John Youn [Thu, 17 Dec 2015 19:16:31 +0000 (11:16 -0800)]
usb: dwc2: Fix dr_mode validation
The dr_mode parameter was being checked against how the dwc2 module
was being configured at compile time. But it wasn't checked against
the hardware capabilities, nor were the hardware capabilities checked
against the compilation parameters.
This commit adds those checks and adjusts dr_mode to an appropriate
value, if needed. If the hardware capabilities and module compilation
do not match then we fail as it wouldn't be possible to run properly.
The hardware, module, and dr_mode, can each be set to host, device,
or otg. Check that all these values are compatible and adjust the
value of dr_mode if possible.
The following table summarizes the behavior:
actual
HW MOD dr_mode dr_mode
------------------------------
HST HST any : HST
HST DEV any : ---
HST OTG any : HST
DEV HST any : ---
DEV DEV any : DEV
DEV OTG any : DEV
OTG HST any : HST
OTG DEV any : DEV
OTG OTG any : dr_mode
Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
John Youn [Thu, 17 Dec 2015 19:16:17 +0000 (11:16 -0800)]
usb: dwc2: Add functions to check the HW OTG config
Added functions to query the GHWCFG2.OTG_MODE. This tells us whether the
controller hardware is configured for OTG, device-only, or host-only.
Signed-off-by: John Youn <johnyoun@synopsys.com> Tested-by: Douglas Anderson <dianders@chromium.org> Reviewed-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Felipe Balbi <balbi@ti.com>
John Youn [Thu, 17 Dec 2015 19:16:03 +0000 (11:16 -0800)]
usb: dwc2: Add dwc2_core_reset()
dwc2_core_reset() was previously renamed to
dwc2_core_reset_and_dr_force_mode(). Now add back dwc2_core_reset() which
performs only a basic core reset without forcing the mode.
Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
John Youn [Thu, 17 Dec 2015 19:15:49 +0000 (11:15 -0800)]
usb: dwc2: Rename dwc2_core_reset()
Renamed dwc2_core_reset() to dwc2_core_reset_and_force_dr_mode(). This
describes what it is doing more accurately. This is in preparation of
introducing a plain dwc2_core_reset() function that only performs the
reset and doesn't force the mode.
Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
John Youn [Thu, 17 Dec 2015 19:15:35 +0000 (11:15 -0800)]
usb: dwc2: Reorder AHBIDLE and CSFTRST in dwc2_core_reset()
According to the databook, the core soft reset should be done before
checking for AHBIDLE. The gadget version of core reset had it correct
but the hcd version did not. This fixes the hcd version.
Signed-off-by: John Youn <johnyoun@synopsys.com> Reviewed-by: Douglas Anderson <dianders@chromium.org> Tested-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Thu, 17 Dec 2015 19:15:21 +0000 (11:15 -0800)]
usb: dwc2: Avoid more calls to dwc2_core_reset()
Calls to dwc2_core_reset() are currently very slow, taking at least
150ms (possibly more). It behooves us to take as many of these calls
out as possible.
It turns out that the calls in dwc2_fs_phy_init() and dwc2_hs_phy_init()
should (as documented in the code) only be needed if we need to do a PHY
SELECT. That means that if we see that we can avoid the PHY SELECT then
we can avoid the reset.
This patch appears to successfully bypass two resets (one per USB
device) on rk3288-based ARM Chromebooks.
Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Yunzhi Li [Thu, 17 Dec 2015 19:15:08 +0000 (11:15 -0800)]
usb: dwc2: reduce dwc2 driver probe time
I found that the probe function of dwc2 driver takes much time
when kernel boot up. There are many long delays in the probe
function these take almost 1 second.
This patch trying to reduce unnecessary delay time.
In dwc2_core_reset() I see it use two at least 20ms delays to
wait AHB idle and core soft reset, but dwc2 data book said that
dwc2 core soft reset and AHB idle just need a few clocks (I think
it refers to AHB clock, and AHB clock run at 150MHz in my RK3288
board), so 20ms is too long, delay 1us for wait AHB idle and soft
reset is enough.
And in dwc2_get_hwparams() it takes 150ms to wait ForceHostMode
and ForceDeviceMode valid but in data book it said software must
wait at least 25ms before the change to take effect, so I reduce
this time to 25ms~50ms. By the way, is there any state bit show
that the force mode take effect ? Could we poll curmod bit for
figuring out if the change take effect ?
It seems that usleep_range() at boot time will pick the longest
value in the range. In dwc2_core_reset() there is a very long
delay takes 200ms, and this function run twice when probe, could
any one tell me is this delay time resonable ?
I have tried this patch in my RK3288-evb board. It works well.
Signed-off-by: Yunzhi Li <lyz@rock-chips.com> Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Thu, 17 Dec 2015 19:14:54 +0000 (11:14 -0800)]
usb: dwc2: Speed dwc2_get_hwparams() on some host-only ports
On some host-only DWC2 ports (like the one in rk3288) when we set
GUSBCFG_FORCEHOSTMODE in GUSBCFG and then read back, we don't see the
bit set. Presumably that's because the port is always forced to HOST
mode so there's no reason to implement these status bits.
