QCA2066 is a PCI based DBS device. It is very similar to WCN6855
overall: they share the same PCI device ID, the same major and
minor version numbers, the same register address, and same HAL
descriptors etc. The most significant difference is that QCA2066
supports 3-antenna configuration while WCN6855 does not. To differentiate
them, subversion numbers are used. Currently four numbers are used
by QCA2066: 0x1019A0E1, 0x1019B0E1, 0x1019C0E1 and 0x1019D0E1.
Tested-on: QCA2066 hw2.1 PCI WLAN.HSP.1.1-03737-QCAHSPSWPL_V2_SILICONZ_CE-1
Tested-on: WCN6855 hw2.1 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3
Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240109021336.4143-3-quic_bqiang@quicinc.com
In ath11k_pci_probe() there is a switch statement that, based
upon the PCI device ID, assigns pci_ops. After the switch,
ath11k_pcic_register_pci_ops() is called to register the pci_ops.
Unfortunately, this registration is too late if any of the cases
in the switch need to perform operations that require the pci_ops
to already be registered. In particular, an upcoming patch for
QCA2066 needs to call ath11k_pcic_read32().
To address this issue, call ath11k_pcic_register_pci_ops() from
each case instead of doing so after the switch. That way the ops
will be registered if any subsequent operations within the case
processing require the ops to be present.
Tested-on: WCN6855 hw2.1 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3
Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240109021336.4143-2-quic_bqiang@quicinc.com
Provide address list to mac80211 so user doesn't need to specify addresses when
a second interface is added because the address can be allocated from the list
by mac80211.
The derived addresses have LAA (Local Administered Address) bit set, and only
the first byte is changed. Take the 00:03:7f:xx:xx:xx as example to derive:
addresses[0] is unchanged, it's still 00:03:7f:xx:xx:xx,
addresses[1] is 02:03:7f:xx:xx:xx,
addresses[2] is 12:03:7f:xx:xx:xx,
addresses[3] is 22:03:7f:xx:xx:xx,
addresses[4] is 32:03:7f:xx:xx:xx.
However as only 3 addresses are reported now, so addresses[3] and addresses[4]
aren't actually derived.
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3
Signed-off-by: Carl Huang <quic_cjhuang@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20230714023801.2621802-3-quic_cjhuang@quicinc.com
Add hardware parameter support_dual_stations to indicate whether 2 station
interfaces are supported. For chips which support this feature, limit total
number of AP interface and mesh point to 1. The max interfaces are 3 for such
chips.
The chips affected are:
QCA6390 hw2.0
WCN6855 hw2.0
WCN6855 hw2.1
Other chips are not affected.
For affected chips, remove radar_detect_widths because now
num_different_channels is set to 2. radar_detect_widths can be set only when
num_different_channels is 1, see mac80211 function wiphy_verify_combinations
for details. This means that in affectected chips DFS cannot be enabled in AP
mode.
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3
Signed-off-by: Carl Huang <quic_cjhuang@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20230714023801.2621802-2-quic_cjhuang@quicinc.com
Specify v2m_serial0 as the device for boot console output through the
stdout-path in the chosen node.
Signed-off-by: Ole P. Orhagen <ole.orhagen@northern.tech>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20240209112807.1345164-2-ole.orhagen@northern.tech
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
And use the stored addresses in rtl8192cu instead of hardcoding them.
This is what the vendor drivers do.
Perhaps this is not strictly necessary for RTL8192CU devices. However,
the dual mac version of RTL8192DU has two USB interfaces, each with its
own set of endpoints. Hardcoding their addresses in the upcoming
rtl8192du driver would require making some assumptions which I'm not
qualified to make.
Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/7b6a602a-6101-4bab-958d-bcff4d565b40@gmail.com
rtl8192cu handles 1T1R devices (RTL8188CUS), 1T2R devices (RTL8191CU),
and 2T2R devices (RTL8192CU). The 2T2R devices were incorrectly detected
as 1T2R because of a mistake in the IS_92C_1T2R macro.
The visible effect of this is that the firmware was allowed to use
TX rates only up to MCS7.
Fix the IS_92C_1T2R macro.
