This loop ends on -1 so the error message will never be printed.
Fixes: 4bcf59a5de ("serial: 8250: 8250_omap: Account for data in flight during DMA teardown")
Reviewed-by: Alexander Sverdlin <alexander.sverdlin@gmail.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/YIpd+kOpXKMpEXPf@mwanda
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This fixes the following checkpatch.pl warnings:
WARNING: Missing a blank line after declarations
WARNING: braces {} are not necessary for any arm of this statement
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/257ffd691b4a062ad017333c9430d69da6dbd29a.1619594713.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use 'hlist_for_each_entry' instead of hand writing it.
This saves a few lines of code.
The comment about warning generated by some gcc version is also removed.
The way 'hlist_for_each_entry' is written should prevent such a warning to
be emitted.
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/14024ddeb2b3a8c5b0138b5ba5083f54d00164a9.1619594713.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The variable id being initialized with a value that is never read
and it is being updated later with a new value. The initialization is
redundant and can be removed. Since id is just being used in a for-loop
inside a local scope, move the declaration of id to that scope.
Reviewed-by: Kevin Hilman <khilman@baylibre.com>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Addresses-Coverity: ("Unused value")
Link: https://lore.kernel.org/r/20210426101106.9122-1-colin.king@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Our SoC's have always had a NS16650A UART core and older SoC's would
have a compatible string of: 'compatible = ""ns16550a"' and use the
8250_of driver. Our newer SoC's have added enhancements to the base
core to add support for DMA and accurate high speed baud rates and use
this newer 8250_bcm7271 driver. The Device Tree node for our enhanced
UARTs has a compatible string of: 'compatible = "brcm,bcm7271-uart",
"ns16550a"''. With both drivers running and the link order setup so
that the 8250_bcm7217 driver is initialized before the 8250_of driver,
we should bind the 8250_bcm7271 driver to the enhanced UART, or for
upstream kernels that don't have the 8250_bcm7271 driver, we bind to
the 8250_of driver.
The problem is that when both the 8250_of and 8250_bcm7271 drivers
were running, occasionally the 8250_of driver would be bound to the
enhanced UART instead of the 8250_bcm7271 driver. This was happening
because we use SCMI based clocks which come up late in initialization
and cause probe DEFER's when the two drivers get their clocks.
Occasionally the SCMI clock would become ready between the 8250_bcm7271
probe and the 8250_of probe and the 8250_of driver would be bound. To
fix this we decided to config only our 8250_bcm7271 driver and added
"ns16665a0" to the compatible string so the driver would work on our
older system.
This commit has of_platform_serial_probe() check specifically for the
"brcm,bcm7271-uart" and whether its companion driver is enabled. If it
is the case, and the clock provider is not ready, we want to make sure
that when the 8250_bcm7271.c driver returns EPROBE_DEFER, we are not
getting the UART registered via 8250_of.c.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Jim Quinlan <jim2101024@gmail.com>
Signed-off-by: Al Cooper <alcooperx@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20210423183206.3917725-1-f.fainelli@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The variable line being initialized with a value that is never read
and it is being updated later with a new value. The initialization is
redundant and can be removed.
Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Addresses-Coverity: ("Unused value")
Link: https://lore.kernel.org/r/20210423081317.318352-1-colin.king@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When there is more than one serial port present, the property 'label'
allows a custom name to be used for briefly describe the usage or position
of each serial port.
Without this "label" property, many dtbs_check warnings similar to the
following are reported:
arch/arm64/boot/dts/hisilicon/hi3660-hikey960.dt.yaml: \
serial@ffd74000: Additional properties are not allowed ('label' was unexpected)
From schema: Documentation/devicetree/bindings/serial/pl011.yaml
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Link: https://lore.kernel.org/r/20210506132049.1513-3-thunder.leizhen@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The /schemas/serial.yaml from dt-schema only has the property $nodename,
whereas the kernel-defined /schemas/serial/serial.yaml contains more
useful properties, support for more complex application scenarios.
