already populated.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Wed, 6 May 2020 15:47:45 +0000 (15:47 +0000)]
hsr: remove WARN_ONCE() in hsr_fill_frame_info()
When VLAN frame is being sent, hsr calls WARN_ONCE() because hsr doesn't
support VLAN. But using WARN_ONCE() is overdoing.
Using netdev_warn_once() is enough.
Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ioana Ciornei [Tue, 5 May 2020 20:14:29 +0000 (23:14 +0300)]
soc: fsl: dpio: properly compute the consumer index
Mask the consumer index before using it. Without this, we would be
writing frame descriptors beyond the ring size supported by the QBMAN
block.
Fixes: 3b2abda7d28c ("soc: fsl: dpio: Replace QMAN array mode with ring mode enqueue") Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com> Acked-by: Li Yang <leoyang.li@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Tue, 5 May 2020 19:20:56 +0000 (22:20 +0300)]
net: dsa: sja1105: implement tc-gate using time-triggered virtual links
Restrict the TTEthernet hardware support on this switch to operate as
closely as possible to IEEE 802.1Qci as possible. This means that it can
perform PTP-time-based ingress admission control on streams identified
by {DMAC, VID, PCP}, which is useful when trying to ensure the
determinism of traffic scheduled via IEEE 802.1Qbv.
The oddity comes from the fact that in hardware (and in TTEthernet at
large), virtual links always need a full-blown action, including not
only the type of policing, but also the list of destination ports. So in
practice, a single tc-gate action will result in all packets getting
dropped. Additional actions (either "trap" or "redirect") need to be
specified in the same filter rule such that the conforming packets are
actually forwarded somewhere.
Apart from the VL Lookup, Policing and Forwarding tables which need to
be programmed for each flow (virtual link), the Schedule engine also
needs to be told to open/close the admission gates for each individual
virtual link. A fairly accurate (and detailed) description of how that
works is already present in sja1105_tas.c, since it is already used to
trigger the egress gates for the tc-taprio offload (IEEE 802.1Qbv). Key
point here, we remember that the schedule engine supports 8
"subschedules" (execution threads that iterate through the global
schedule in parallel, and that no 2 hardware threads must execute a
schedule entry at the same time). For tc-taprio, each egress port used
one of these 8 subschedules, leaving a total of 4 subschedules unused.
In principle we could have allocated 1 subschedule for the tc-gate
offload of each ingress port, but actually the schedules of all virtual
links installed on each ingress port would have needed to be merged
together, before they could have been programmed to hardware. So
simplify our life and just merge the entire tc-gate configuration, for
all virtual links on all ingress ports, into a single subschedule. Be
sure to check that against the usual hardware scheduling conflicts, and
program it to hardware alongside any tc-taprio subschedule that may be
present.
The following scenarios were tested:
1. Quantitative testing:
tc qdisc add dev swp2 clsact
tc filter add dev swp2 ingress flower skip_sw \
dst_mac 42:be:24:9b:76:20 \
action gate index 1 base-time 0 \
sched-entry OPEN 1200 -1 -1 \
sched-entry CLOSE 1200 -1 -1 \
action trap
Vladimir Oltean [Tue, 5 May 2020 19:20:55 +0000 (22:20 +0300)]
net: dsa: sja1105: support flow-based redirection via virtual links
Implement tc-flower offloads for redirect, trap and drop using
non-critical virtual links.
Commands which were tested to work are:
# Send frames received on swp2 with a DA of 42:be:24:9b:76:20 to the
# CPU and to swp3. This type of key (DA only) when the port's VLAN
# awareness state is off.
tc qdisc add dev swp2 clsact
tc filter add dev swp2 ingress flower skip_sw dst_mac 42:be:24:9b:76:20 \
action mirred egress redirect dev swp3 \
action trap
# Drop frames received on swp2 with a DA of 42:be:24:9b:76:20, a VID
# of 100 and a PCP of 0.
tc filter add dev swp2 ingress protocol 802.1Q flower skip_sw \
dst_mac 42:be:24:9b:76:20 vlan_id 100 vlan_prio 0 action drop
Under the hood, all rules match on DMAC, VID and PCP, but when VLAN
filtering is disabled, those are set internally by the driver to the
port-based defaults. Because we would be put in an awkward situation if
the user were to change the VLAN filtering state while there are active
rules (packets would no longer match on the specified keys), we simply
deny changing vlan_filtering unless the list of flows offloaded via
virtual links is empty. Then the user can re-add new rules.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Tue, 5 May 2020 19:20:54 +0000 (22:20 +0300)]
net: dsa: sja1105: make room for virtual link parsing in flower offload
Virtual links are a sja1105 hardware concept of executing various flow
actions based on a key extracted from the frame's DMAC, VID and PCP.