Now my 2T2R device has much better upload speed.
Before: 46 Mbps.
After: 82 Mbps.
Also fix a debug message which was printing "RF_1T1R" even for 1T2R
chips.
Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/ed960059-5c77-422d-ac4e-fe9fc9d0d296@gmail.com
The rcu_read_unlock() is accidentally added, and sparse warn:
drivers/net/wireless/realtek/rtw89/fw.c:2807:17:
warning: context imbalance in 'rtw89_fw_h2c_assoc_cmac_tbl_g7' - unexpected unlock
Fixes: b82730bf57 ("wifi: cfg80211/mac80211: move puncturing into chandef")
Cc: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240213122556.9593-1-pkshih@realtek.com
We are going to allow MCC (multi-channel concurrency) on RTL8922A.
So, increase 8922a::support_chanctx_num up to 2 first.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240213073514.23796-6-pkshih@realtek.com
On Wi-Fi 7 chips, concurrent stuffs are supported by FW MRC series
(multi-role concurrent) functions. And, driver has implemented the
corresponding SW handling in patches in front of this one. Now, we
extend SW MCC (multi-channel concurrent) flow to work on Wi-Fi 7
chips.
In SW point of view, things look as below.
| SW | | FW func |
| | | H2C/C2H |
--------------------------------------------
| | ax |
| | /----| FW MCC func |
| MCC | -- chip --+ |
| | \----| FW MRC func |
| | be |
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240213073514.23796-5-pkshih@realtek.com
Implement MRC (multiple role concurrent) H2C commands. Mainly deal with
H2C format, LE type built from CPU value, default setting on some fields,
and then sending the command to FW. Besides, MRC start, MRC delete, and
MRC request TSF need to wait for a report from C2H events.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240213073514.23796-4-pkshih@realtek.com
Add handling of MRC (multiple role concurrent) C2H events including
TSF report and status report. Parse report data and then complete the
corresponding H2C commands, which will be implemented in the following.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240213073514.23796-3-pkshih@realtek.com
For Wi-Fi 7 chips, FW supports MRC (multi-role concurrent) functions
including H2C commands and C2H events. We can consider FW MRC functions
as a superset of FW MCC (multi-channel concurrent) functions. And, MRC
functions can take MLO things into account.
Basically before MLO, SW can also manipulate FW MRC to work original
SW MCC flow. So, we add them first and implement the handling in the
following. And then, SW MCC will call different series of FW functions
according to chip later.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240213073514.23796-2-pkshih@realtek.com
Previously, the driver created a net device (typically wlan0) as soon
as the module was loaded. This commit changes the driver to follow
normal Linux convention of creating the net device only when bus
probing detects a supported chip.
Signed-off-by: David Mosberger-Tang <davidm@egauge.net>
Tested-By: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240212202057.3468714-1-davidm@egauge.net
Use id-gpios and vbus-gpios instead.
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Acked-by: Shawn Guo <shawnguo@kernel.org>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20240215093214.796821-3-alexander.stein@ew.tq-group.com
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
wilc_netdev_cleanup currently triggers a KASAN warning, which can be
observed on interface registration error path, or simply by
removing the module/unbinding device from driver:
echo spi0.1 > /sys/bus/spi/drivers/wilc1000_spi/unbind
==================================================================
BUG: KASAN: slab-use-after-free in wilc_netdev_cleanup+0x508/0x5cc
Read of size 4 at addr c54d1ce8 by task sh/86
CPU: 0 PID: 86 Comm: sh Not tainted 6.8.0-rc1+ #117
Hardware name: Atmel SAMA5
unwind_backtrace from show_stack+0x18/0x1c
show_stack from dump_stack_lvl+0x34/0x58
dump_stack_lvl from print_report+0x154/0x500
print_report from kasan_report+0xac/0xd8
kasan_report from wilc_netdev_cleanup+0x508/0x5cc
wilc_netdev_cleanup from wilc_bus_remove+0xc8/0xec
wilc_bus_remove from spi_remove+0x8c/0xac
spi_remove from device_release_driver_internal+0x434/0x5f8
device_release_driver_internal from unbind_store+0xbc/0x108
unbind_store from kernfs_fop_write_iter+0x398/0x584
kernfs_fop_write_iter from vfs_write+0x728/0xf88
vfs_write from ksys_write+0x110/0x1e4
ksys_write from ret_fast_syscall+0x0/0x1c
[...]