For example:
1) The property "current-speed" in fsl-lx2160a.dtsi
2) The subnode "bluetooth" in hi3660-hikey960.dts
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Link: https://lore.kernel.org/r/20210506132049.1513-2-thunder.leizhen@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Previously this driver's use of devm_ioremap_resource() led to
duplicated calls to __release_region() when unbinding it (one from
serial8250_release_std_resource() and one from devres_release_all()),
the second of which resulted in a warning message:
# echo 1e787000.serial > /sys/bus/platform/drivers/aspeed-vuart/unbind
[33091.774200] Trying to free nonexistent resource <000000001e787000-000000001e78703f>
With this change the driver uses the generic serial8250 code's
UPF_IOREMAP to take care of the register mapping automatically instead
of doing its own devm_ioremap_resource(), thus avoiding the duplicate
__release_region() on unbind.
In doing this we eliminate vuart->regs, since it merely duplicates
vuart->port->port.membase, which we now use for our I/O accesses.
Reported-by: Andrew Jeffery <andrew@aj.id.au>
Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
Link: https://lore.kernel.org/r/20210510014231.647-4-zev@bewilderbeest.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Previously this had only been initialized if we hit the throttling path
in aspeed_vuart_handle_irq(); moving it to the probe function is a
slight consistency improvement and avoids redundant reinitialization in
the interrupt handler. It also serves as preparation for converting the
driver's I/O accesses to use port->port.membase instead of its own
vuart->regs.
Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
Link: https://lore.kernel.org/r/20210510014231.647-3-zev@bewilderbeest.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is a small prepatory step for changing the way this driver does
its I/O accesses.
Reviewed-by: Andrew Jeffery <andrew@aj.id.au>
Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
Link: https://lore.kernel.org/r/20210510014231.647-2-zev@bewilderbeest.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With commit 312c004d36 ("[PATCH] driver core: replace "hotplug" by
"uevent"") already in the tree over a decade, update the name of
FW_ACTION defines to follow semantics, and reflect what the defines are
really meant for, i.e. whether or not generate user space event.
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Link: https://lore.kernel.org/r/20210425020024.28057-1-shawn.guo@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The value of 'ret' is not used which reported by svace, so just return
instead of break.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: Tian Tao <tiantao6@hisilicon.com>
Link: https://lore.kernel.org/r/1619484575-26098-1-git-send-email-tiantao6@hisilicon.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
SYSRQ doesn't work with DMA. This is because there is no error
indication whether a symbol had a framing error or not. Actually,
this is not completely correct, there is a bit in the data register
which is set in this case, but we'd have to read change the DMA access
to 16 bit and we'd need to post process the data, thus make the DMA
pointless in the first place.
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-10-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The LPUART can loop the RX and TX signal. Add support for it.
Please note, this was only tested on the 32 bit version of the LPUART.
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-9-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The LPUART doesn't have the ability to control the RTS or CTS line
manually. Instead it will set it automatically when data is send or
handle it when data is received. Thus drop the wrong code in set_mctrl.
For the 32 bit version this was already done in the commit 2b30efe2e8
("tty: serial: lpuart: Remove unnecessary code from set_mctrl"). Keep
the 8-bit version in sync and remove it there, too.
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-8-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The wrong code in set_mctrl() was already removed in commit 2b30efe2e8
("tty: serial: lpuart: Remove unnecessary code from set_mctrl"), but the
code in get_mctrl() wasn't removed. It will not return the state of the
RTS or CTS line but whether automatic flow control is enabled, which is
wrong for the get_mctrl(). Thus remove it.
Fixes: 2b30efe2e8 ("tty: serial: lpuart: Remove unnecessary code from set_mctrl")
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-7-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Although there is already sysrq characters handling, a break condition
was never detected. Add support for it.
The LPUART can't distinguish between a framing error and a break
condition. We assume it is a break if the received data is all zero.
