Set the V4L2_CAP_IO_MC capability flag and remove the driver specific
vidioc_enum_input, vidioc_g_input and vidioc_s_input callbacks for the
media controller enabled part of the driver. Also add support mbus_code
filtering for format enumeration.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The VIDIOC_ENUM_FMT ioctl enumerates all formats supported by a video
node. For MC-centric devices, its behaviour has always been ill-defined,
with drivers implementing one of the following behaviours:
- No support for VIDIOC_ENUM_FMT at all
- Enumerating all formats supported by the video node, regardless of the
configuration of the pipeline
- Enumerating formats supported by the video node for the active
configuration of the connected subdevice
The first behaviour is obviously useless for applications. The second
behaviour provides the most information, but doesn't offer a way to find
what formats are compatible with a given pipeline configuration. The
third behaviour fixes that, but with the drawback that applications
can't enumerate all supported formats anymore, and have to modify the
active configuration of the pipeline to enumerate formats.
The situation is messy as none of the implemented behaviours are ideal,
and userspace can't predict what will happen as the behaviour is
driver-specific.
To fix this, let's extend the VIDIOC_ENUM_FMT with a missing capability:
enumerating pixel formats for a given media bus code. The media bus code
is passed through the v4l2_fmtdesc structure in a new mbus_code field
(repurposed from the reserved fields). With this capability in place,
applications can enumerate pixel formats for a given media bus code
without modifying the active configuration of the device.
The current behaviour of the ioctl is preserved when the new mbus_code
field is set to 0, ensuring compatibility with existing userspace. The
API extension is documented as mandatory for MC-centric devices (as
advertised through the V4L2_CAP_IO_MC capability), allowing applications
and compliance tools to easily determine the availability of the
VIDIOC_ENUM_FMT extension.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Replace initialization of static const v4l2_fmtdesc instances that
specify every struct member with designated initializers. This allows
not zeroing the reserved fields explicitly, and will avoid a need to
patch these drivers every time a reserved field is repurposed.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Add a video device capability flag to indicate that its inputs and/or
outputs are controlled by the Media Controller instead of the V4L2 API.
When this flag is set, ioctl for enum inputs and outputs are
automatically enabled and programmed to call a helper function.
Suggested-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
This is required to bring Mali450 gpu out of reset.
Cc: Peter Griffin <peter.griffin@linaro.org>
Cc: Enrico Weigelt <info@metux.net>
Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
[jstultz: Added comment, Fix void return build issue
Reported-by: kbuild test robot <lkp@intel.com>]
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Add the reset line of the internal DAC found on the amlogic gxl SoC family
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
i.MX8MP is a new SoC of i.MX8M family, it has same src IP
inside but with different module layout, add support for it.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
The driver now supports i.MX8MP, so update bindings accordingly.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
i.MX8MN can reuse i.MX8MQ's reset driver, update the compatible
property and related info to support i.MX8MN.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
PM8150L WLED supports the following:
- Two modulators and each sink can use any of the modulator
- Multiple CABC selection options from which one can be selected/enabled
- Multiple brightness width selection (12 bits to 15 bits)
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
This is a rewrite of this[1] Wiki page with further enhancements. The
doc also includes a section on debugging problems in nested
environments, among other improvements.
[1] https://www.linux-kvm.org/page/Nested_Guests
Signed-off-by: Kashyap Chamarthy <kchamart@redhat.com>
Message-Id: <20200505112839.30534-1-kchamart@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The mute key, is broken. All the consumer keys on the keyboard USB
interface work normally, except for mute which only sends press events
and never sends release events.
The touchpad key sends the otherwise unused input report with a report-id
of 5 on the touchpad interface. It too only sends press events. This also
requires extra special handling since the multi-touch touchpad events and
the KEY_F21 events for the touchpad toggle must not be send from the same
input_dev (userspace cannot handle this).
