Lukas Wunner [Thu, 12 May 2022 08:42:04 +0000 (10:42 +0200)]
usbnet: smsc95xx: Avoid link settings race on interrupt reception
When a PHY interrupt is signaled, the SMSC LAN95xx driver updates the
MAC full duplex mode and PHY flow control registers based on cached data
in struct phy_device:
smsc95xx_status() # raises EVENT_LINK_RESET
usbnet_deferred_kevent()
smsc95xx_link_reset() # uses cached data in phydev
Simultaneously, phylib polls link status once per second and updates
that cached data:
phy_state_machine()
phy_check_link_status()
phy_read_status()
lan87xx_read_status()
genphy_read_status() # updates cached data in phydev
If smsc95xx_link_reset() wins the race against genphy_read_status(),
the registers may be updated based on stale data.
E.g. if the link was previously down, phydev->duplex is set to
DUPLEX_UNKNOWN and that's what smsc95xx_link_reset() will use, even
though genphy_read_status() may update it to DUPLEX_FULL afterwards.
PHY interrupts are currently only enabled on suspend to trigger wakeup,
so the impact of the race is limited, but we're about to enable them
perpetually.
Avoid the race by delaying execution of smsc95xx_link_reset() until
phy_state_machine() has done its job and calls back via
smsc95xx_handle_link_change().
Signaling EVENT_LINK_RESET on wakeup is not necessary because phylib
picks up link status changes through polling. So drop the declaration
of a ->link_reset() callback.
Note that the semicolon on a line by itself added in smsc95xx_status()
is a placeholder for a function call which will be added in a subsequent
commit. That function call will actually handle the INT_ENP_PHY_INT_
interrupt.
Tested-by: Oleksij Rempel <o.rempel@pengutronix.de> # LAN9514/9512/9500 Tested-by: Ferry Toth <fntoth@gmail.com> # LAN9514 Signed-off-by: Lukas Wunner <lukas@wunner.de> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Lukas Wunner [Thu, 12 May 2022 08:42:03 +0000 (10:42 +0200)]
usbnet: smsc95xx: Don't reset PHY behind PHY driver's back
smsc95xx_reset() resets the PHY behind the PHY driver's back, which
seems like a bad idea generally. Remove that portion of the function.
We're about to use PHY interrupts instead of polling to detect link
changes on SMSC LAN95xx chips. Because smsc95xx_reset() is called from
usbnet_open(), PHY interrupt settings are lost whenever the net_device
is brought up.
There are two other callers of smsc95xx_reset(), namely smsc95xx_bind()
and smsc95xx_reset_resume(), and both may indeed benefit from a PHY
reset. However they already perform one through their calls to
phy_connect_direct() and phy_init_hw().
Tested-by: Oleksij Rempel <o.rempel@pengutronix.de> # LAN9514/9512/9500 Tested-by: Ferry Toth <fntoth@gmail.com> # LAN9514 Signed-off-by: Lukas Wunner <lukas@wunner.de> Cc: Martyn Welch <martyn.welch@collabora.com> Cc: Gabriel Hojda <ghojda@yo2urs.ro> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Upon receiving data from the Interrupt Endpoint, the SMSC LAN95xx driver
attempts to clear the signaled interrupts by writing "all ones" to the
Interrupt Status Register.
However the driver only ever enables a single type of interrupt, namely
the PHY Interrupt. And according to page 119 of the LAN950x datasheet,
its bit in the Interrupt Status Register is read-only. There's no other
way to clear it than in a separate PHY register:
The commit made binding and unbinding of USB Ethernet asymmetrical:
Before, usbnet_probe() first invoked the ->bind() callback and then
register_netdev(). usbnet_disconnect() mirrored that by first invoking
unregister_netdev() and then ->unbind().
Since the commit, the order in usbnet_disconnect() is reversed and no
longer mirrors usbnet_probe().
One consequence is that a PHY disconnected (and stopped) in ->unbind()
is afterwards stopped once more by unregister_netdev() as it closes the
netdev before unregistering. That necessitates a contortion in ->stop()
because the PHY may only be stopped if it hasn't already been
disconnected.
Reverting the commit allows making the call to phy_stop() unconditional
in ->stop().