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-6-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Instead of uart_handle_sysrq_char() use uart_prepare_sysrq_char() and
uart_unlock_and_check_sysrq(). This will call handle_sysrq() without
holding the port lock, which in turn let us drop the spin_trylock hack.
Suggested-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-5-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>From commit 75f4e830fa ("serial: do not restore interrupt state in
sysrq helper"):
Since commit 81e2073c17 ("genirq: Disable interrupts for force
threaded handlers") interrupt handlers that are not explicitly
requested as threaded are always called with interrupts disabled and
there is no need to save the interrupt state when taking the port
lock.
Apply this also to fsl_lpuart in prepartion for sysrq handling with
uart_unlock_and_check_sysrq().
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-4-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use the corresponding macro instead of the magic number. While at it,
drop the useless cast to "unsigned char".
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-3-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
lpuart_rx_dma_startup() is used for both the 8 bit and the 32 bit
version of the LPUART. Modify the UARTCR only for the 8 bit version.
Fixes: f4eef224a0 ("serial: fsl_lpuart: add sysrq support when using dma")
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20210512141255.18277-2-michael@walle.cc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We should be very careful about the register values that will be used
for division or modulo operations, althrough the possibility that the
UARTBAUD register value is zero is very low, but we had better to deal
with the "bad data" of hardware in advance to avoid division or modulo
by zero leading to undefined kernel behavior.
Signed-off-by: Sherry Sun <sherry.sun@nxp.com>
Link: https://lore.kernel.org/r/20210427021226.27468-1-sherry.sun@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use map_urb_for_dma() to improve the dma map code for single step
set feature request urb in test mode.
Signed-off-by: Li Jun <jun.li@nxp.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Link: https://lore.kernel.org/r/1620452039-11694-3-git-send-email-jun.li@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It is needed at USB Certification test for Embedded Host 2.0, and
the detail is at CH6.4.1.1 of On-The-Go and Embedded Host Supplement
to the USB Revision 2.0 Specification. Since other USB 2.0 capable
host like XHCI also need it, so move it to HCD core.
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
Link: https://lore.kernel.org/r/1620452039-11694-1-git-send-email-jun.li@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There's no separate low power (LP) version of Elkhart Lake, thus
this patch updates the PCI Device ID DEFINE to indicate this.
Acked-by: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Raymond Tan <raymond.tan@intel.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Link: https://lore.kernel.org/r/20210512135901.28495-1-heikki.krogerus@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There are two occurrances where objects are being free'd via
a put call and yet they are being referenced after this. Fix these
by adding in the missing continue statement so that the put on the
end of the loop is skipped over.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Addresses-Coverity: ("Use after free")
Fixes: 11e8f5fd22 ("drm: Add simpledrm driver")
Link: https://patchwork.freedesktop.org/patch/msgid/20210512203051.299026-1-colin.king@canonical.com
This may happen if the port becomes resume status exactly
when usb_port_resume() gets port status, it still need provide
a TRSMCRY time before access the device.
CC: <stable@vger.kernel.org>
Reported-by: Tianping Fang <tianping.fang@mediatek.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
Link: https://lore.kernel.org/r/20210512020738.52961-1-chunfeng.yun@mediatek.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add the missing unlock before return from function devm_aperture_acquire()
in the error handling case.
Signed-off-by: Zou Wei <zou_wei@huawei.com>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 730e7992dc ("drm/aperture: Add infrastructure for aperture ownership")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Maxime Ripard <maxime@cerno.tech>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Link: https://patchwork.freedesktop.org/patch/msgid/1620895564-52367-1-git-send-email-zou_wei@huawei.com
struct resource start and end fields are not always long long,
so using %llx to print them can cause build warnings (below).
Fix these by using the special "%pr" for printing struct resource info.