Allocated by task 1:
kasan_save_track+0x30/0x5c
__kasan_kmalloc+0x8c/0x94
__kmalloc_node+0x1cc/0x3e4
kvmalloc_node+0x48/0x180
alloc_netdev_mqs+0x68/0x11dc
alloc_etherdev_mqs+0x28/0x34
wilc_netdev_ifc_init+0x34/0x8ec
wilc_cfg80211_init+0x690/0x910
wilc_bus_probe+0xe0/0x4a0
spi_probe+0x158/0x1b0
really_probe+0x270/0xdf4
__driver_probe_device+0x1dc/0x580
driver_probe_device+0x60/0x140
__driver_attach+0x228/0x5d4
bus_for_each_dev+0x13c/0x1a8
bus_add_driver+0x2a0/0x608
driver_register+0x24c/0x578
do_one_initcall+0x180/0x310
kernel_init_freeable+0x424/0x484
kernel_init+0x20/0x148
ret_from_fork+0x14/0x28
Freed by task 86:
kasan_save_track+0x30/0x5c
kasan_save_free_info+0x38/0x58
__kasan_slab_free+0xe4/0x140
kfree+0xb0/0x238
device_release+0xc0/0x2a8
kobject_put+0x1d4/0x46c
netdev_run_todo+0x8fc/0x11d0
wilc_netdev_cleanup+0x1e4/0x5cc
wilc_bus_remove+0xc8/0xec
spi_remove+0x8c/0xac
device_release_driver_internal+0x434/0x5f8
unbind_store+0xbc/0x108
kernfs_fop_write_iter+0x398/0x584
vfs_write+0x728/0xf88
ksys_write+0x110/0x1e4
ret_fast_syscall+0x0/0x1c
[...]
David Mosberger-Tan initial investigation [1] showed that this
use-after-free is due to netdevice unregistration during vif list
traversal. When unregistering a net device, since the needs_free_netdev has
been set to true during registration, the netdevice object is also freed,
and as a consequence, the corresponding vif object too, since it is
attached to it as private netdevice data. The next occurrence of the loop
then tries to access freed vif pointer to the list to move forward in the
list.
Fix this use-after-free thanks to two mechanisms:
- navigate in the list with list_for_each_entry_safe, which allows to
safely modify the list as we go through each element. For each element,
remove it from the list with list_del_rcu
- make sure to wait for RCU grace period end after each vif removal to make
sure it is safe to free the corresponding vif too (through
unregister_netdev)
Since we are in a RCU "modifier" path (not a "reader" path), and because
such path is expected not to be concurrent to any other modifier (we are
using the vif_mutex lock), we do not need to use RCU list API, that's why
we can benefit from list_for_each_entry_safe.
[1] https://lore.kernel.org/linux-wireless/ab077dbe58b1ea5de0a3b2ca21f275a07af967d2.camel@egauge.net/
Fixes: 8399918f30 ("staging: wilc1000: use RCU list to maintain vif interfaces list")
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240212-wilc_rework_deinit-v1-1-9203ae56c27f@bootlin.com
When SOF_TIMESTAMPING_OPT_ID is used to ambiguate timestamped datagrams,
the sk_tskey can become unpredictable in case of any error happened
during sendmsg(). Move increment later in the code and make decrement of
sk_tskey in error path. This solution is still racy in case of multiple
threads doing snedmsg() over the very same socket in parallel, but still
makes error path much more predictable.
Fixes: 09c2d251b7 ("net-timestamp: add key to disambiguate concurrent datagrams")
Reported-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Vadim Fedorenko <vadfed@meta.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20240213110428.1681540-1-vadfed@meta.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Currently, we are checking whether the PHY package mode matches the
individual PHY interface modes at PHY package probe time, but at that time
we only know the PHY package mode and not the individual PHY interface
modes as of_get_phy_mode() that populates it will only get called once the
netdev to which PHY-s are attached to is being probed and thus this check
will always fail and return -EINVAL.