Since we know dwc2_core_reset() is always called before
dwc2_get_hwparams() and we know dwc2_core_reset() should have set
GUSBCFG_FORCEHOSTMODE whenever hsotg->dr_mode == USB_DR_MODE_HOST, we
can just check hsotg->dr_mode to decide that we can skip the delays in
dwc2_get_hwparams().
Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Thu, 17 Dec 2015 19:14:40 +0000 (11:14 -0800)]
usb: dwc2: Avoid double-reset at boot time
In (usb: dwc2: reset dwc2 core before dwc2_get_hwparams()) we added an
extra reset to the probe path for the dwc2 USB controllers. This
allowed proper detection of parameters even if the firmware had already
used the USB part.
Unfortunately, this extra reset is quite slow and is affecting boot
speed. We can avoid the double-reset by skipping the extra reset that
would happen just after the one we added. Logic that explains why this
is safe:
* As of the CL mentioned above, we now always call dwc2_core_reset() in
dwc2_driver_probe() before dwc2_hcd_init().
* The only caller of dwc2_hcd_init() is dwc2_driver_probe(), so we're
guaranteed that dwc2_core_reset() was called before dwc2_hdc_init().
* dwc2_hdc_init() is the only caller that passes an irq other than -1 to
dwc2_core_init(). Thus if dwc2_core_init() is called with an irq
other than -1 we're guaranteed that dwc2_core_reset was called before
dwc2_core_init().
...this allows us to remove the dwc2_core_reset() in dwc2_core_init() if
irq is not < 0.
Note that since "irq" wasn't used in the function dwc2_core_init()
anyway and since select_phy was always set at exactly the same times we
could avoid the reset, we remove "irq" and rename "select_phy" to
"initial_setup" and adjust the callers accordingly.
Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Yunzhi Li [Thu, 17 Dec 2015 19:14:26 +0000 (11:14 -0800)]
usb: dwc2: reset dwc2 core before dwc2_get_hwparams()
We initiate dwc2 usb controller in BIOS, dwc2_core_reset() should
be called before dwc2_get_hwparams() to reset core registers to
default value. Without this the FIFO setting might be incorrect
because calculating FIFO size need power-on value of
GRXFSIZ/GNPTXFSIZ/HPTXFSIZ registers.
This patch could avoid warnning massage like in rk3288 platform:
[ 2.074764] dwc2 ff580000.usb: 256 invalid for
host_perio_tx_fifo_size. Check HW configuration.
Signed-off-by: Yunzhi Li <lyz@rock-chips.com> Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Thu, 17 Dec 2015 19:14:12 +0000 (11:14 -0800)]
usb: dwc2: Restore GUSBCFG in dwc2_get_hwparams()
Previously dwc2_get_hwparams() was changing GUSBCFG and not putting it
back the way it was (specifically it set and cleared FORCEHOSTMODE).
Since we want to move dwc2_core_reset() _before_ dwc2_get_hwparams() we
should make sure dwc2_get_hwparams() isn't messing with things in a
permanent way.
Since we're now looking at GUSBCFG, it's obvious that we shouldn't need
all the extra delays if FORCEHOSTMODE was already set. This will avoid
some delays for any ports that have forced host mode.
Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Mon, 7 Dec 2015 11:07:31 +0000 (12:07 +0100)]
usb: dwc2: gadget: don't overwrite DCTL register on NAKEFF interrupts
When receiving GINTSTS_GINNAKEFF or GINTSTS_GOUTNAKEFF interrupt,
DCTL will be overwritten with DCTL_CGOUTNAK or DCTL_CGNPINNAK values.
Instead of overwriting it, write only needed bits.
It could cause an issue if GINTSTS_GINNAKEFF or GINTSTS_GOUTNAKEFF
interrupt is received after dwc2 disabled pullup by writing
DCTL_SFTDISCON bit.
Pullup will then be re-enabled whereas it should not.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Du, Changbin [Fri, 4 Dec 2015 07:38:23 +0000 (15:38 +0800)]
usb: dwc2: fix transfer stop programming for out endpoint
To stop an out endpoint, software should set sets the Global OUT NAK,
but not the Global Non-periodic IN NAK. This driver bug leads the out-ep
failed be in disabled state with below error.
dwc2_hsotg_ep_stop_xfr: timeout DOEPCTL.EPDisable
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Du, Changbin <changbin.du@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Baolin Wang [Sat, 21 Nov 2015 07:44:53 +0000 (15:44 +0800)]
usb: gadget: Add the console support for usb-to-serial port
It dose not work when we want to use the usb-to-serial port based
on one usb gadget as a console. Thus this patch adds the console
initialization to support this request.
To avoid the re-entrance when transferring data with usb endpoint,
it introduces a kthread to do the IO transmission.
Signed-off-by: Baolin Wang <baolin.wang@linaro.org> Signed-off-by: Felipe Balbi <balbi@ti.com>
Felipe Balbi [Thu, 17 Dec 2015 15:55:41 +0000 (09:55 -0600)]
usb: of: fix build breakage on !OF
If OF is disabled, we will try to define a stub for
of_usb_get_dr_mode_by_phy(), however that missed a
static inline annotation which made us redefine the
stub over and over again. Fix that.
Fixes: 98bfb3946695 ("usb: of: add an api to get
dr_mode by the phy node") Signed-off-by: Felipe Balbi <balbi@ti.com>
David Eccher [Fri, 11 Dec 2015 21:13:55 +0000 (22:13 +0100)]
usb: gadget: inode.c: fix unbalanced spin_lock in ep0_write
Fix bad unlock balance: ep0_write enter with the locks locked from
inode.c:1769, hence it must exit with spinlock held to avoid double
unlock in dev_config.