../drivers/gpu/drm/tiny/simpledrm.c: In function ‘simpledrm_device_init_mm’:
../include/drm/drm_print.h:412:32: warning: format ‘%llx’ expects argument of type ‘long long unsigned int’, but argument 3 has type ‘resource_size_t {aka unsigned int}’ [-Wformat=]
../drivers/gpu/drm/tiny/simpledrm.c:533:54: note: format string is defined here
drm_err(dev, "could not acquire memory range [0x%llx:0x%llx]: error %d\n",
~~~^
%x
../include/drm/drm_print.h:412:32: warning: format ‘%llx’ expects argument of type ‘long long unsigned int’, but argument 4 has type ‘resource_size_t {aka unsigned int}’ [-Wformat=]
../drivers/gpu/drm/tiny/simpledrm.c:533:61: note: format string is defined here
drm_err(dev, "could not acquire memory range [0x%llx:0x%llx]: error %d\n",
~~~^
%x
Fixes: 4aae79f77e ("drm/simpledrm: Acquire memory aperture for framebuffer")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210512233459.19534-1-rdunlap@infradead.org
This is a regression introduced by 1373fefc62 ("usb: typec: tcpm:
Allow slow charging loops to comply to pSnkStby")
When Source advertises Rp-default, tcpm would request 500mA when in
SINK_DISCOVERY, Type-C spec advises the sink to follow BC1.2 current
limits when Rp-default is advertised.
[12750.503381] Requesting mux state 1, usb-role 2, orientation 1
[12750.503837] state change SNK_ATTACHED -> SNK_STARTUP [rev3 NONE_AMS]
[12751.003891] state change SNK_STARTUP -> SNK_DISCOVERY
[12751.003900] Setting voltage/current limit 5000 mV 500 mA
This patch restores the behavior where the tcpm would request 0mA when
Rp-default is advertised by the source.
[ 73.174252] Requesting mux state 1, usb-role 2, orientation 1
[ 73.174749] state change SNK_ATTACHED -> SNK_STARTUP [rev3 NONE_AMS]
[ 73.674800] state change SNK_STARTUP -> SNK_DISCOVERY
[ 73.674808] Setting voltage/current limit 5000 mV 0 mA
During SNK_DISCOVERY, Cap the current limit to PD_P_SNK_STDBY_MW / 5 only
for slow_charger_loop case.
Fixes: 1373fefc62 ("usb: typec: tcpm: Allow slow charging loops to comply to pSnkStby")
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Signed-off-by: Badhri Jagan Sridharan <badhri@google.com>
Link: https://lore.kernel.org/r/20210510211756.3346954-1-badhri@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The pdev field got removed from struct drm_device recently. Replace
the invalid reference with an upcast from the struct's dev field.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Fixes: b347e04452 ("drm: Remove pdev field from struct drm_device")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Maxime Ripard <maxime@cerno.tech>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Link: https://patchwork.freedesktop.org/patch/msgid/20210512185527.26050-1-tzimmermann@suse.de
It should have been an OVERLAY from the beginning. The documentation
stipulates that there should be an unique PRIMARY plane per CRTC.
Fixes: fc1acf317b ("drm/ingenic: Add support for the IPU")
Cc: <stable@vger.kernel.org> # 5.8+
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Simon Ser <contact@emersion.fr>
Link: https://patchwork.freedesktop.org/patch/msgid/20210329175046.214629-2-paul@crapouillou.net
One of AMD xhci controller require reset on resume.
Occasionally AMD xhci controller does not respond to
Stop endpoint command.
Once the issue happens controller goes into bad state
and in that case controller needs to be reset.
Cc: <stable@vger.kernel.org>
Signed-off-by: Sandeep Singh <sandeep.singh@amd.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20210512080816.866037-6-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On some devices (specifically the SC8180x based Surface Pro X with
QCOM04A6) HC halt / xhci_halt() times out during boot. Manually binding
the xhci-hcd driver at some point later does not exhibit this behavior.
To work around this, double XHCI_MAX_HALT_USEC, which also resolves this
issue.
