These are two new handlers for MLO. As the configurations done in
bss_info_changed() are now split into two separate flows, use MLO
specific implementation instead of common functions with the non-MLO
code.
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104949.7b238cae0895.Ieb87f204787fb1c7cb7562e1cbf54ef518d87123@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
In newer firmware APIs the firmware is responsible for tracking
the DTIM period and other beacon timing, so we don't need to
wait with setting associated. In real MLO operation, mac80211
isn't tracking this anyway, and connections wouldn't work.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104949.02354241fede.Id957bed3851fdf1fe902d79a1b0338c6d80bc0e1@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The firmware now allows adding a link that's not yet bound
to a PHY context. Make use of that to align the driver with
mac80211's API expectations. For now, just add the link at
the same time as the MAC since we don't yet have real MLD
support, but that'll obviously change later. This fixes an
issue with apStaId tracking in the firmware.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104949.097e5008b637.I4e75c6c11e21c08d28ff6a066be36629d3975db6@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
HE/EHT support is reported to FW if there's at least one link
that supports it. Configure beacon separately for each link.
Don't send the beacon template before adding the MAC.
Co-developed-by: Johannes Berg <johannes.berg@intel.com>
Co-developed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104949.5ef4efeda2dd.I6ebda2b71c964b9aa63240c9fa1ee0d28099fe6e@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The LINK cmd host api has been updated. Align the driver to the
new changes. Also, temporary use mac_id for link_id.
Using the phy_id as the link_id is wrong since we might have 2 macs
operating on the same phy - in this case we will have 2 different
links (one for each mac) with the same link_id. On the other hand,
since we don't have MLO implemented yet, we won't have 2 different
links of the same mac. Therefore, we can use the mac_id as the
link_id.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.78ae716884fe.Icfeb2794d9652baaccf9b0cdddbd751d0db4f952@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Currently we're setting the sta->addr as the peers address only if
the iftype is NL80211_IFTYPE_AP, otherwise we are setting the bssid to
be the addr. This causes bugs in TDLS. Fix this by always using the
sta->addr.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.4c104c3074c4.I78912bb85251033e60db99a65165890779203612@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
During CSA the PHY used by the link is changing, So the driver
needs to modify the links phy to the FW. Currently the driver is doing
it by removing the link and adding a new one with the new phy_id, but the
FW expects the link only to be modified. Fix this by modifying the links
phy_id instead of removing it and adding a new one
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.c07ca7aace29.I4ed5c77f4afe1b5da19322734e2f84d51aa541ad@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The patch linked below fixes the crash on queue removal bug only
for the non-MLD API. Do the same for the MLD API.
Fixes: c5a976cf6a ("wifi: iwlwifi: modify new queue allocation command")
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.527dace26147.Ia215df5833634f95688a979f39fae70c1ac4e027@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
This bitmap indicates what fields of the cmd got changed.
A field will be ignored by the FW if the corresponding flag wasn't set.
There are a few cases in which we currently set the wrong bits when
sending this cmd, which caused FW asserts. Fix this by setting the
correct bits in each case.
Fixes: 1ab2663233 ("wifi: iwlwifi: mvm: Add an add_interface() callback for mld mode")
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.19ddbee0c98d.I595abb79d0419c9a21e5234303c2c3fd5290a52a@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
These flows are the same in both MLD API and the current API,
except for the commands that are being sent during this flows.
Instead of checking each time before calling these floews
what API we use and then call the correct function, call always the old
one, which in turn will call the new one in case we're using the MLD
API.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.5692d8dea9be.Ib1882b2c2f0b0603abc4b7d4a0ecc45cd1fbf9a7@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
This flow is almost the same for both MLD and non-MLD modes,
except for some function calls. Therefore there is no reason to
add an MLD version of this flow. Instead - put the parts that are unique
for each mode in helper functions, and in the next patch each version of
this flow will call the common part with pointers to its specific
helper functions.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.61bc077a7f3c.Ia3aa81d3293792bf8f80528dbc67a711ce334b32@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
In MLD mode we have a new STA cmd. As a result, it is also changes
the flows of adding/updating/removing and handling state of
a station. Add these flows.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.b5548cfd8fe3.I70f9c8f3c95e18d5c9af0a5681e0830893509531@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
WE can't mix between the new MLD API and the old API.