This commit adds special handlig for both, fixing these keys not working.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
The multi-touch touchpad found on the Medion Akoya E1239T's keyboard-dock,
uses the same custom multi-touch protocol as the Asus keyboard-docks
(same chipset vendor, Integrated Technology Express / ITE).
Add support for this using the existing multi-touch touchpad support in
the hid-asus driver.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Add the report_size to struct asus_touchpad_info instead of calculating it.
This is a preparation patch for adding support for the multi-touch touchpad
found on the Medion Akoya E1239T's keyboard-dock, which uses the same
custom multi-touch protocol as the Asus keyboard-docks (same chipset
vendor, Integrated Technology Express / ITE).
The only difference in that the Akoya E1239T keyboard-dock's input-reports
have a 5 byte footer instead of a 1 byte footer, which requires the
report_size to be configurable per touchpad-model.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Add a hid_is_using_ll_driver(usb_hid_driver) check to ensure that the
parent device is an usb_interface, before casting the parent device
pointer to an usb_interface pointer with to_usb_interface().
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Before this commit the code was trying to keep the mapping for the
left mouse-button, do avoid the hidinput_has_been_populated() check
in hid-input.c from triggering and causing the touchpad input_dev ro
get cleaned up.
But the hidinput_has_been_populated() check happens after the
input_configured callback which sets up all the input bit for the
touchpad, so there is no need to preserve the left button mapping.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Make asus_input_mapping() only set EV_REP if we are adding a mapping.
The T100CHI bluetooth keyboard dock has a few input reports for which
we do not create any mappings (these input-reports are present in the
descriptors but never send).
The hid-asus code relies on the HID core not creating input devices for
input-reports without any mappings. But the present of the EV_REP but
counts as a mapping causing 6 /dev/input/event# nodes to be created for
the T100CHI bluetooth keyboard dock. This change brings the amount of
created /dev/input/event# nodes / input-devices down to 4.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Add WLED5 specific bindings.
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Acked-by: Daniel Thompson <daniel.thompson@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Add wled_cabc_config, wled_sync_toggle, wled_ovp_fault_status
and wled_ovp_delay and wled_auto_detection_required callback
functions to prepare the driver for adding WLED5 support.
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Convert the qcom-wled bindings from .txt to .yaml format.
Also replace PM8941 to WLED3 and PMI8998 to WLED4.
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
Signed-off-by: Subbaraman Narayanamurthy <subbaram@codeaurora.org>
Acked-by: Daniel Thompson <daniel.thompson@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Fix errors reported by dt_binding_check, due to missing required
regulators in the example node.
Fixes: f06b9fc9a8 ("dt-bindings: phy: Add binding for qcom,usb-snps-femto-v2")
Signed-off-by: Wesley Cheng <wcheng@codeaurora.org>
Reported-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/1588753589-22673-1-git-send-email-wcheng@codeaurora.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Fix the following coccicheck warning:
drivers/net/wireless/realtek/rtlwifi/rtl8188ee/rf.c:476:6-14: WARNING:
Comparison to bool
drivers/net/wireless/realtek/rtlwifi/rtl8188ee/rf.c:54:5-22: WARNING:
Comparison to bool
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504113321.41118-1-yanaijie@huawei.com
Fix the following coccicheck warning:
drivers/net/wireless/broadcom/b43/pio.c:768:10-25: WARNING: Comparison
of 0/1 to bool variable
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504113311.41026-1-yanaijie@huawei.com
Fix the following coccicheck warning:
drivers/net/wireless/broadcom/b43/phy_n.c:5510:19-32: WARNING:
Comparison of 0/1 to bool variable
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504113300.40895-1-yanaijie@huawei.com
Currently there is a check if priv is null when calling lbtf_remove_card
but not in a previous call to if_usb_reset_dev that can also dereference
priv. Fix this by also only calling lbtf_remove_card if priv is null.
It is noteable that there don't seem to be any bugs reported that the
null pointer dereference has ever occurred, so I'm not sure if the null
check is required, but since we're doing a null check anyway it should
be done for both function calls.