Cc: <stable@vger.kernel.org>
Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20210512080816.866037-5-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
'xhci_urb_enqueue()' is passed a 'mem_flags' argument, because "URBs may be
submitted in interrupt context" (see comment related to 'usb_submit_urb()'
in 'drivers/usb/core/urb.c')
So this flag should be used in all the calling chain.
Up to now, 'xhci_check_maxpacket()' which is only called from
'xhci_urb_enqueue()', uses GFP_KERNEL.
Be safe and pass the mem_flags to this function as well.
Fixes: ddba5cd0ae ("xhci: Use command structures when queuing commands on the command ring")
Cc: <stable@vger.kernel.org>
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20210512080816.866037-4-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 9ebf300078 ("xhci: Fix halted endpoint at stop endpoint command
completion") in 5.12 changes how cancelled URBs are given back.
To cancel a URB xhci driver needs to stop the endpoint first.
To clear a halted endpoint xhci driver needs to reset the endpoint.
In rare cases when an endpoint halt (error) races with a endpoint stop we
need to clear the reset before removing, and giving back the cancelled URB.
The above change in 5.12 takes care of this, but it also relies on the
reset endpoint completion handler to give back the cancelled URBs.
There are cases when driver refuses to queue reset endpoint commands,
for example when a link suddenly goes to an inactive error state.
In this case the cancelled URB is never given back.
Fix this by giving back the URB in the stop endpoint if queuing a reset
endpoint command fails.
Fixes: 9ebf300078 ("xhci: Fix halted endpoint at stop endpoint command completion")
CC: <stable@vger.kernel.org> # 5.12
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20210512080816.866037-3-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the same way as Intel Tiger Lake TCSS (Type-C Subsystem) the Alder Lake
TCSS xHCI needs to be runtime suspended whenever possible to allow the
TCSS hardware block to enter D3cold and thus save energy.
Cc: stable@vger.kernel.org
Signed-off-by: Abhijeet Rao <abhijeet.rao@intel.com>
Signed-off-by: Nikunj A. Dadhania <nikunj.dadhania@intel.com>
Signed-off-by: Azhar Shaikh <azhar.shaikh@intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/20210512080816.866037-2-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Call tick_nohz_task_switch() slightly earlier after the context switch
to benefit from disabled IRQs. This way the function doesn't need to
disable them once more.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210512232924.150322-10-frederic@kernel.org
When the tick dependency of a task is updated, we want it to aknowledge
the new state and restart the tick if needed. If the task is not
running, we don't need to kick it because it will observe the new
dependency upon scheduling in. But if the task is running, we may need
to send an IPI to it so that it gets notified.
Unfortunately we don't have the means to check if a task is running
in a race free way. Checking p->on_cpu in a synchronized way against
p->tick_dep_mask would imply adding a full barrier between
prepare_task_switch() and tick_nohz_task_switch(), which we want to
avoid in this fast-path.
Therefore we blindly fire an IPI to the task's CPU.
Meanwhile we can check if the task is queued on the CPU rq because
p->on_rq is always set to TASK_ON_RQ_QUEUED _before_ schedule() and its
full barrier that precedes tick_nohz_task_switch(). And if the task is
queued on a nohz_full CPU, it also has fair chances to be running as the
isolation constraints prescribe running single tasks on full dynticks
CPUs.
So use this as a trick to check if we can spare an IPI toward a
non-running task.
NOTE: For the ordering to be correct, it is assumed that we never
deactivate a task while it is running, the only exception being the task
deactivating itself while scheduling out.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210512232924.150322-9-frederic@kernel.org
Rather than waking up all nohz_full CPUs on the system, only wake up
the target CPUs of member threads of the signal.
Reduces interruptions to nohz_full CPUs.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210512232924.150322-8-frederic@kernel.org
When adding a tick dependency to a task, its necessary to
wake up the CPU where the task resides to reevaluate tick
dependencies on that CPU.