So, lets move this check to .config_init_once as at that point individual
PHY interface modes should be populated.
Fixes: d1cb613efb ("net: phy: qcom: add support for QCA807x PHY Family")
Signed-off-by: Robert Marko <robimarko@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20240212115043.1725918-1-robimarko@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Before consolidation of commit 4965a68780 ("arch: define the
ARCH_DMA_ADDR_T_64BIT config symbol in lib/Kconfig"), the config ARCH_ATU
was used to control the state of the config ARCH_DMA_ADDR_T_64BIT. After
this consolidation, the config ARCH_ATU has been without use and effect.
Remove this obsolete config.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andreas Larsson <andreas@gaisler.com>
Link: https://lore.kernel.org/r/20231211083029.22078-1-lukas.bulwahn@gmail.com
It's currently possible to change the mesh ID when the
interface isn't yet in mesh mode, at the same time as
changing it into mesh mode. This leads to an overwrite
of data in the wdev->u union for the interface type it
currently has, causing cfg80211_change_iface() to do
wrong things when switching.
We could probably allow setting an interface to mesh
while setting the mesh ID at the same time by doing a
different order of operations here, but realistically
there's no userspace that's going to do this, so just
disallow changes in iftype when setting mesh ID.
Cc: stable@vger.kernel.org
Fixes: 29cbe68c51 ("cfg80211/mac80211: add mesh join/leave commands")
Reported-by: syzbot+dd4779978217b1973180@syzkaller.appspotmail.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Even if that's the same as IEEE80211_MAX_SSID_LEN, we really
should just use IEEE80211_MAX_MESH_ID_LEN for mesh, rather
than having the BUILD_BUG_ON()s.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
mac80211 might (due to an unavoidable race) cancel a ROC that has already
expired. In that case the driver should not send the session protection
cmd to cancel the ROC.
When session protection is supported, the te_data::id field is reused
to save the configuration id. Check it before sending the cmd.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240205211151.30176bf869d9.Id811c20d3746b870cbe0c946bbfe1c0ab0a290cb@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
This is set when a P2P ROC ends, and uses as an indication inside
iwl_mvm_roc_done_wk that the resources used for this ROC (sta/link)
needs to be flushed/deactivated (respectively).
But we also have IWL_MVM_STATUS_ROC_RUNNING, which is set whenever
P2P ROC starts, and is not even used in iwl_mvm_roc_done_wk.
Use IWL_MVM_STATUS_ROC_RUNNING as an indicator, and remove the redundant
bit.
While at it, add a call to synchronize_net also for the
AUX ROC case, which is missing in the existing code.
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240205211151.0494f75de311.Ic4aacacf7581a5c9046c4f1df87cbb67470853e7@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
As per section 35.3.1 Starting Reception for the MIPI CSI-2 Input on the
latest hardware manual (R01UH0914EJ0145 Rev.1.45) we need to supply all
the clocks and then release the CRU resets. Currently we are releasing
the resets and then supplying the clocks. So, fix the start reception
procedure.
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Link: https://lore.kernel.org/r/20240213181233.242316-6-biju.das.jz@bp.renesas.com
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
As per section 35.3.1 Starting Reception for the MIPI CSI-2 Input on the
latest hardware manual (R01UH0914EJ0145 Rev.1.45) we need to disable the
vclk before enabling the LINK reception and enable the vclk after enabling
the link Reception. So restructure vclk handling as per the HW manual.
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Link: https://lore.kernel.org/r/20240213181233.242316-5-biju.das.jz@bp.renesas.com
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
As per section 35.3.1 Starting Reception for the MIPI CSI-2 Input on the
latest hardware manual (R01UH0914EJ0140 Rev.1.40) it is mentioned that
initialize the AXI master first and then initialize the image processing.
Fix the start procedure as per the hardware manual.
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Link: https://lore.kernel.org/r/20240213181233.242316-4-biju.das.jz@bp.renesas.com
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
As per section 35.3.1 Starting Reception for the MIPI CSI-2 Input on the
latest hardware manual (R01UH0914EJ0140 Rev.1.40) it is mentioned that
after DPHY reset, we need to wait for 1 msec or more before start
receiving data from the sensor. So add a delay after pre_streamon().