I.e. - we can't send one of the new cmds and then one of the old ones.
This will cause a FW assert. So we need an indication what API should be
used. We use the new API if:
1. FW supports it
2. We are registered to mac80211 with the new MLD ops
Add an indication which will only be true if both conditions are true.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.5756b0907403.I0adce36d1783cce23d0e080e3c4a8953db33b515@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Split iwl_mvm_sta into general and link specific parts. As a first
step, all link dependent parameters reside in deflink.
The change was done mostly using the spatch below with some manual
adjustments.
@iwl_mvm_sta@
struct iwl_mvm_sta *s;
identifier var = {sta_id, lq_sta, avg_energy};
@@
(
s->
- var
+ deflink.var
)
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.34eace06d583.I1f8c5e919a71b21030460fbdd220d42401b688b1@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
In MLO, some fields of iwl_mvm_vif should be defined in the
context of a link. Define a separate structure for these fields and
add a deflink object to hold it as part of iwl_mvm_vif. Non-MLO legacy
code will use only deflink object while MLO related code will use the
corresponding link from the link array.
It follows the strategy applied in mac80211 for introducing MLO
changes.
The below spatch takes care of updating all driver code to access
fields separated into MLD specific data structure via deflink (need
to convert all references to the fields listed in var to deflink.var
and also to take care of calls like iwl_mvm_vif_from_mac80211(vif)->field).
@iwl_mld_vif@
struct iwl_mvm_vif *v;
struct ieee80211_vif *vv;
identifier fn;
identifier var = {bssid, ap_sta_id, bcast_sta, mcast_sta,
beacon_stats, smps_requests, probe_resp_data,
he_ru_2mhz_block, cab_queue, phy_ctxt,
queue_params};
@@
(
v->
- var
+ deflink.var
|
fn(vv)->
- var
+ deflink.var
)
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230328104948.4896576f0a9f.Ifaf0187c96b9fe52b24bd629331165831a877691@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Some extreme test environment may keep IC temperature very low or very high
during system boot stage. For stability concern, we add thermal voltage
compenstation if needed no matter svs bank phase is in init02 or mon mode.
Signed-off-by: Roger Lu <roger.lu@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Link: https://lore.kernel.org/r/20230202124104.16504-4-roger.lu@mediatek.com
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
clang with W=1 reports
net/mac80211/rc80211_minstrel_ht.c:1711:6: error: variable
'n_supported' set but not used [-Werror,-Wunused-but-set-variable]
int n_supported = 0;
^
This variable is not used so remove it.
Signed-off-by: Tom Rix <trix@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230325132610.1334820-1-trix@redhat.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Avoid potential data corruption issues caused by uninitialized driver
private data structures.
Reported-by: Brian Coverstone <brian@mainsequence.net>
Fixes: 6a9d1b91f3 ("mac80211: add pre-RCU-sync sta removal driver operation")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230324120924.38412-3-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Adjust the network header to point at the correct payload offset
Fixes: 986e43b19a ("wifi: mac80211: fix receiving A-MSDU frames on mesh interfaces")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230324120924.38412-2-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
When forwarding is set to 0, frames are typically sent with ttl=1.
Move the ttl decrement check below the check for local receive in order to
fix packet drops.