Signed-off-by: David Eccher <d.eccher@gmail.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Simon Horman [Wed, 16 Dec 2015 04:51:52 +0000 (13:51 +0900)]
usb: renesas_usbhs: add device tree support for r8a779[23]
Simply document new compatibility string.
As a previous patch adds a generic R-Car Gen2 compatibility string
there appears to be no need for a driver updates.
Signed-off-by: Simon Horman <horms+renesas@verge.net.au> Acked-by: Rob Herring <robh@kernel.org> Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Add fallback compatibility strings for R-Car Gen2 and Gen3.
This is in keeping with the fallback scheme being adopted wherever
appropriate for drivers for Renesas SoCs.
Signed-off-by: Simon Horman <horms+renesas@verge.net.au> Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Mike Looijmans [Mon, 30 Nov 2015 11:18:23 +0000 (12:18 +0100)]
usb: gadget: ether: Allow changing the MTU
The gadget ethernet driver supports changing the MTU, but only allows this
when the USB cable is removed. The comment indicates that this is because
the "peer won't know". Even if the network link is still down and only the
USB link is established, the driver won't allow the change.
Other network interfaces allow changing the MTU any time, and don't force
the link to be disabled. This makes perfect sense, because in order to be
able to negotiate the MTU, the link needs to be up.
Remove the restriction so that it is now actually possible to change the
MTU (e.g. using "ifconfig usb0 mtu 15000") without having to manually pull
the plug or change the driver's default setting.
This is especially important after commit bba787a860fa
("usb: gadget: ether: Allow jumbo frames")
Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl> Signed-off-by: Felipe Balbi <balbi@ti.com>
This patch introduces pre-allocation of IN endpoint USB requests. This
improves on latency (requires no usb request allocation on transmit) and avoid
several potential probles on allocating too many usb requests (which involves
DMA pool allocation problems).
This implementation also handles better multiple MIDI Gadget ports, always
processing the last processed MIDI substream if the last USB request wasn't
enought to handle the whole stream.
Signed-off-by: Felipe F. Tonello <eu@felipetonello.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
usb: gadget: f_midi: fail if set_alt fails to allocate requests
This ensures that the midi function will only work if the proper number of
IN and OUT requrests are allocated. Otherwise the function will work with less
requests then what the user wants.
Signed-off-by: Felipe F. Tonello <eu@felipetonello.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Tony Lindgren [Tue, 1 Dec 2015 05:37:13 +0000 (21:37 -0800)]
usb: musb: Fix unbalanced pm_runtime_enable
When reloading omap2430 kernel module we get a warning about
unbalanced pm_runtime_enable. Let's fix this. Note that we
need to do this after the child musb-core platform_device is
removed because of pm_runtime_irq_safe being set at the child.
Cc: Bin Liu <b-liu@ti.com> Cc: Felipe Balbi <balbi@ti.com> Cc: Kishon Vijay Abraham I <kishon@ti.com> Cc: NeilBrown <neil@brown.name> Reviewed-by: Kishon Vijay Abraham I <kishon@ti.com> Signed-off-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Tony Lindgren [Tue, 1 Dec 2015 05:37:12 +0000 (21:37 -0800)]
usb: musb: core: Fix handling of the phy notifications
We currently can't unload omap2430 MUSB platform glue driver module and
this cause issues for fixing the MUSB code further. The reason we can't
remove omap2430 is because it uses the PHY functions and also exports the
omap_musb_mailbox function that some PHY drivers are using.
Let's fix the issue by exporting a more generic musb_mailbox function
from the MUSB core and allow platform glue layers to register phy_callback
function as needed.
And now we can now also get rid of the include/linux/musb-omap.h.
Cc: Bin Liu <b-liu@ti.com> Cc: Felipe Balbi <balbi@ti.com> Cc: Kishon Vijay Abraham I <kishon@ti.com> Cc: NeilBrown <neil@brown.name> Reviewed-by: Kishon Vijay Abraham I <kishon@ti.com> Signed-off-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Ruslan Bilovol [Mon, 23 Nov 2015 08:56:38 +0000 (09:56 +0100)]
usb: gadget: udc-core: independent registration of gadgets and gadget drivers
Change behavior during registration of gadgets and
gadget drivers in udc-core. Instead of previous
approach when for successful probe of usb gadget driver
at least one usb gadget should be already registered
use another one where gadget drivers and gadgets
can be registered in udc-core independently.
Independent registration of gadgets and gadget drivers
is useful for built-in into kernel gadget and gadget
driver case - because it's possible that gadget is
really probed only on late_init stage (due to deferred
probe) whereas gadget driver's probe is silently failed
on module_init stage due to no any UDC added.
Also it is useful for modules case - now there is no
difference what module to insert first: gadget module
or gadget driver one.