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Link: https://lore.kernel.org/r/20240213181233.242316-3-biju.das.jz@bp.renesas.com
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Replace the old SET_RUNTIME_PM_OPS() helpers with its modern alternative
RUNTIME_PM_OPS(). The usage of pm_ptr and RUNTIME_PM_OPS() allows the
compiler to see where it's used but still drop the dead code. After this
we can get rid of the unnecessary '__maybe_unused' annotations on PM
functions.
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Link: https://lore.kernel.org/r/20240213181233.242316-2-biju.das.jz@bp.renesas.com
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
There's a conflict already and some upcoming changes
also depend on changes in wireless for being conflict-
free, so pull wireless in to make all that easier.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
clang-16 warns about a cast between incompatible function types:
drivers/gpu/drm/xe/xe_range_fence.c:155:10: error: cast from 'void (*)(const void *)' to 'void (*)(struct xe_range_fence *)' converts to incompatible function type [-Werror,-Wcast-function-type-strict]
155 | .free = (void (*)(struct xe_range_fence *rfence)) kfree,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Avoid this with a trivial helper function that calls kfree() here.
v2:
- s/* rfence/*rfence/ (Thomas)
Fixes: 845f64bdbf ("drm/xe: Introduce a range-fence utility")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240213095719.454865-1-arnd@kernel.org
(cherry picked from commit f2c9364db5)
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
shmem ensures the memory is cleared on allocation, however here we are
using TTM, which doesn't natively support shmem (other than for swap),
but instead just allocates normal system memory. And we only zero such
memory for userspace allocations. In the case of intel_fbdev we are
missing the memset_io() since display path incorrectly thinks object is
shmem based.
Fixes: 44e694958b ("drm/xe/display: Implement display support")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Suraj Kandpal <suraj.kandpal@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240205153110.38340-2-matthew.auld@intel.com
(cherry picked from commit 63fb531fbf)
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
The function xe_vm_prepare_vma was blindly accepting zero as the
number of fences and forwarded that to drm_exec_prepare_obj.
However, that leads to an out-of-bounds shift in the
dma_resv_reserve_fences() and while one could argue that the
dma_resv code should be robust against that, avoid attempting
to reserve zero fences.
Relevant stack trace:
[773.183188] ------------[ cut here ]------------
[773.183199] UBSAN: shift-out-of-bounds in ../include/linux/log2.h:57:13
[773.183241] shift exponent 64 is too large for 64-bit type 'long unsigned int'
[773.183254] CPU: 2 PID: 1816 Comm: xe_evict Tainted: G U 6.8.0-rc3-xe #1
[773.183256] Hardware name: ASUS System Product Name/PRIME Z690-P D4, BIOS 2014 10/14/2022
[773.183257] Call Trace:
[773.183258] <TASK>
[773.183260] dump_stack_lvl+0xaf/0xd0
[773.183266] dump_stack+0x10/0x20
[773.183283] ubsan_epilogue+0x9/0x40
[773.183286] __ubsan_handle_shift_out_of_bounds+0x10f/0x170
[773.183293] dma_resv_reserve_fences.cold+0x2b/0x48
[773.183295] ? ww_mutex_lock+0x3c/0x110
[773.183301] drm_exec_prepare_obj+0x45/0x60 [drm_exec]
[773.183313] xe_vm_prepare_vma+0x33/0x70 [xe]
[773.183375] xe_vma_destroy_unlocked+0x55/0xa0 [xe]
[773.183427] xe_vm_close_and_put+0x526/0x940 [xe]
Fixes: 2714d50936 ("drm/xe: Convert pagefaulting code to use drm_exec")
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240208132115.3132-1-thomas.hellstrom@linux.intel.com
(cherry picked from commit eb538b5574)
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
It is still possible to compile-test a kernel without CONFIG_COMMON_CLK
for some ancient ARM boards or other architectures, but this causes a
link failure in the qcom-cpufreq-hw driver:
ERROR: modpost: "devm_clk_hw_register" [drivers/cpufreq/qcom-cpufreq-hw.ko] undefined!