Tested-by: Oleksij Rempel <o.rempel@pengutronix.de> # LAN9514/9512/9500 Tested-by: Ferry Toth <fntoth@gmail.com> # LAN9514 Signed-off-by: Lukas Wunner <lukas@wunner.de> Acked-by: Oliver Neukum <oneukum@suse.com> Cc: Martyn Welch <martyn.welch@collabora.com> Cc: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Remove .owner field if calls are used which set it automatically.
./drivers/net/ethernet/sunplus/spl2sw_driver.c:569:3-8: No need to set
.owner here. The core will do it.
Reported-by: Abaci Robot <abaci@linux.alibaba.com> Signed-off-by: Yang Li <yang.lee@linux.alibaba.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jie Wang [Thu, 12 May 2022 06:56:31 +0000 (14:56 +0800)]
net: page_pool: add page allocation stats for two fast page allocate path
Currently If use page pool allocation stats to analysis a RX performance
degradation problem. These stats only count for pages allocate from
page_pool_alloc_pages. But nic drivers such as hns3 use
page_pool_dev_alloc_frag to allocate pages, so page stats in this API
should also be counted.
Signed-off-by: Jie Wang <wangjie125@huawei.com> Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jiapeng Chong [Thu, 12 May 2022 06:09:05 +0000 (14:09 +0800)]
net: ethernet: Use swap() instead of open coding it
Clean the following coccicheck warning:
./drivers/net/ethernet/sunplus/spl2sw_driver.c:217:27-28: WARNING
opportunity for swap().
./drivers/net/ethernet/sunplus/spl2sw_driver.c:222:27-28: WARNING
opportunity for swap().
Reported-by: Abaci Robot <abaci@linux.alibaba.com> Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com> Signed-off-by: David S. Miller <davem@davemloft.net>
drm/i915: Fix CFI violation with show_dynamic_id()
When an attribute group is created with sysfs_create_group(), the
->sysfs_ops() callback is set to kobj_sysfs_ops, which sets the ->show()
callback to kobj_attr_show(). kobj_attr_show() uses container_of() to
get the ->show() callback from the attribute it was passed, meaning the
->show() callback needs to be the same type as the ->show() callback in
'struct kobj_attribute'.
However, show_dynamic_id() has the type of the ->show() callback in
'struct device_attribute', which causes a CFI violation when opening the
'id' sysfs node under drm/card0/metrics. This happens to work because
the layout of 'struct kobj_attribute' and 'struct device_attribute' are
the same, so the container_of() cast happens to allow the ->show()
callback to still work.
Change the type of show_dynamic_id() to match the ->show() callback in
'struct kobj_attributes' and update the type of sysfs_metric_id to
match, which resolves the CFI violation.
Arnd Bergmann [Fri, 13 May 2022 09:31:21 +0000 (11:31 +0200)]
Merge tag 'visconti-arm-dt-for-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/iwamatsu/linux-visconti into arm/dt
Visconti device tree updates for 5.19
- Update the clock providers for PCIe host controller
- Update the clock providers for ethernet device
- Update the clock providers for SPI
- Update the clock providers for watchdog timer
- Update the clock providers for I2C
- Update the clock providers for UART
- Add clock controller support for TMPV7708
* tag 'visconti-arm-dt-for-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/iwamatsu/linux-visconti:
arm64: dts: visconti: Update the clock providers for PCIe host controller
arm64: dts: visconti: Update the clock providers for ethernet device
arm64: dts: visconti: Update the clock providers for SPI
arm64: dts: visconti: Update the clock providers for watchdog timer
arm64: dts: visconti: Update the clock providers for I2C
arm64: dts: visconti: Update the clock providers for UART
arm64: dts: visconti: Add clock controller support for TMPV7708
Hans Verkuil [Fri, 6 May 2022 12:46:46 +0000 (14:46 +0200)]
media: cec-adap.c: reconfigure if the PA changes during configuration
If the physical address changes (i.e. becomes invalid, then valid again)
while the adapter is still claiming free logical addresses, then trigger
a reconfiguration since any claimed LAs may now be stale.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Arnd Bergmann [Fri, 13 May 2022 09:29:11 +0000 (11:29 +0200)]
Merge tag 'v5.19-rockchip-dts32-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into arm/dt
Naming and flag corrections.