Tested-by: Maxime Ripard <maxime.ripard@free-electrons.com> Signed-off-by: Ruslan Bilovol <ruslan.bilovol@gmail.com>
[simplified code as requested by Alan Stern and Felipe Balbi,
fixed checkpatch issues] Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Tested-by: Peter Chen <peter.chen@freescale.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Now when last user of usb_udc_attach_driver() is switched
to passing UDC name via usb_gadget_driver struct, it's safe
to remove this function
Tested-by: Maxime Ripard <maxime.ripard@free-electrons.com> Signed-off-by: Ruslan Bilovol <ruslan.bilovol@gmail.com> Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Tested-by: Peter Chen <peter.chen@freescale.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Ruslan Bilovol [Mon, 23 Nov 2015 08:56:36 +0000 (09:56 +0100)]
usb: gadget: configfs: pass UDC name via usb_gadget_driver struct
Now when udc-core supports binding to specific UDC by passing
its name via 'udc_name' member of usb_gadget_driver struct,
switch to this generic approach.
Tested-by: Maxime Ripard <maxime.ripard@free-electrons.com> Signed-off-by: Ruslan Bilovol <ruslan.bilovol@gmail.com>
[rebased and fixed checkpatch issues] Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Tested-by: Peter Chen <peter.chen@freescale.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Ruslan Bilovol [Mon, 23 Nov 2015 08:56:35 +0000 (09:56 +0100)]
usb: gadget: bind UDC by name passed via usb_gadget_driver structure
Introduce new 'udc_name' member to usb_gadget_driver structure.
The 'udc_name' is a name of UDC that usb_gadget_driver should
be bound to. If udc_name is NULL, it will be bound to any
available UDC.
Tested-by: Maxime Ripard <maxime.ripard@free-electrons.com> Signed-off-by: Ruslan Bilovol <ruslan.bilovol@gmail.com> Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Tested-by: Peter Chen <peter.chen@freescale.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Bin Liu [Tue, 8 Dec 2015 16:31:50 +0000 (10:31 -0600)]
usb: phy: phy-am335x: bypass first VBUS sensing for host-only mode
To prevent VBUS contention, the am335x MUSB phy senses VBUS first before
transitioning to host mode. However, for host-only mode, VBUS could be
directly tied to 5V power rail which could prevent MUSB transitions to
host mode.
This change receives dr_mode of the controller then bypass the first
VBUS sensing for host-only mode, so that MUSB can work in host mode
event if VBUS is tied to 5V.
Signed-off-by: Bin Liu <b-liu@ti.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Bin Liu [Fri, 20 Nov 2015 22:13:07 +0000 (16:13 -0600)]
usb: phy: correct the am335x phy header filename
The filename of am35x-phy-control.h is confusing. The header is used
by the am335x phy driver, but the filename refers to am35x. Even worse
there is indeed another device called am35x but it does not use this
header at all.
Signed-off-by: Bin Liu <b-liu@ti.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
usb: phy: Remove unused Renesas R-Car (Gen1) USB PHY driver
As of commit 3d7608e4c169af03 ("ARM: shmobile: bockw: remove legacy
board file and config"), the Renesas R-Car (Gen1) USB PHY driver is no
longer used.
In theory it could still be used on R-Car Gen1 SoCs, but that would
require adding DT support to the driver. Instead, a new driver using the
generic PHY framework should be written, as was done for R-Car Gen2.
Remove the driver for good.
Acked-by: Simon Horman <horms+renesas@verge.net.au> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Felipe Balbi <balbi@ti.com>
usb: gadget: lpc32xxx_udc: clean up and sort include directives out
Remove mach/irq.h from the list of included headers, there is no
compilation dependency on this include file, and the change is needed
to prevent a compilation failure, when mach/irq.h is removed.
Additionally remove other unneeded includes and sort out their order.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Felipe Balbi [Wed, 2 Dec 2015 16:06:45 +0000 (10:06 -0600)]
usb: dwc3: gadget: handle request->zero
So far, dwc3 has always missed request->zero
handling for every endpoint. Let's implement
that so we can handle cases where transfer must
be finished with a ZLP.
Note that dwc3 is a little special. Even though
we're dealing with a ZLP, we still need a buffer
of wMaxPacketSize bytes; to hide that detail from
every gadget driver, we have a preallocated buffer
of 1024 bytes (biggest bulk size) to use (and
share) among all endpoints.
Reported-by: Ravi B <ravibabu@ti.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Fri, 20 Nov 2015 17:06:28 +0000 (09:06 -0800)]
usb: dwc2: host: Clear interrupts before handling them
In general it is wise to clear interrupts before processing them. If
you don't do that, you can get:
1. Interrupt happens
2. You look at system state and process interrupt
3. A new interrupt happens
4. You clear interrupt without processing it.
This patch was actually a first attempt to fix missing device insertions
as described in (usb: dwc2: host: Fix missing device insertions) and it
did solve some of the signal bouncing problems but not all of
them (which is why I submitted the other patch). Specifically, this
patch itself would sometimes change:
1. hardware sees connect
2. hardware sees disconnect
3. hardware sees connect
4. dwc2_port_intr() - clears connect interrupt
5. dwc2_handle_common_intr() - calls dwc2_hcd_disconnect()
...to:
1. hardware sees connect
2. hardware sees disconnect
3. dwc2_port_intr() - clears connect interrupt
4. hardware sees connect
5. dwc2_handle_common_intr() - calls dwc2_hcd_disconnect()
...but with different timing then sometimes we'd still miss cable
insertions.
In any case, though this patch doesn't fix any (known) problems, it
still seems wise as a general policy to clear interrupt before handling
them.