However the current code wakes up all nohz_full CPUs, which
is unnecessary.
Switch to waking up a single CPU, by using ordering of writes
to task->cpu and task->tick_dep_mask.
[ mingo: Minor readability edit. ]
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210512232924.150322-7-frederic@kernel.org
CONFIG_NO_HZ_FULL behaves just like CONFIG_NO_HZ_IDLE by default.
Reassure distros about it.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210512232924.150322-6-frederic@kernel.org
The idle_exittime field of tick_sched is used to record the time when
the idle state was left. but currently the idle_exittime is updated in
the function tick_nohz_restart_sched_tick(), which is not always in idle
state when nohz_full is configured:
tick_irq_exit
tick_nohz_irq_exit
tick_nohz_full_update_tick
tick_nohz_restart_sched_tick
ts->idle_exittime = now;
It's thus overwritten by mistake on nohz_full tick restart. Move the
update to the appropriate idle exit path instead.
Signed-off-by: Yunfeng Ye <yeyunfeng@huawei.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210512232924.150322-5-frederic@kernel.org
The vtime_accounting_enabled_this_cpu() early check already makes what
follows as dead code in the case of CONFIG_VIRT_CPU_ACCOUNTING_NATIVE.
No need to keep the ifdeferry around.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210512232924.150322-4-frederic@kernel.org
In nohz_full mode, switching from idle to a task will unconditionally
issue a tick restart. If the task is alone in the runqueue or is the
highest priority, the tick will fire once then eventually stop. But that
alone is still undesired noise.
Therefore, only restart the tick on idle exit when it's strictly
necessary.
Signed-off-by: Yunfeng Ye <yeyunfeng@huawei.com>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20210512232924.150322-3-frederic@kernel.org
When tick_nohz_full_cpu() is called with smp_processor_id(), the latter
is unconditionally evaluated whether the static key is on or off. It is
not necessary in the off-case though, so make sure the cpu expression
is executed at the last moment.
Illustrate with the following test function:
int tick_nohz_test(void)
{
return tick_nohz_full_cpu(smp_processor_id());
}
The resulting code before was:
mov %gs:0x7eea92d1(%rip),%eax # smp_processor_id() fetch
nopl 0x0(%rax,%rax,1)
xor %eax,%eax
retq
cmpb $0x0,0x29d393a(%rip) # <tick_nohz_full_running>
je tick_nohz_test+0x29 # jump to below eax clear
mov %eax,%eax
bt %rax,0x29d3936(%rip) # <tick_nohz_full_mask>
setb %al
movzbl %al,%eax
retq
xor %eax,%eax
retq
Now it becomes:
nopl 0x0(%rax,%rax,1)
xor %eax,%eax
retq
cmpb $0x0,0x29d3871(%rip) # <tick_nohz_full_running>
je tick_nohz_test+0x29 # jump to below eax clear
mov %gs:0x7eea91f0(%rip),%eax # smp_processor_id() fetch, after static key
mov %eax,%eax
bt %rax,0x29d3866(%rip) # <tick_nohz_full_mask>
setb %al
movzbl %al,%eax
retq
xor %eax,%eax
retq
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210512232924.150322-2-frederic@kernel.org
We have a mismatch between RCU and isolation -- in relation to what is
considered the maximum valid CPU number.
This matters because nohz_full= and rcu_nocbs= are joined at the hip; in
fact the former will enforce the latter. So we don't want a CPU mask to
be valid for one and denied for the other.
The difference 1st appeared as of v4.15; further details are below.
As it is confusing to anyone who isn't looking at the code regularly, a
reminder is in order; three values exist here:
CONFIG_NR_CPUS - compiled in maximum cap on number of CPUs supported.
nr_cpu_ids - possible # of CPUs (typically reflects what ACPI says)
cpus_present - actual number of present/detected/installed CPUs.