Addresses-Coverity: ("Dereference before null check")
Fixes: baa0280f08 ("libertas_tf: don't defer firmware loading until start()")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200501173900.296658-1-colin.king@canonical.com
gcc-10 reports a warning for mwifiex_cmd_802_11_key_material_v1:
drivers/net/wireless/marvell/mwifiex/sta_cmd.c: In function 'mwifiex_cmd_802_11_key_material_v1':
cc1: warning: writing 16 bytes into a region of size 0 [-Wstringop-overflow=]
In file included from drivers/net/wireless/marvell/mwifiex/sta_cmd.c:23:
drivers/net/wireless/marvell/mwifiex/fw.h:993:9: note: at offset 0 to object 'action' with size 2 declared here
993 | __le16 action;
| ^~~~~~
As the warning makes no sense, I reported it as a bug for gcc. In the
meantime using a temporary pointer for the key data makes the code easier
to read and stops the warning.
Fixes: 5e6e3a92b9 ("wireless: mwifiex: initial commit for Marvell mwifiex driver")
Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94881
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200430213101.135134-4-arnd@arndb.de
The variable 'rtlpriv->rfkill.rfkill_state' is bool and can directly
assigned to bool values.
Fix the following coccicheck warning:
drivers/net/wireless/realtek/rtlwifi/core.c:1725:14-42: WARNING:
Comparison of 0/1 to bool variable
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200429140924.7750-1-yanaijie@huawei.com
Fix the following coccicheck warning:
drivers/net/wireless/broadcom/brcm80211/brcmsmac/main.c:1060:6-12:
WARNING: Comparison to bool
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504113357.41422-1-yanaijie@huawei.com
It was observed that P2P Cert. 5.1.19: DEVUT responds to Service
Discovery request failed due to DUT did not send GAS Comeback Response
after receiving request from test bed P2P peer. To fix this issue,
we need to add P2P Action Frame retry delay to enhance P2P connection
under VSDB and noisy environment, since the peer can be in other
channels under VSDB.
Signed-off-by: Justin Li <Justin.Li@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/1588572453-194663-4-git-send-email-wright.feng@cypress.com
P2P group formation fails since either peer is not able to send go-neg
confirm or dut is not able to send go-neg response. To fix this, retry
limit should be increased and dwell time check should be added.
Signed-off-by: Joseph Chuang <joseph.chuang@cypress.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1588572453-194663-3-git-send-email-wright.feng@cypress.com
With RSDB feature, firmware is able to support two P2P-AGO or two
P2P-GC at the same time. So we add the second p2p connection type
to map to the second P2P connection bsscfg.
Signed-off-by: Wright Feng <wright.feng@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/1588572453-194663-2-git-send-email-wright.feng@cypress.com
the related system resources were not released when pci_iomap() return
error in the rtw_pci_io_mapping() function. add pci_release_regions() to
fix it.
Fixes: e3037485c6 ("rtw88: new Realtek 802.11ac driver")
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Dejin Zheng <zhengdejin5@gmail.com>
Acked-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504083442.3033-1-zhengdejin5@gmail.com
Interface configuration table is used to configure PCI PHY that are
normally decided by design or bootstrap pin, and driver can do additional
settings by this table.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-9-yhchuang@realtek.com
DIG sets only one IG register for most chips, but 8723D need to set
additional register for CCK rate.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-8-yhchuang@realtek.com
This ops is used to do statistics of false alarm periodically, and then
fine tune RX initial gain to adaptive different circumstance.
There are three steps, hold/get/reset counter, to retrieve false alarm
counters that consist of CCK and OFDM. In addition to false alarm
counters, it also collects CRC ok/error counters of CCK, OFDM and HT.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-7-yhchuang@realtek.com
Some chips are not able to receive LDPC packets. Add an attribute
to rtw_chip_info to determine if the LDPC capability in [ht/vht]_cap
should be advertised or not.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-6-yhchuang@realtek.com
TX status report of 8723D differs from 8822B/8822C, it uses
C2H_CCX_TX_RPT (0x03) with different format. With sequence number
and TX status, driver can know if certain packet was transmitted
successfully.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-5-yhchuang@realtek.com
Set MAC/BB/RF register according to specified channel. The function
rtw_set_channel_mac() is used to set MAC registers, but 8723D only need
some of them.