* tag 'v5.19-rockchip-dts32-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
ARM: dts: rockchip: rename pcfg_pull_default node name on rk3036
ARM: dts: rockchip: use generic node name for dma rk3036/rk322x
ARM: dts: rockchip: correct interrupt flags on rk3188 boards
Hans Verkuil [Fri, 6 May 2022 07:43:25 +0000 (09:43 +0200)]
media: cec-adap.c: fix is_configuring state
If an adapter is trying to claim a free logical address then it is
in the 'is_configuring' state. If during that process the cable is
disconnected (HPD goes low, which in turn invalidates the physical
address), then cec_adap_unconfigure() is called, and that set the
is_configuring boolean to false, even though the thread that's
trying to claim an LA is still running.
Don't touch the is_configuring bool in cec_adap_unconfigure(), it
will eventually be cleared by the thread. By making that change
the cec_config_log_addr() function also had to change: it was
aborting if is_configuring became false (since that is what
cec_adap_unconfigure() did), but that no longer works. Instead
check if the physical address is invalid. That is a much
more appropriate check anyway.
This fixes a bug where the the adapter could be disabled even
though the device was still configuring. This could cause POLL
transmits to time out.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Arnd Bergmann [Fri, 13 May 2022 09:28:40 +0000 (11:28 +0200)]
Merge tag 'v5.19-rockchip-dts64-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into arm/dt
New peripherals supported on rk356x: sfc, usb3, sata and
the video-decoder on rk3328.
RK3399 received some improvements and nodes for the memory controller.
Additional peripherals for PineNote, Gru and BananaPi-R2-Pro.
New boards are the Firefly Station M2, Pine64 SoQuartz SOM and
Quartz64 model B as well as the Radxa Rock3 model A.
* tag 'v5.19-rockchip-dts64-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip: (32 commits)
arm64: dts: rockchip: enable otg/drd operation of usb_host0_xhci in rk356x
arm64: dts: rockchip: rename HDMI ref clock to 'ref' on rk3399
arm64: dts: rockchip: add dts for Firefly Station M2 rk3566
arm64: dts: rockchip: add SoQuartz CM4IO dts
arm64: dts: rockchip: add Pine64 Quartz64-B device tree
dt-bindings: arm: rockchip: Add Firefly Station M2
dt-bindings: arm: rockchip: Add Pine64 SoQuartz SoM
dt-bindings: arm: rockchip: Add Pine64 Quartz64 Model B
arm64: dts: rockchip: enable usb hub on the radxa rock3 model a
arm64: dts: rockchip: add usb3 support to the radxa rock3 model a
arm64: dts: rockchip: add rk356x sfc support
arm64: dts: rockchip: Add USB and TCPC to rk3566-pinenote
arm64: dts: rockchip: Add accelerometer to rk3566-pinenote
arm64: dts: rockchip: add an input enable pinconf to rk3399
arm64: dts: rockchip: Add vdec support for RK3328
arm64: dts: rockchip: Rename vdec_mmu node for RK3328
arm64: dts: rockchip: Enable dmc and dfi nodes on gru
arm64: dts: rockchip: Add dfi and dmc nodes to rk3399
arm64: dts: rockchip: add clocks property to cru nodes rk3399
arm64: dts: rockchip: use generic node name for pmucru on rk3399
...
Hans Verkuil [Fri, 6 May 2022 07:39:22 +0000 (09:39 +0200)]
media: cec-adap.c: stop trying LAs on CEC_TX_STATUS_TIMEOUT
If, while trying to claim a free logical address, a POLL message
times out, then abort this process. A CEC_TX_STATUS_TIMEOUT
should be handled the same as a CEC_TX_STATUS_ABORTED.
This avoids a situation where transmits time out due to a
driver or hardware bug and it takes ages before the attempt
to find available free logical addresses finishes.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Fri, 6 May 2022 07:25:55 +0000 (09:25 +0200)]
media: cec-adap.c: don't unconfigure if already unconfigured
The __cec_s_log_addrs() function can configure or unconfigure the
adapter. The ioctl handler in cec-api.c will prevent it from being
called to configure the adapter if it was already configured (or in
the process of configuring). But it can still be called to unconfigure
an already unconfigured adapter, and it didn't check for that.
This can cause cec_activate_cnt_dec() to be called too often, causing
a WARN_ON.
Instead first check if adap->log_addrs.num_log_addrs == 0 and return
since in that case the adapter is already unconfigured.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Tue, 10 May 2022 11:59:48 +0000 (13:59 +0200)]
media: cec-pin.c: don't zero work_pin_num_events in adap_enable
It's OK to keep the pending pin events when disabling or
enabling the 'adapter'. Zeroing this can cause a race condition
if this happens when the pin kthread is handling a pin event
and calls atomic_dec later, causing work_pin_num_events to become
negative.