Reported-by: Thomas Hühn <thomas.huehn@hs-nordhausen.de>
Reported-by: Nick Hainke <vincent@systemli.org>
Fixes: 986e43b19a ("wifi: mac80211: fix receiving A-MSDU frames on mesh interfaces")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230326151709.17743-1-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
There is a 64-bit division in iwl_mvm_get_crosstimestamp_fw(), which
results in a link failure when building 32-bit architectures with clang:
ld.lld: error: undefined symbol: __udivdi3
>>> referenced by ptp.c
>>> drivers/net/wireless/intel/iwlwifi/mvm/ptp.o:(iwl_mvm_phc_get_crosstimestamp) in archive vmlinux.a
GCC has optimizations for division by a constant that clang does not
implement, so this issue is not visible when building with GCC.
Use the 64-bit division helper div_u64(), which takes a u64 dividend and
u32 divisor, which matches this situation and prevents the emission of a
libcall for the division.
Fixes: 21fb8da6eb ("wifi: iwlwifi: mvm: read synced time from firmware if supported")
Reported-by: Arnd Bergmann <arnd@arndb.de>
Link: https://github.com/ClangBuiltLinux/linux/issues/1826
Reported-by: "kernelci.org bot" <bot@kernelci.org>
Link: https://lore.kernel.org/6423173a.620a0220.3d5cc.6358@mx.google.com/
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
rx->sta->amsdu_mesh_control is being passed to ieee80211_amsdu_to_8023s
without checking rx->sta. Since it doesn't make sense to accept A-MSDU
packets without a sta, simply add a check earlier.
Fixes: 6e4c0d0460 ("wifi: mac80211: add a workaround for receiving non-standard mesh A-MSDU")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230330090001.60750-2-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The logic in acpi_is_processor_usable() requires the online capable
bit be set for hotpluggable CPUs. The online capable bit has been
introduced in ACPI 6.3.
However, for ACPI revisions < 6.3 which do not support that bit, CPUs
should be reported as usable, not the other way around.
Reverse the check.
[ bp: Rewrite commit message. ]
Fixes: e2869bd7af ("x86/acpi/boot: Do not register processors that cannot be onlined for x2APIC")
Suggested-by: Miguel Luis <miguel.luis@oracle.com>
Suggested-by: Boris Ostrovsky <boris.ovstrosky@oracle.com>
Signed-off-by: Eric DeVolder <eric.devolder@oracle.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Tested-by: David R <david@unsolicited.net>
Cc: <stable@kernel.org>
Link: https://lore.kernel.org/r/20230327191026.3454-2-eric.devolder@oracle.com
ACPI 6.3 introduced the online capable bit, and also introduced MADT
version 5.
Latter was used to distinguish whether the offset storing online capable
could be used. However ACPI 6.2b has MADT version "45" which is for
an errata version of the ACPI 6.2 spec. This means that the Linux code
for detecting availability of MADT will mistakenly flag ACPI 6.2b as
supporting online capable which is inaccurate as it's an ACPI 6.3 feature.
Instead use the FADT major and minor revision fields to distinguish this.
[ bp: Massage. ]
Fixes: aa06e20f1b ("x86/ACPI: Don't add CPUs that are not online capable")
Reported-by: Eric DeVolder <eric.devolder@oracle.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@kernel.org>
Link: https://lore.kernel.org/r/943d2445-84df-d939-f578-5d8240d342cc@unsolicited.net
Arseniy Krasnov says:
====================
fix header length on skb merging
this patchset fixes appending newly arrived skbuff to the last skbuff of
the socket's queue during rx path. Problem fires when we are trying to
append data to skbuff which was already processed in dequeue callback
at least once. Dequeue callback calls function 'skb_pull()' which changes
'skb->len'. In current implementation 'skb->len' is used to update length
in header of last skbuff after new data was copied to it. This is bug,
because value in header is used to calculate 'rx_bytes'/'fwd_cnt' and
thus must be constant during skbuff lifetime. Here is example, we have
two skbuffs: skb0 with length 10 and skb1 with length 4.