ERROR: modpost: "devm_of_clk_add_hw_provider" [drivers/cpufreq/qcom-cpufreq-hw.ko] undefined!
ERROR: modpost: "of_clk_hw_onecell_get" [drivers/cpufreq/qcom-cpufreq-hw.ko] undefined!
Add a Kconfig dependency here to make sure this always work. Apparently
this bug has been in the kernel for a while without me running into it
on randconfig builds as COMMON_CLK is almost always enabled.
I have cross-checked by building an allmodconfig kernel with COMMON_CLK
disabled, which showed no other driver having this problem.
Fixes: 4370232c72 ("cpufreq: qcom-hw: Add CPU clock provider support")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
clang-16 warns about a cast between incompatible function types:
drivers/gpu/drm/xe/xe_range_fence.c:155:10: error: cast from 'void (*)(const void *)' to 'void (*)(struct xe_range_fence *)' converts to incompatible function type [-Werror,-Wcast-function-type-strict]
155 | .free = (void (*)(struct xe_range_fence *rfence)) kfree,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Avoid this with a trivial helper function that calls kfree() here.
v2:
- s/* rfence/*rfence/ (Thomas)
Fixes: 845f64bdbf ("drm/xe: Introduce a range-fence utility")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240213095719.454865-1-arnd@kernel.org
The documentation for default_values mentions high/low which can be
confusing, particularly when the ACTIVE_LOW flag is set.
Replace high/low with active/inactive to clarify that the values are
logical not physical.
Similarly, clarify the interpretation of values in struct gpiohandle_data.
Signed-off-by: Kent Gibson <warthog618@gmail.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
There are two legacy, deprecated functions - gpiod_to_chip() and
gpio_device_get_chip() - that still have users in tree. They return the
address of the SRCU-protected chip outside of the read-only critical
sections. They are inherently dangerous and the users should convert to
safer alternatives. Let's explicitly silence lockdep warnings by using
rcu_dereference_check(ptr, 1). While at it: reuse
gpio_device_get_chip() in gpiod_to_chip().
Fixes: d83cee3d2b ("gpio: protect the pointer to gpio_chip in gpio_device with SRCU")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202402122234.d85cca9b-lkp@intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Lockdep with CONFIG_PROVE_RCU enabled reports false positives about
suspicious rcu_dereference() usage. Let's silence it by using
srcu_dereference() which is the correct helper with SRCU.
Fixes: d83cee3d2b ("gpio: protect the pointer to gpio_chip in gpio_device with SRCU")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202402122234.d85cca9b-lkp@intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
We never dereference the chip pointer in character device code so we can
use the lighter rcu_access_pointer() helper. This also makes lockep
happier as it no longer complains about suspicious rcu_dereference()
usage.
Fixes: d83cee3d2b ("gpio: protect the pointer to gpio_chip in gpio_device with SRCU")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202402122234.d85cca9b-lkp@intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
gpiod_hog() may be called without the gpio_device SRCU read lock taken
so we need to do it here as well. It's alright if someone else is
already holding the lock as SRCU read critical sections can be nested.
Fixes: d83cee3d2b ("gpio: protect the pointer to gpio_chip in gpio_device with SRCU")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202402122234.d85cca9b-lkp@intel.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
The Snapdragon 670 uses the Qualcomm driver for CPU frequency scaling.
Block this driver from loading on it so the driver does not pollute
dmesg with an error.
Signed-off-by: Richard Acayan <mailingradian@gmail.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Test runners on debug kernels occasionally fail with:
# # RUN tls_err.13_aes_gcm.poll_partial_rec_async ...
# # tls.c:1883:poll_partial_rec_async:Expected poll(&pfd, 1, 5) (0) == 1 (1)
# # tls.c:1870:poll_partial_rec_async:Expected status (256) == 0 (0)
# # poll_partial_rec_async: Test failed at step #17
# # FAIL tls_err.13_aes_gcm.poll_partial_rec_async
# not ok 699 tls_err.13_aes_gcm.poll_partial_rec_async
# # FAILED: 698 / 699 tests passed.
This points to the second poll() in the test which is expected
to wait for the sender to send the rest of the data.