Just leave pending events in the queue, they'll be read eventually.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Fri, 6 May 2022 07:15:40 +0000 (09:15 +0200)]
media: cec-pin.c: disabling the adapter cannot call kthread_stop
When the adap_enable callback is called the adap->lock is held.
When disabling the adapter it attempts to stop the kthread that
deals with receiving and transmitting messages. However, kthread_stop
waits for the thread to stop, so all that time adap->lock is held.
Unfortunately, the kernel thread itself can call functions that take
that same lock, so a deadlock can occur.
Change the logic to keep the kernel thread running and instead when
disabling the adapter, just set the pin to high, go to idle and then
to state OFF and disable the interrupt. Only stop the kernel thread
when the adapter is deleted.
This way disabling the adapter will not wait for anything and the
deadlock is avoided.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Vladis Dronov [Sun, 8 May 2022 13:09:44 +0000 (15:09 +0200)]
crypto: s390 - add crypto library interface for ChaCha20
Implement a crypto library interface for the s390-native ChaCha20 cipher
algorithm. This allows us to stop to select CRYPTO_CHACHA20 and instead
select CRYPTO_ARCH_HAVE_LIB_CHACHA. This allows BIG_KEYS=y not to build
a whole ChaCha20 crypto infrastructure as a built-in, but build a smaller
CRYPTO_LIB_CHACHA instead.
Make CRYPTO_CHACHA_S390 config entry to look like similar ones on other
architectures. Remove CRYPTO_ALGAPI select as anyway it is selected by
CRYPTO_SKCIPHER.
Add a new test module and a test script for ChaCha20 cipher and its
interfaces. Here are test results on an idle z15 machine:
crypto: cryptd - Protect per-CPU resource by disabling BH.
The access to cryptd_queue::cpu_queue is synchronized by disabling
preemption in cryptd_enqueue_request() and disabling BH in
cryptd_queue_worker(). This implies that access is allowed from BH.
If cryptd_enqueue_request() is invoked from preemptible context _and_
soft interrupt then this can lead to list corruption since
cryptd_enqueue_request() is not protected against access from
soft interrupt.
Replace get_cpu() in cryptd_enqueue_request() with local_bh_disable()
to ensure BH is always disabled.
Remove preempt_disable() from cryptd_queue_worker() since it is not
needed because local_bh_disable() ensures synchronisation.
Fixes: 254eff771441 ("crypto: cryptd - Per-CPU thread implementation...") Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Corentin Labbe [Mon, 2 May 2022 20:19:28 +0000 (20:19 +0000)]
crypto: sun8i-ce - rework debugging
The "Fallback for xxx" message is annoying, remove it and store the
information in the debugfs.
Let's add more precise fallback stats and display it better.
Signed-off-by: Corentin Labbe <clabbe@baylibre.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Corentin Labbe [Mon, 2 May 2022 20:19:27 +0000 (20:19 +0000)]
crypto: sun8i-ce - use sg_nents_for_len
When testing with some large SG list, the sun8i-ce drivers always
fallback even if it can handle it.
So use sg_nents_for_len() which permits to see less SGs than needed.
Signed-off-by: Corentin Labbe <clabbe@baylibre.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Corentin Labbe [Mon, 2 May 2022 20:19:18 +0000 (20:19 +0000)]
crypto: sun8i-ss - use sg_nents_for_len
When testing with some large SG list, the sun8i-ss drivers always
fallback even if it can handle it.
So use sg_nents_for_len() which permits to see less SGs than needed.
Signed-off-by: Corentin Labbe <clabbe@baylibre.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Corentin Labbe [Mon, 2 May 2022 20:19:14 +0000 (20:19 +0000)]
crypto: sun8i-ss - rework handling of IV
sun8i-ss fail handling IVs when doing decryption of multiple SGs in-place.
It should backup the last block of each SG source for using it later as
IVs.
In the same time remove allocation on requests path for storing all
IVs.
Corentin Labbe [Mon, 2 May 2022 20:19:13 +0000 (20:19 +0000)]
crypto: sun4i-ss - do not allocate backup IV on requests
Instead of allocate memory on each requests, it is easier to
pre-allocate buffer for backup IV.