Note that for dwc2_handle_usb_port_intr(), instead of moving the clear
of PRTINT to the beginning of the function we remove it completely. The
only way to clear PRTINT is to clear the sources that set it in the
first place.
Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Fri, 20 Nov 2015 17:06:27 +0000 (09:06 -0800)]
usb: dwc2: host: Add missing spinlock in dwc2_hcd_reset_func()
The dwc2_hcd_reset_func() function is only ever called directly by a
delayed work function. As such no locks are already held when the
function is called.
Doing a read-modify-write of CPU registers and setting fields in the
main hsotg data structure is a bad idea without locks. Let's add
locks.
The bug was found by code inspection only. It turns out that the
dwc2_hcd_reset_func() is only ever called today if the
"host_support_fs_ls_low_power" parameter is enabled and no code in
mainline enables that parameter. Thus no known issues in mainline are
fixed by this patch, but it's still probably wise to fix the function.
Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Felipe Balbi <balbi@ti.com>
Bin Liu [Tue, 3 Nov 2015 17:51:15 +0000 (11:51 -0600)]
usb: of: add an api to get dr_mode by the phy node
Some USB phy drivers have different handling for the controller in each
dr_mode. But the phy driver does not have visibility to the dr_mode of
the controller.
This adds an api to return the dr_mode of the controller which
associates the given phy node.
Signed-off-by: Bin Liu <b-liu@ti.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Peter Chen [Thu, 19 Nov 2015 07:02:19 +0000 (15:02 +0800)]
usb: gadget: f_sourcesink: quit if usb_ep_queue returns error
Since now, we may have more than one request during the test, and
it is better we just quit once the error occurs instead of try
queueing further requests.
Signed-off-by: Peter Chen <peter.chen@freescale.com> Suggested-by: Krzysztof Opasiak <k.opasiak@samsung.com> Reviewed-by: Krzysztof Opasiak <k.opasiak@samsung.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Igor Kotrasinski [Tue, 20 Oct 2015 16:33:13 +0000 (18:33 +0200)]
usb: gadget: composite: remove redundant bcdUSB setting in legacy
Since composite now overwrites bcdUSB for any gadget, remove
setting it in legacy gadgets. All legacy gadgets set 0x0200, the
same as the value additionally set by composite, so there is no
behaviour change.
Signed-off-by: Igor Kotrasinski <i.kotrasinsk@samsung.com>
Rebase onto current balbi/next
Signed-off-by: Krzysztof Opasiak <k.opasiak@samsung.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Wed, 11 Nov 2015 18:33:52 +0000 (10:33 -0800)]
usb: dwc2: host: Support immediate retries for split transactions
In some cases, like when you've got a "Microsoft Wireless Keyboard 2000"
connected to dwc2 with a hub, expected that we'll get some transfer
errors sometimes. The controller is expected to try at least 3 times
before giving up. See figure "Figure A-67. Normal HS CSPLIT 3 Strikes
Smash" in the USB spec.
The dwc2 controller has a way to support this by using the "EC_MC"
field. The Raspberry Pi driver has logic for setting this right. See
fiq_fsm_queue_split_transaction() in their "dwc_otg_hcd.c". Let's use
the same logic.
After making this change, we no longer get dropped characters from the
above mentioned keyboard. Other devices on the same bus as the keyboard
also behave more properly.
Thanks for Julius Werner for the expert analysis and suggestions.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Felipe Balbi <balbi@ti.com>
Douglas Anderson [Thu, 19 Nov 2015 21:23:14 +0000 (13:23 -0800)]
usb: dwc2: host: Fix missing device insertions
If you've got your interrupt signals bouncing a bit as you insert your
USB device, you might end up in a state when the device is connected but
the driver doesn't know it.
Specifically, the observed order is:
1. hardware sees connect
2. hardware sees disconnect
3. hardware sees connect
4. dwc2_port_intr() - clears connect interrupt
5. dwc2_handle_common_intr() - calls dwc2_hcd_disconnect()
Now you'll be stuck with the cable plugged in and no further interrupts
coming in but the driver will think we're disconnected.
We'll fix this by checking for the missing connect interrupt and
re-connecting after the disconnect is posted. We don't skip the
disconnect because if there is a transitory disconnect we really want to
de-enumerate and re-enumerate.
Notes:
1. As part of this change we add a "force" parameter to
dwc2_hcd_disconnect() so that when we're unloading the module we
avoid the new behavior. The need for this was pointed out by John
Youn.
2. The bit of code needed at the end of dwc2_hcd_disconnect() is
exactly the same bit of code from dwc2_port_intr(). To avoid
duplication, we refactor that code out into a new function
dwc2_hcd_connect().
Signed-off-by: Douglas Anderson <dianders@chromium.org> Acked-by: John Youn <johnyoun@synopsys.com> Tested-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Fri, 20 Nov 2015 10:49:31 +0000 (11:49 +0100)]
usb: dwc2: host: use kmem cache to allocate descriptors
Kmem caches help to get correct boundary for descriptor buffers
which need to be 512 bytes aligned for dwc2 controller.