Apparently under some conditions that doesn't happen within 5ms,
bump the timeout to 20ms.
Fixes: 23fcb62bc1 ("selftests: tls: add tests for poll behavior")
Link: https://lore.kernel.org/r/20240213142055.395564-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
To fully support initializing the LAG support code, a DDP package that
extracts the logical port from the metadata is required. If such a
package is not present, there could be difficulties in supporting some
bond types.
Add a check into the initialization flow that will bypass the new paths
if any of the support pieces are missing.
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Fixes: df006dd4b1 ("ice: Add initial support framework for LAG")
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Reviewed-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Sujai Buvaneswaran <sujai.buvaneswaran@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Link: https://lore.kernel.org/r/20240213183957.1483857-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Using this_cpu_has_cap() has the potential to go wrong when
used system-wide on a preemptible kernel. Instead, use the
__system_matches_cap() helper when checking for FEAT_NV in the
FEAT_NV1 probing helper.
Fixes: 3673d01a2f ("arm64: cpufeatures: Only check for NV1 if NV is present")
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Link: https://lore.kernel.org/kvmarm/86bk8k5ts3.wl-maz@kernel.org/
Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
that's how we love it ;-)
iwlwifi:
- correct A3 in A-MSDUs
- fix crash when operating as AP and running out of station
slots to use
- clear link ID to correct some later checks against it
- fix error codes in SAR table loading
- fix error path in PPAG table read
mac80211:
- reload a pointer after SKB may have changed
(only in certain monitor inject mode scenarios)
-----BEGIN PGP SIGNATURE-----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=h/fM
-----END PGP SIGNATURE-----
Merge tag 'wireless-2024-02-14' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless
Johannes Berg says:
====================
Valentine's day edition, with just few fixes because
that's how we love it ;-)
iwlwifi:
- correct A3 in A-MSDUs
- fix crash when operating as AP and running out of station
slots to use
- clear link ID to correct some later checks against it
- fix error codes in SAR table loading
- fix error path in PPAG table read
mac80211:
- reload a pointer after SKB may have changed
(only in certain monitor inject mode scenarios)
* tag 'wireless-2024-02-14' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless:
wifi: iwlwifi: mvm: fix a crash when we run out of stations
wifi: iwlwifi: uninitialized variable in iwl_acpi_get_ppag_table()
wifi: iwlwifi: Fix some error codes
wifi: iwlwifi: clear link_id in time_event
wifi: iwlwifi: mvm: use correct address 3 in A-MSDU
wifi: mac80211: reload info pointer in ieee80211_tx_dequeue()
====================
Link: https://lore.kernel.org/r/20240214184326.132813-3-johannes@sipsolutions.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If we hit CQ ring overflow when attempting to post a multishot accept
completion, we don't properly save the result or return code. This
results in losing the accepted fd value.
Instead, we return the result from the poll operation that triggered
the accept retry. This is generally POLLIN|POLLPRI|POLLRDNORM|POLLRDBAND
which is 0xc3, or 195, which looks like a valid file descriptor, but it
really has no connection to that.
Handle this like we do for other multishot completions - assign the
result, and return IOU_STOP_MULTISHOT to cancel any further completions
from this request when overflow is hit. This preserves the result, as we
should, and tells the application that the request needs to be re-armed.
Cc: stable@vger.kernel.org
Fixes: 515e269612 ("io_uring: revert "io_uring fix multishot accept ordering"")
Link: https://github.com/axboe/liburing/issues/1062
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The context persistence code does things like send super high priority
heartbeat pulses to ensure any leaked context can still be pre-empted
and thus isn't a total denial of service but only a minor denial of
service. Unfortunately, it wasn't bothering to restart the heartbeat
worker with a fresh timeout. Thus, if a persistent context happened to
be closed just before the heartbeat was going to go ping anyway then
the forced pulse would get a negligble execution time. And as the
forced pulse is super high priority, the worker thread's next step is
a reset. Which means a potentially innocent system randomly goes boom
when attempting to close a context. So, force a re-schedule of the
worker thread with the appropriate timeout.
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240110210216.4125092-1-John.C.Harrison@Intel.com