This made error path easier. Signed-off-by: Corentin Labbe <clabbe@baylibre.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tetsuo Handa [Mon, 2 May 2022 03:49:04 +0000 (05:49 +0200)]
media: imon: reorganize serialization
Since usb_register_dev() from imon_init_display() from imon_probe() holds
minor_rwsem while display_open() which holds driver_lock and ictx->lock is
called with minor_rwsem held from usb_open(), holding driver_lock or
ictx->lock when calling usb_register_dev() causes circular locking
dependency problem.
Since usb_deregister_dev() from imon_disconnect() holds minor_rwsem while
display_open() which holds driver_lock is called with minor_rwsem held,
holding driver_lock when calling usb_deregister_dev() also causes circular
locking dependency problem.
Sean Young explained that the problem is there are imon devices which have
two usb interfaces, even though it is one device. The probe and disconnect
function of both usb interfaces can run concurrently.
Alan Stern responded that the driver and USB cores guarantee that when an
interface is probed, both the interface and its USB device are locked.
Ditto for when the disconnect callback gets run. So concurrent probing/
disconnection of multiple interfaces on the same device is not possible.
Therefore, we don't need locks for handling race between imon_probe() and
imon_disconnect(). But we still need to handle race between display_open()
/vfd_write()/lcd_write()/display_close() and imon_disconnect(), for
disconnect event can happen while file descriptors are in use.
Since "struct file"->private_data is set by display_open(), vfd_write()/
lcd_write()/display_close() can assume that "struct file"->private_data
is not NULL even after usb_set_intfdata(interface, NULL) was called.
Replace insufficiently held driver_lock with refcount_t based management.
Add a boolean flag for recording whether imon_disconnect() was already
called. Use RCU for accessing this boolean flag and refcount_t.
Since the boolean flag for imon_disconnect() is shared, disconnect event
on either intf0 or intf1 affects both interfaces. But I assume that this
change does not matter, for usually disconnect event would not happen
while interfaces are in use.
Link: https://syzkaller.appspot.com/bug?extid=c558267ad910fc494497 Reported-by: syzbot <syzbot+c558267ad910fc494497@syzkaller.appspotmail.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Tested-by: syzbot <syzbot+c558267ad910fc494497@syzkaller.appspotmail.com> Cc: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Ondrej Mosnacek [Tue, 3 May 2022 11:50:10 +0000 (13:50 +0200)]
crypto: qcom-rng - fix infinite loop on requests not multiple of WORD_SZ
The commit referenced in the Fixes tag removed the 'break' from the else
branch in qcom_rng_read(), causing an infinite loop whenever 'max' is
not a multiple of WORD_SZ. This can be reproduced e.g. by running:
kcapi-rng -b 67 >/dev/null
There are many ways to fix this without adding back the 'break', but
they all seem more awkward than simply adding it back, so do just that.
Tested on a machine with Qualcomm Amberwing processor.
Fixes: a680b1832ced ("crypto: qcom-rng - ensure buffer for generate is completely filled") Cc: stable@vger.kernel.org Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Reviewed-by: Brian Masney <bmasney@redhat.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Oliver Neukum [Thu, 28 Apr 2022 13:34:55 +0000 (15:34 +0200)]
media: imon: drop references only after device is no longer used
The point of using get/put_device() is to keep references
for as long as the device may be in use. That means dropping
them must be the penultimate action right before freeing the memory.
Signed-off-by: Oliver Neukum <oneukum@suse.com> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
media: hantro: HEVC: unconditionnaly set pps_{cb/cr}_qp_offset values
Always set pps_cb_qp_offset and pps_cr_qp_offset values in Hantro/G2
register whatever is V4L2_HEVC_PPS_FLAG_PPS_SLICE_CHROMA_QP_OFFSETS_PRESENT
flag value.
The vendor code does the same to set these values.
This fixes conformance test CAINIT_G_SHARP_3.
Fluster HEVC score is increase by one with this patch.
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com> Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
PoC shall be int the range of -2^31 to 2^31 -1
(HEVC spec section 8.3.1 Decoding process for picture order count).
The current way to know if an entry in reference picture array is free
is to test if PoC = UNUSED_REF. Since UNUSED_REF is defined as '-1' that
could lead to decode issue if one PoC also equal '-1'.
PoC with value = '-1' exists in conformance test SLIST_B_Sony_9.