Currently the tc-flower offload code supports only parsing the DMAC if
that is the broadcast MAC address, and the VLAN PCP. Extract the key
parsing logic from the L2 policers functionality and move it into its
own function, after adding extra logic for matching on any DMAC and VID.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Tue, 5 May 2020 19:20:52 +0000 (22:20 +0300)]
net: dsa: introduce a dsa_port_from_netdev public helper
As its implementation shows, this is synonimous with calling
dsa_slave_dev_check followed by dsa_slave_to_port, so it is quite simple
already and provides functionality which is already there.
However there is now a need for these functions outside dsa_priv.h, for
example in drivers that perform mirroring and redirection through
tc-flower offloads (they are given raw access to the flow_cls_offload
structure), where they need to call this function on act->dev.
But simply exporting dsa_slave_to_port would make it non-inline and
would result in an extra function call in the hotpath, as can be seen
for example in sja1105:
Because we want to avoid possible performance regressions, introduce
this new function which is designed to be public.
Suggested-by: Vivien Didelot <vivien.didelot@gmail.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Vivien Didelot <vivien.didelot@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 7 May 2020 20:22:35 +0000 (13:22 -0700)]
Merge tag 'wireless-drivers-next-2020-05-07' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next
Kalle Valo says:
====================
wireless-drivers-next patches for v5.8
First set of patches for v5.8. Changes all over, ath10k apparently
seeing most new features this time. rtw88 also had lots of changes due
to preparation for new hardware support.
In this pull request there's also a new macro to include/linux/iopoll:
read_poll_timeout_atomic(). This is needed by rtw88 for atomic
polling.
Major changes:
ath11k
* add debugfs file for testing ADDBA and DELBA
* add 802.11 encapsulation offload on hardware support
* add htt_peer_stats_reset debugfs file
ath10k
* enable VHT160 and VHT80+80 modes
* enable radar detection in secondary segment
* sdio: disable TX complete indication to improve throughput
* sdio: decrease power consumption
* sdio: add HTT TX bundle support to increase throughput
* sdio: add rx bitrate reporting
ath9k
* improvements to AR9002 calibration logic
carl9170
* remove buggy P2P_GO support
p54usb
* add support for AirVasT USB stick
rtw88
* add support for antenna configuration
ti wlcore
* add support for AES_CMAC cipher
iwlwifi
* support for a few new FW API versions
* new hw configs
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 7 May 2020 20:21:12 +0000 (13:21 -0700)]
Merge branch 'Add-QRTR-MHI-client-driver'
Manivannan Sadhasivam says:
====================
Add QRTR MHI client driver
Here is the series adding MHI client driver support to Qualcomm IPC router
protocol. MHI is a newly added bus to kernel which is used to communicate to
external modems over a physical interface like PCI-E. This driver is used to
transfer the QMI messages between the host processor and external modems over
the "IPCR" channel.
For QRTR, this driver is just another driver acting as a transport layer like
SMD.
Currently this driver is needed to control the QCA6390 WLAN device from ath11k.
The ath11k MHI controller driver will take care of booting up QCA6390 and
bringing it to operating state. Later, this driver will be used to transfer QMI
messages over the MHI-IPCR channel.
The second patch of this series removes the ARCH_QCOM dependency for QRTR. This
is needed because the QRTR driver will be used with x86 machines as well to talk
to devices like QCA6390.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Manivannan Sadhasivam [Thu, 7 May 2020 12:53:06 +0000 (18:23 +0530)]
net: qrtr: Do not depend on ARCH_QCOM
IPC Router protocol is also used by external modems for exchanging the QMI
messages. Hence, it doesn't always depend on Qualcomm platforms. One such
instance is the QCA6390 WLAN device connected to x86 machine.
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Manivannan Sadhasivam [Thu, 7 May 2020 12:53:05 +0000 (18:23 +0530)]
net: qrtr: Add MHI transport layer
MHI is the transport layer used for communicating to the external modems.
Hence, this commit adds MHI transport layer support to QRTR for
transferring the QMI messages over IPC Router.
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Geert Uytterhoeven [Thu, 7 May 2020 11:42:05 +0000 (13:42 +0200)]
via-rhine: Add platform dependencies
The VIA Rhine Ethernet interface is only present on PCI devices or
VIA/WonderMedia VT8500/WM85xx SoCs. Add platform dependencies to the
VIA_RHINE config symbol, to avoid asking the user about it when
configuring a kernel without PCI or VT8500/WM85xx support.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: David S. Miller <davem@davemloft.net>
Po Liu [Thu, 7 May 2020 10:57:38 +0000 (18:57 +0800)]
net:enetc: bug fix for qos sfi operate space after freed
'Dan Carpenter' reported:
This code frees "sfi" and then dereferences it on the next line:
> kfree(sfi);
> clear_bit(sfi->index, epsfp.psfp_sfi_bitmap);
This "sfi->index" should be "index".
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Po Liu <Po.Liu@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / users / jedix / linux-maple.git / log