Two kmem caches are needed for generic descriptors and for
hs isochronous descriptors which doesn't have same size.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Fri, 20 Nov 2015 10:49:29 +0000 (11:49 +0100)]
usb: dwc2: host: avoid usage of dma_alloc_coherent with irqs disabled
Use Streaming DMA mappings to handle cache coherency of frame list and
descriptor list. Cache are always flushed before controller access it
or before cpu access it.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
usb: dwc2: host: enable descriptor dma for fs devices
As descriptor dma mode does not support split transfers, it can't be
enabled for high speed devices. Add a core parameter to enable it for
full speed devices.
Ensure frame list and descriptor list are correctly freed during
disconnect.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Mian Yousaf Kaukab <yousaf.kaukab@intel.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Deepa Dinamani [Wed, 4 Nov 2015 23:29:11 +0000 (15:29 -0800)]
usb: misc: usbtest: Remove timeval usage
timeval is deprecated and not y2038 safe. Its size also changes according
to 32 bit/ 64 bit compilation. Replace it with 32 and 64 bit versions of
its individual fields, giving two ioctls with different code values.
The two ioctls are necessary to maintain the 32 bit and 64 bit userspace
compatibility with a 64/32 bit kernel.
Change unsigned to __u32 types for a definitive userspace interface.
This is in accordance with the psABI that the unsigned type is always
32 bits.
Also use motonic timer instead of real time to ensure positive delta
values.
Refactor usbtest_ioctl for readability to isolate the handling of the
testing timing measurement.
The official testusb userspace tool can be changed in a separate patch
to reflect the __u32 changes as well. It can use the usbtest_param_32
struct, since 32 bit seconds is long enough for test durations.
usb: renesas_usbhs: Modify ep.caps.type_xxx and usb_ep_maxpacket_limit()
This patch modifies the ep.caps.type_{iso,bulk,int} setting and
the second argument of usb_ep_maxpacket_limit() using
the dparam.pipe_configs.
In the previous code, all the type_{iso,bulk,int} were set to true.
However, to avoid waste time for finding suitable pipe in usb_ep_enable(),
this driver should set correct type.
Also the second argument of usb_ep_maxpacket_limit() was set to 512
even if the pipe is isochronous or interrupt. So, this driver could
not bind a gadget driver like the g_audio driver.
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
The current code has info->bufnmb_last to calculate the BUFNMB bits of
PIPEBUF register. However, since the bufnmb_last is initialized in
the usbhs_pipe_init() only, this driver is possible to set unexpected
value to the register if usb_ep_{enable,disable}() are called many times.
So, this patch modifies the pipe configuration via struct
renesas_usbhs_driver_param to simplify the code. Also this patch changes:
- a double buffer configuration
- isochronous buffer size from 512 to 1024
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Thu, 5 Nov 2015 08:41:43 +0000 (09:41 +0100)]
usb: dwc2: host: program descriptor for next frame
Isochronous descriptor is currently programmed for the frame
after the last descriptor was programmed.
If the last descriptor frame underrun, then current descriptor must
take this into account and must be programmed on the current frame + 1.
This overrun usually happens when system is loaded and dwc2 can't init
descriptor list in time.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Thu, 5 Nov 2015 08:41:40 +0000 (09:41 +0100)]
usb: dwc2: host: fix use of qtd after free in desc dma mode
When completing non isoc xfer, dwc2_complete_non_isoc_xfer_ddma()
is relying on qtd->n_desc to process the corresponding number of
descriptors.
During the processing of these descriptors, qtd could be unlinked
and freed if xfer is done and urb is no more in progress.
In this case, dwc2_complete_non_isoc_xfer_ddma() will read again
qtd->n_desc whereas qtd has been freed. This will lead to unpredictable
results since qtd->n_desc is no more valid value.
To avoid this error, return a result != 0 in dwc2_process_non_isoc_desc(),
so that dwc2_complete_non_isoc_xfer_ddma() stops desc processing.
This has been seen with Slub debug enabled.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Thu, 5 Nov 2015 08:41:39 +0000 (09:41 +0100)]
usb: dwc2: host: rework isochronous halt path
When a channel is halted because of urb dequeue during transfer
completion, no other qtds must be scheduled until halt is done.
Moreover, all in progress qtds must be given back.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Thu, 5 Nov 2015 08:41:38 +0000 (09:41 +0100)]
usb: dwc2: host: set active bit in isochronous descriptors
Active bit must be enabled in all scheduled descriptors. Else transfer
never start.
Remove previous code which was not correctly configuring descriptors.
Active bit was set before calling dwc2_fill_host_isoc_dma_desc() which
is erasing dma_desc->status.
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Gregory Herrero [Thu, 5 Nov 2015 08:41:37 +0000 (09:41 +0100)]
usb: dwc2: host: ensure filling of isoc desc is correctly done
Increment qtd->isoc_frame_index_last before testing it, else below
check will never be true and IOC (Interrupt On Complete) bit for
last frame will never be set in descriptor status.
/* Set IOC for each descriptor corresponding to last frame of URB */
if (qtd->isoc_frame_index_last == qtd->urb->packet_count)
dma_desc->status |= HOST_DMA_IOC;
Acked-by: John Youn <johnyoun@synopsys.com> Signed-off-by: Gregory Herrero <gregory.herrero@intel.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Peter Chen [Thu, 19 Nov 2015 07:02:16 +0000 (15:02 +0800)]
usb: gadget: f_sourcesink: add queue depth
Add queue depth for both iso and bulk transfer, with more queues, we
can do performance and stress test using sourcesink, and update g_zero
accordingly.