1) skb0 arrives, hdr->len == skb->len == 10, rx_bytes == 10
2) Read 3 bytes from skb0, skb->len == 7, hdr->len == 10, rx_bytes == 10
3) skb1 arrives, hdr->len == skb->len == 4, rx_bytes == 14
4) Append skb1 to skb0, skb0 now has skb->len == 11, hdr->len == 11.
But value of 11 in header is invalid.
5) Read whole skb0, update rx_bytes by 11 from skb0's header.
6) At this moment rx_bytes == 3, but socket's queue is empty.
This bug starts to fire since:
commit
0777061657 ("virtio/vsock: don't use skbuff state to account credit")
In fact, it presents before, but didn't triggered due to a little bit
buggy implementation of credit calculation logic. So i'll use Fixes tag
for it.
I really forgot about this branch in rx path when implemented patch
0777061657.
This patchset contains 3 patches:
1) Fix itself.
2) Patch with WARN_ONCE() to catch such problems in future.
3) Patch with test which triggers skb appending logic. It looks like
simple test with several 'send()' and 'recv()', but it checks, that
skbuff appending works ok.
====================
Link: https://lore.kernel.org/r/0683cc6e-5130-484c-1105-ef2eb792d355@sberdevices.ru
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
This adds test which checks case when data of newly received skbuff is
appended to the last skbuff in the socket's queue. It looks like simple
test with 'send()' and 'recv()', but internally it triggers logic which
appends one received skbuff to another. Test checks that this feature
works correctly.
This test is actual only for virtio transport.
Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
This adds WARN_ONCE() and return from stream dequeue callback when
socket's queue is empty, but 'rx_bytes' still non-zero. This allows
the detection of potential bugs due to packet merging (see previous
patch).
Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
This fixes appending newly arrived skbuff to the last skbuff of the
socket's queue. Problem fires when we are trying to append data to skbuff
which was already processed in dequeue callback at least once. Dequeue
callback calls function 'skb_pull()' which changes 'skb->len'. In current
implementation 'skb->len' is used to update length in header of the last
skbuff after new data was copied to it. This is bug, because value in
header is used to calculate 'rx_bytes'/'fwd_cnt' and thus must be not
be changed during skbuff's lifetime.
Bug starts to fire since:
commit 0777061657
("virtio/vsock: don't use skbuff state to account credit")
It presents before, but didn't triggered due to a little bit buggy
implementation of credit calculation logic. So use Fixes tag for it.
Fixes: 0777061657 ("virtio/vsock: don't use skbuff state to account credit")
Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The mISDN_class_release() is not needed at all, as the class structure
is static, and it does not actually do anything either, so it is safe to
remove as struct class does not require a release callback.
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20230329060127.2688492-1-gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Override the PMIC-default voltage constraints for VGPU and VSRAM_GPU
with the platform specific vmin/vmax for the highest possible SoC
binning.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Suggested-by: Chen-Yu Tsai <wenst@chromium.org>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Link: https://lore.kernel.org/r/20230301095523.428461-20-angelogioacchino.delregno@collabora.com
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
Override the PMIC-default voltage constraints for VGPU and VSRAM_GPU
with the platform specific vmin/vmax for the highest possible SoC
binning.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Suggested-by: Chen-Yu Tsai <wenst@chromium.org>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Link: https://lore.kernel.org/r/20230301095523.428461-19-angelogioacchino.delregno@collabora.com
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
Add a GPU node for MT8186 SoC but keep it disabled.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Link: https://lore.kernel.org/r/20230301095523.428461-18-angelogioacchino.delregno@collabora.com
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
Enable the Mali-G57 found on this platform with the open-source
Panfrost driver.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Link: https://lore.kernel.org/r/20230301095523.428461-17-angelogioacchino.delregno@collabora.com
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
Add GPU support through panfrost for the Mali-G57 GPU on MT8195
with its OPP table but keep it in disabled state.
This is expected to be enabled only on boards which make use of
the GPU.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Link: https://lore.kernel.org/r/20230301095523.428461-16-angelogioacchino.delregno@collabora.com
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com>