For channel 14, we need to set different CCK DFIR values, so restore the
values when channel 1 to 13 is selected.
Spur calibration is needed in channel 13 and 14, and we do notch if spur
is over threshold.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-4-yhchuang@realtek.com
This ops is used to parse RX descriptor to know the length of received
packet and containing PHY status. If PHY status is existing, the order is
RX descriptor, PHY status and then packet.
There are two types of PHY status, named CCK and OFDM. Their size are the
same, but formats are different.
struct ieee80211_rx_status is also filled depends on above information.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-3-yhchuang@realtek.com
To improve user experience in field, we need DIG to adjust RX initial gain
depends on field situation. Since each chip has its own register address,
this commit defines 8723d specific address.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200504105010.10780-2-yhchuang@realtek.com
Fix the following coccicheck warning:
drivers/net/wireless/ray_cs.c:2797:5-14: WARNING: Comparison of 0/1 to
bool variable
drivers/net/wireless/ray_cs.c:2798:2-11: WARNING: Assignment of 0/1 to
bool variable
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200426103709.6730-1-yanaijie@huawei.com
The 'blocked' is a bool variable, and '==' expression itself is bool
too. So no need to convert it to 0/1.
This fixes the following coccicheck warning:
drivers/net/wireless/realtek/rtlwifi/base.c:508:13-41: WARNING:
Comparison of 0/1 to bool variable
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200426094115.23294-1-yanaijie@huawei.com
Axe a memory allocation failure log message. This message is useless and
incorrect (vmalloc is not used here for the memory allocation)
This has been like that since the very beginning of this driver in
commit 43f66a6ce8 ("Add ipw2200 wireless driver.")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200424154527.27309-1-christophe.jaillet@wanadoo.fr
It's incorrect to use usleep in atomic context.
Switch to a macro which uses udelay instead of usleep to prevent the issue.
Fixes: 6343a6d4b2 ("rtw88: Add delay on polling h2c command status bit")
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200423073007.3566-1-kai.heng.feng@canonical.com
Like read_poll_timeout, an atomic variant for multiple parameter read
function can be useful.
Will be used by a later patch.
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200424184918.30360-1-kai.heng.feng@canonical.com
Luke Nelson says:
====================
This patch series introduces a set of optimizations to the BPF JIT
on RV64. The optimizations are related to the verifier zero-extension
optimization and BPF_JMP BPF_K.
We tested the optimizations on a QEMU riscv64 virt machine, using
lib/test_bpf and test_verifier, and formally verified their correctness
using Serval.
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
This patch optimizes BPF_JSET BPF_K by using a RISC-V andi instruction
when the BPF immediate fits in 12 bits, instead of first loading the
immediate to a temporary register.
Examples of generated code with and without this optimization:
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: and t1,a0,t1
28: bnez t1,0x30
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: mv t2,a0
28: slli t2,t2,0x20
2c: srli t2,t2,0x20
30: slli t1,t1,0x20
34: srli t1,t1,0x20
38: and t1,t2,t1
3c: bnez t1,0x44
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
In these examples, because the upper 32 bits of the sign-extended
immediate are 0, BPF_JMP BPF_JSET and BPF_JMP32 BPF_JSET are equivalent
and therefore the JIT produces identical code for them.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-5-luke.r.nels@gmail.com
This patch adds an optimization to BPF_JMP (32- and 64-bit) BPF_K for
when the BPF immediate is zero.
When the immediate is zero, the code can directly use the RISC-V zero
register instead of loading a zero immediate to a temporary register
first.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-4-luke.r.nels@gmail.com