Change the way unused entries are managed in reference pictures array to
avoid using PoC to detect then.
This patch doesn't change fluster HEVC score.
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com> Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
media: hantro: HEVC: Fix tile info buffer value computation
Use pps->column_width_minus1[j] + 1 as value for the tile info buffer
instead of pps->column_width_minus1[j + 1].
The patch fixes DBLK_E_VIXS_2, DBLK_F_VIXS_2, DBLK_G_VIXS_2,
SAO_B_MediaTek_5, TILES_A_Cisco_2 and TILES_B_Cisco_1 tests in fluster.
Fixes: cb5dd5a0fa51 ("media: hantro: Introduce G2/HEVC decoder") Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hantro decoder doesn't take care of the requested and aligned size
of the capture buffer.
Stop using the bitstream width/height and use capture frame size
stored in the context to get the correct values.
hantro_hevc_chroma_offset() and hantro_hevc_motion_vectors_offset()
are only used in hantro_g2_hevc_dec.c so take the opportunity
to move them here.
fluster HEVC score goes up from 77 to 85 successful tests (over 147)
with this patch.
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
media: hantro: Use post processor scaling capacities
Hantro G2 post processor is able to down scale decoded frames
by a factor of 2, 4 or 8.
Add enum_framesizes() ops to postproc_ops structure to enumerate the
possible output sizes for a given input resolution.
For G2 post-processor use fsize->index (from 0 to 3) as power of 2
divisor. As described in v4l2 documentation return -EINVAL when scaling
down isn't possible.
fluster scores:
77/147 for HEVC
143/303 for VP9
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Ming Qian [Mon, 9 May 2022 07:52:25 +0000 (09:52 +0200)]
media: amphion: G/S_PARM only for encoder's output queue
G/S_PARM doesn't make sense for the capture queue of a stateful encoder,
unless V4L2_FMT_FLAG_ENC_CAP_FRAME_INTERVAL is set to reserve hardware
resources.
Otherwise it will fail the v4l2-compliance
Signed-off-by: Ming Qian <ming.qian@nxp.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Eugen Hristev [Tue, 3 May 2022 08:44:18 +0000 (10:44 +0200)]
media: atmel: atmel-isc-base: use mutex to lock awb workq from streaming
The AWB workqueue runs in a kernel thread and needs to be synchronized
w.r.t. the streaming status.
It is possible that streaming is stopped while the AWB workq is running.
In this case it is likely that the check for vb2_start_streaming_called is
done at one point in time, but the AWB computations are done later,
including a call to isc_update_profile, which requires streaming to be
started.
Thus , isc_update_profile will fail if during this operation sequence the
streaming was stopped.
To solve this issue, a mutex is added, that will serialize the awb work and
streaming stopping, with the mention that either streaming is stopped
completely including termination of the last frame is done, and after that
the AWB work can check stream status and stop; either first AWB work is
completed and after that the streaming can stop correctly.
The awb spin lock cannot be used since this spinlock is taken in the same
context and using it in the stop streaming will result in a recursion BUG.
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com> Reviewed-by: Jacopo Mondi <jacopo@jmondi.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Eugen Hristev [Tue, 3 May 2022 08:44:17 +0000 (10:44 +0200)]
media: atmel: atmel-sama5d2-isc: fix wrong mask in YUYV format check
While this does not happen in production, this check should be done
versus the mask, as checking with the YCYC value may not include
some bits that may be set.
It is correct and safe to check the whole mask.
Eugen Hristev [Tue, 3 May 2022 08:44:14 +0000 (10:44 +0200)]
media: atmel: atmel-isc-base: use streaming status when queueing buffers
During experiments with libcamera, it looks like vb2_is_streaming returns
true before our start streaming is called.
Order of operations is streamon -> queue -> start_streaming
ISC would have started the DMA immediately when a buffer is being added
to the vbqueue if the queue is streaming.
It is more safe to start the DMA after the start streaming of the driver is
called.
Thus, even if vb2queue is streaming, add the buffer to the dma queue of the
driver instead of actually starting the DMA process, if the start streaming
has not been called yet.
Tho achieve this, we have to use vb2_start_streaming_called instead of
vb2_is_streaming.