Reviewed-by: Krzysztof Opasiak <k.opasiak@samsung.com> Signed-off-by: Peter Chen <peter.chen@freescale.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
Felipe Balbi [Wed, 18 Nov 2015 19:15:20 +0000 (13:15 -0600)]
usb: dwc3: add generic OF glue layer
For simple platforms which merely enable some clocks
and populate its children, we can use this generic
glue layer to avoid boilerplate code duplication.
For now this supports Qcom and Xilinx, but if we
find a way to add generic handling of regulators and
optional PHYs, we can absorb exynos as well.
Tested-by: Subbaraya Sundeep Bhatta <subbaraya.sundeep.bhatta@xilinx.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
In dwc3_cleanup_done_reqs() we expect that all
iterations of our while (1) loop will find a valid
struct dwc3_request *. In case we don't, we're
dumping a WARN_ON_ONCE() splat so that people report
the failure.
By moving our sanity checks our internal function
__dwc3_gadget_ep_queue() we can simplify the
externally visible API while also making sure that
callers of __dwc3_gadget_ep_queue() also make use of
the same checks.
Peter Zijlstra [Sun, 13 Dec 2015 21:11:16 +0000 (22:11 +0100)]
sched/wait: Fix the signal handling fix
Jan Stancek reported that I wrecked things for him by fixing things for
Vladimir :/
His report was due to an UNINTERRUPTIBLE wait getting -EINTR, which
should not be possible, however my previous patch made this possible by
unconditionally checking signal_pending().
We cannot use current->state as was done previously, because the
instruction after the store to that variable it can be changed. We must
instead pass the initial state along and use that.
Fixes: 68985633bccb ("sched/wait: Fix signal handling in bit wait helpers") Reported-by: Jan Stancek <jstancek@redhat.com> Reported-by: Chris Mason <clm@fb.com> Tested-by: Jan Stancek <jstancek@redhat.com> Tested-by: Vladimir Murzin <vladimir.murzin@arm.com> Tested-by: Chris Mason <clm@fb.com> Reviewed-by: Paul Turner <pjt@google.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: tglx@linutronix.de Cc: Oleg Nesterov <oleg@redhat.com> Cc: hpa@zytor.com Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Sun, 13 Dec 2015 20:41:10 +0000 (12:41 -0800)]
Merge branch 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull timer fixlets from Thomas Gleixner:
"Two trivial fixes which add missing header fileas and forward
declarations so the code will compile even when the magic include
chains are different"
* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip/gic-v3: Add missing include for barrier.h
irqchip/gic-v3: Add missing struct device_node declaration
Linus Torvalds [Sun, 13 Dec 2015 20:29:22 +0000 (12:29 -0800)]
Merge tag 'char-misc-4.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc
Pull fpga driver fixes from Greg KH:
"Only two small fpga driver fixes here, both have been in linux-next
for a while, and resolve some reported issues"
* tag 'char-misc-4.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
fpga manager: Fix firmware resource leak on error
fpga manager: remove label
Linus Torvalds [Sun, 13 Dec 2015 19:58:18 +0000 (11:58 -0800)]
Merge tag 'usb-4.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Pull USB driver fixes from Greg KH:
"Here are a number of small USB fixes for 4.4-rc5. All of them have
been in linux-next. The majority are gadget and phy issues, with a
few new quirks and device ids added as well"
* tag 'usb-4.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (32 commits)
USB: add quirk for devices with broken LPM
xhci: fix usb2 resume timing and races.
usb: musb: fail with error when no DMA controller set
usb: gadget: uvc: fix permissions of configfs attributes
usb: musb: core: Fix pm runtime for deferred probe
usb: phy: msm: fix a possible NULL dereference
USB: host: ohci-at91: fix a crash in ohci_hcd_at91_overcurrent_irq
usb: Quiet down false peer failure messages
usb: xhci: fix config fail of FS hub behind a HS hub with MTT
xhci: Fix memory leak in xhci_pme_acpi_rtd3_enable()
usb: Use the USB_SS_MULT() macro to decode burst multiplier for log message
USB: whci-hcd: add check for dma mapping error
usb: core : hub: Fix BOS 'NULL pointer' kernel panic
USB: quirks: Apply ALWAYS_POLL to all ELAN devices
usb-storage: Fix scsi-sd failure "Invalid field in cdb" for USB adapter JMicron
USB: quirks: Fix another ELAN touchscreen
usb: dwc3: gadget: don't prestart interrupt endpoints
USB: serial: Another Infineon flash loader USB ID
USB: cdc_acm: Ignore Infineon Flash Loader utility
USB: cp210x: Remove CP2110 ID from compatibility list
...
Linus Torvalds [Sun, 13 Dec 2015 00:43:44 +0000 (16:43 -0800)]
Merge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM SoC fixes from Arnd Bergmann:
"Here are a bunch of small bug fixes for various ARM platforms, nothing
really sticks out this week, most of either fixes bugs in code that
was just added in 4.4, or that has been broken for many years without
anyone noticing.
at91/sama5d2:
- fix sama5de hardware setup of sd/mmc interface
- proper selection of pinctrl drivers. PIO4 is necessary for sama5d2
imx:
- Fix vf610 SAI clock configuration bug which is discovered by the
newly added master mode support in SAI audio driver.