For this example, I'll refer to NR_CPUS=64 from "make defconfig" and
nr_cpu_ids=6 for ACPI reporting on a board that could run a six core,
and present=4 for a quad that is physically in the socket. From dmesg:
smpboot: Allowing 6 CPUs, 2 hotplug CPUs
setup_percpu: NR_CPUS:64 nr_cpumask_bits:64 nr_cpu_ids:6 nr_node_ids:1
rcu: RCU restricting CPUs from NR_CPUS=64 to nr_cpu_ids=6.
smp: Brought up 1 node, 4 CPUs
And from userspace, see:
paul@trash:/sys/devices/system/cpu$ cat present
0-3
paul@trash:/sys/devices/system/cpu$ cat possible
0-5
paul@trash:/sys/devices/system/cpu$ cat kernel_max
63
Everything is fine if we boot 5x5 for rcu/nohz:
Command line: BOOT_IMAGE=/boot/bzImage nohz_full=2-5 rcu_nocbs=2-5 root=/dev/sda1 ro
NO_HZ: Full dynticks CPUs: 2-5.
rcu: Offload RCU callbacks from CPUs: 2-5.
..even though there is no CPU 4 or 5. Both RCU and nohz_full are OK.
Now we push that > 6 but less than NR_CPU and with 15x15 we get:
Command line: BOOT_IMAGE=/boot/bzImage rcu_nocbs=2-15 nohz_full=2-15 root=/dev/sda1 ro
rcu: Note: kernel parameter 'rcu_nocbs=', 'nohz_full', or 'isolcpus=' contains nonexistent CPUs.
rcu: Offload RCU callbacks from CPUs: 2-5.
These are both functionally equivalent, as we are only changing flags on
phantom CPUs that don't exist, but note the kernel interpretation changes.
And worse, it only changes for one of the two - which is the problem.
RCU doesn't care if you want to restrict the flags on phantom CPUs but
clearly nohz_full does after this change from v4.15.
edb9382175: ("sched/isolation: Move isolcpus= handling to the housekeeping code")
- if (cpulist_parse(str, non_housekeeping_mask) < 0) {
- pr_warn("Housekeeping: Incorrect nohz_full cpumask\n");
+ err = cpulist_parse(str, non_housekeeping_mask);
+ if (err < 0 || cpumask_last(non_housekeeping_mask) >= nr_cpu_ids) {
+ pr_warn("Housekeeping: nohz_full= or isolcpus= incorrect CPU range\n");
To be clear, the sanity check on "possible" (nr_cpu_ids) is new here.
The goal was reasonable ; not wanting housekeeping to land on a
not-possible CPU, but note two things:
1) this is an exclusion list, not an inclusion list; we are tracking
non_housekeeping CPUs; not ones who are explicitly assigned housekeeping
2) we went one further in 9219565aa8 ("sched/isolation: Require a present CPU in housekeeping mask")
- ensuring that housekeeping was sanity checking against present and not just possible CPUs.
To be clear, this means the check added in v4.15 is doubly redundant.
And more importantly, overly strict/restrictive.
We care now, because the bitmap boot arg parsing now knows that a value
of "N" is NR_CPUS; the size of the bitmap, but the bitmap code doesn't
know anything about the subtleties of our max/possible/present CPU
specifics as outlined above.
So drop the check added in v4.15 (edb9382175) and make RCU and
nohz_full both in alignment again on NR_CPUS so "N" works for both,
and then they can fall back to nr_cpu_ids internally just as before.
Command line: BOOT_IMAGE=/boot/bzImage nohz_full=2-N rcu_nocbs=2-N root=/dev/sda1 ro
NO_HZ: Full dynticks CPUs: 2-5.
rcu: Offload RCU callbacks from CPUs: 2-5.
As shown above, with this change, RCU and nohz_full are in sync, even
with the use of the "N" placeholder. Same result is achieved with "15".
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
Link: https://lore.kernel.org/r/20210419042659.1134916-1-paul.gortmaker@windriver.com