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com> Reviewed-by: Jacopo Mondi <jacopo@jmondi.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Michael Rodin [Tue, 23 Nov 2021 11:50:36 +0000 (12:50 +0100)]
media: vsp1: Fix offset calculation for plane cropping
The vertical subsampling factor is currently not considered in the
offset calculation for plane cropping done in rpf_configure_partition.
This causes a distortion (shift of the color plane) when formats with
the vsub factor larger than 1 are used (e.g. NV12, see
vsp1_video_formats in vsp1_pipe.c). This commit considers vsub factor
for all planes except plane 0 (luminance).
Drop generalization of the offset calculation to reduce the binary size.
Laurent Pinchart [Fri, 18 Mar 2022 18:01:03 +0000 (19:01 +0100)]
media: vsp1: Don't open-code vb2_fop_release()
Use the vb2_fop_release() helper to replace the open-coded version. The
video->lock is assigned to the queue lock, used by vb2_fop_release(), so
the only functional difference is that v4l2_fh_release() is now called
before vsp1_device_put(). This should be harmless.
Laurent Pinchart [Fri, 18 Mar 2022 17:44:35 +0000 (18:44 +0100)]
media: videobuf2-v4l2: Expose vb2_queue_is_busy() to drivers
vb2 queue ownership is managed by the ioctl handler helpers
(vb2_ioctl_*). There are however use cases where drivers can benefit
from checking queue ownership, for instance when open-coding an ioctl
handler that needs to perform additional checks before calling the
corresponding vb2 operation.
Expose the vb2_queue_is_busy() function in the videobuf2-v4l2.h header,
and change its first argument to a struct vb2_queue pointer as the
function name implies it operates on a queue, not a video_device.
media: atomisp: don't pass a pointer to a local variable
As warned by gcc 12.1:
drivers/staging/media/atomisp/pci/runtime/rmgr/src/rmgr_vbuf.c: In function 'ia_css_rmgr_acq_vbuf':
drivers/staging/media/atomisp/pci/runtime/rmgr/src/rmgr_vbuf.c:275:33: error: storing the address of local variable 'h' in '*handle' [-Werror=dangling-pointer=]
275 | *handle = &h;
| ~~~~~~~~^~~~
drivers/staging/media/atomisp/pci/runtime/rmgr/src/rmgr_vbuf.c:257:40: note: 'h' declared here
257 | struct ia_css_rmgr_vbuf_handle h;
| ^
drivers/staging/media/atomisp/pci/runtime/rmgr/src/rmgr_vbuf.c:257:40: note: 'handle' declared here
cc1: all warnings being treated as errors
The logic uses a temporary struct to update the handler, but,
instead of copying the value to the pointer sent by the caller, it
replaces it with the content with a local variable. That's wrong, and
may lead the caller to use a weird value.
Tom Rix [Sat, 12 Mar 2022 17:30:49 +0000 (18:30 +0100)]
media: stkwebcam: move stk_camera_read_reg() scratch buffer to struct stk_camera
In stk_camera_read_reg() a single byte buffer is alloc-ed and
freed on every function call. Since the size is known,
move the buffer to the struct stk_camera where it will be alloc-ed
and freed once.
Signed-off-by: Tom Rix <trix@redhat.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Let VIDIOC_ENUM_FRAMEINTERVALS return -EINVAL if userspace queries
frame intervals for frame sizes unsupported by the encoder. Fixes the
following v4l2-compliance failure:
fail: v4l2-test-formats.cpp(123): found frame intervals for invalid size 47x16
fail: v4l2-test-formats.cpp(282): node->codec_mask & STATEFUL_ENCODER
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: FAIL
[hverkuil: drop incorrect 'For decoder devices, return -ENOTTY.' in the commit log]
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Document the RK3328 compatible for rockchip-vdec. The driver shares
the same base functionality as the RK3399 hardware so make sure that
the RK3399 compatible is also included in the device tree.
Signed-off-by: Christopher Obbard <chris.obbard@collabora.com> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
I tried to debug streaming in libcamera, where I stumbled upon this.
I asked around on IRC where I was told that this statement in the
documentation is wrong, so I'm submitting a removal.
media: amphion: no need to check return value of debugfs_create functions
When calling debugfs functions, there is no need to ever check the
return value. The function can work or not, but the code logic should
never do something different based on this.
Ming Qian [Fri, 22 Apr 2022 01:36:21 +0000 (03:36 +0200)]
media: amphion: free ctrl handler if error is set and return error
The typical behavior is to add all controls, then at the end check if
hdl->error was set, and if so, v4l2_ctrl_handler_free is called and
the error is returned.
Signed-off-by: Ming Qian <ming.qian@nxp.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Ming Qian [Wed, 20 Apr 2022 11:35:59 +0000 (13:35 +0200)]
media: amphion: handle picture skipped event
For some invalid frames,
especially multiple consecutive invalid frames,
they all can't be decoded,
then the firmware can send picture skipped event
to notify driver that some frames are invalid,
driver can return them with error flag.
Signed-off-by: Ming Qian <ming.qian@nxp.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Ming Qian [Wed, 20 Apr 2022 11:34:54 +0000 (13:34 +0200)]
media: amphion: encoder copy timestamp from output to capture
copy the timestamp using the helper function
V4L2_BUF_FLAG_TIMESTAMP_COPY
To implement this, driver will keep the output buffer until it's
encoded, in previous, driver will return the output buffer immediately
after firmware return it
Signed-off-by: Ming Qian <ming.qian@nxp.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Ming Qian [Wed, 20 Apr 2022 11:34:53 +0000 (13:34 +0200)]
media: amphion: decoder copy timestamp from output to capture
copy the timestamp using the helper function
V4L2_BUF_FLAG_TIMESTAMP_COPY
To implement this, driver will keep the output buffer until it's
decoded, in previous, driver will return the output buffer immediately
after copying data to stream buffer.
After that, there is no need to make a workaround for poll function.
driver can use v4l2_m2m_fop_poll directly.
Also, driver don't need to keep a input threshold
as the buffer count is up to only 32.
Signed-off-by: Ming Qian <ming.qian@nxp.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Randy Dunlap [Wed, 20 Apr 2022 06:36:44 +0000 (08:36 +0200)]
media: make RADIO_ADAPTERS tristate
Fix build errors when RADIO_TEA575X=y, VIDEO_BT848=m, and VIDEO_DEV=m.
The build errors occur due to [in drivers/media/Makefile]:
obj-$(CONFIG_VIDEO_DEV) += radio/
so the (would be) builtin tea575x.o is not being built.
This is also due to drivers/media/radio/Kconfig declaring a bool
Kconfig symbol (RADIO_ADAPTERS) that depends on a tristate (VIDEO_DEV),
so when VIDEO_DEV=m, RADIO_ADAPTERS becomes =y, and then the drivers
that depend on RADIO_ADPATERS can be configured as builtin (=y) or
as loadable modules (=m).
Fix this by converting RADIO_ADAPTERS to a tristate symbol instead
of a bool symbol.
Pavel Skripkin [Fri, 15 Apr 2022 21:24:48 +0000 (23:24 +0200)]
media: pvrusb2: fix array-index-out-of-bounds in pvr2_i2c_core_init
Syzbot reported that -1 is used as array index. The problem was in
missing validation check.
hdw->unit_number is initialized with -1 and then if init table walk fails
this value remains unchanged. Since code blindly uses this member for
array indexing adding sanity check is the easiest fix for that.
hdw->workpoll initialization moved upper to prevent warning in
__flush_work.
Reported-and-tested-by: syzbot+1a247e36149ffd709a9b@syzkaller.appspotmail.com Fixes: d855497edbfb ("V4L/DVB (4228a): pvrusb2 to kernel 2.6.18") Signed-off-by: Pavel Skripkin <paskripkin@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Sebastian Fricke [Tue, 29 Mar 2022 06:56:56 +0000 (08:56 +0200)]
media: docs-rst: Append HEVC specific term
Describe the coding tree unit as replacement for the macroblock in the
HEVC codec. Highlight a key difference of the HEVC codec to predecessors
like AVC(H.264) to give a better overview of the differences between the
coding standards.
[hverkuil: replaced the 'corresponds to' symbol with the full text for clarity]
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com> Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Sebastian Fricke [Sat, 26 Mar 2022 18:37:17 +0000 (19:37 +0100)]
media: staging: media: rkvdec: Update TODO list
VP9 support has been added to the driver by commit f25709c4ff15
("media: rkvdec: Add the VP9 backend").
And the VP9 uABI was merged with commit b88dbe38dca8
("media: uapi: Add VP9 stateless decoder controls").
The remaining codec that keeps this driver in staging is HEVC.
Update the TODO list accordingly.
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com> Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
projects / people / ms / linux.git / log