- Fix buggy L2 cache latency values in vf610 device trees, which may
cause system hang when cpu runs at a higher frequency.
ixp4xx:
- fix prototypes for readl/writel functions
ls2080a:
- use little-endian register access for GPIO and SDHCI
omap:
- Fix clock source for ARM TWD and global timers on am437x
- Always select REGULATOR_FIXED_VOLTAGE for omap2+ instead of when
MACH_OMAP3_PANDORA is selected
- Fix SPI DMA handles for dm816x as only some were mapped
- Fix up mbox cells for dm816x to make mailbox usable
Linus Torvalds [Sat, 12 Dec 2015 21:39:59 +0000 (13:39 -0800)]
Merge tag 'powerpc-4.4-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc fixes from Michael Ellerman:
- opal-irqchip: Fix double endian conversion from Alistair Popple
- cxl: Set endianess of kernel contexts from Frederic Barrat
- sbc8641: drop bogus PHY IRQ entries from DTS file from Paul Gortmaker
- Revert "powerpc/eeh: Don't unfreeze PHB PE after reset" from Andrew
Donnellan
* tag 'powerpc-4.4-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
Revert "powerpc/eeh: Don't unfreeze PHB PE after reset"
powerpc/sbc8641: drop bogus PHY IRQ entries from DTS file
cxl: Set endianess of kernel contexts
powerpc/opal-irqchip: Fix double endian conversion
Linus Torvalds [Sat, 12 Dec 2015 18:44:49 +0000 (10:44 -0800)]
Merge branch 'akpm' (patches from Andrew)
Merge misc fixes from Andrew Morton:
"17 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
MIPS: fix DMA contiguous allocation
sh64: fix __NR_fgetxattr
ocfs2: fix SGID not inherited issue
mm/oom_kill.c: avoid attempting to kill init sharing same memory
drivers/base/memory.c: prohibit offlining of memory blocks with missing sections
tmpfs: fix shmem_evict_inode() warnings on i_blocks
mm/hugetlb.c: fix resv map memory leak for placeholder entries
mm: hugetlb: call huge_pte_alloc() only if ptep is null
kernel: remove stop_machine() Kconfig dependency
mm: kmemleak: mark kmemleak_init prototype as __init
mm: fix kerneldoc on mem_cgroup_replace_page
osd fs: __r4w_get_page rely on PageUptodate for uptodate
MAINTAINERS: make Vladimir co-maintainer of the memory controller
mm, vmstat: allow WQ concurrency to discover memory reclaim doesn't make any progress
mm: fix swapped Movable and Reclaimable in /proc/pagetypeinfo
memcg: fix memory.high target
mm: hugetlb: fix hugepage memory leak caused by wrong reserve count
Linus Torvalds [Sat, 12 Dec 2015 18:34:20 +0000 (10:34 -0800)]
Merge branch 'parisc-4.4-3' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux
Pull parisc fixes from Helge Deller:
"Fix the boot crash on Mako machines with Huge Pages, prevent a panic
with SATA controllers (and others) by correctly calculating the IOMMU
space, hook up the mlock2 syscall and drop unneeded code in the parisc
pci code"
* 'parisc-4.4-3' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
parisc: Disable huge pages on Mako machines
parisc: Wire up mlock2 syscall
parisc: Remove unused pcibios_init_bus()
parisc iommu: fix panic due to trying to allocate too large region
Linus Torvalds [Sat, 12 Dec 2015 18:24:00 +0000 (10:24 -0800)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-block
Pull block layer fixes from Jens Axboe:
"A set of fixes for the current series. This contains:
- A bunch of fixes for lightnvm, should be the last round for this
series. From Matias and Wenwei.
- A writeback detach inode fix from Ilya, also marked for stable.
- A block (though it says SCSI) fix for an OOPS in SCSI runtime power
management.
- Module init error path fixes for null_blk from Minfei"
* 'for-linus' of git://git.kernel.dk/linux-block:
null_blk: Fix error path in module initialization
lightnvm: do not compile in debugging by default
lightnvm: prevent gennvm module unload on use
lightnvm: fix media mgr registration
lightnvm: replace req queue with nvmdev for lld
lightnvm: comments on constants
lightnvm: check mm before use
lightnvm: refactor spin_unlock in gennvm_get_blk
lightnvm: put blks when luns configure failed
lightnvm: use flags in rrpc_get_blk
block: detach bdev inode from its wb in __blkdev_put()
SCSI: Fix NULL pointer dereference in runtime PM
Linus Torvalds [Sat, 12 Dec 2015 18:16:26 +0000 (10:16 -0800)]
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fixes from Catalin Marinas:
- Update the linker script to use L1_CACHE_BYTES instead of hard-coded
64. We recently changed L1_CACHE_BYTES to 128
- Improve race condition reporting on set_pte_at() and change the BUG
to WARN_ONCE. With hardware update of the accessed/dirty state, we
need to ensure that set_pte_at() does not inadvertently override
hardware updated state. The patch also makes the checks ignore
!pte_valid() new entries
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: Improve error reporting on set_pte_at() checks
arm64: update linker script to increased L1_CACHE_BYTES value
Qais Yousef [Fri, 11 Dec 2015 21:41:09 +0000 (13:41 -0800)]
MIPS: fix DMA contiguous allocation
Recent changes to how GFP_ATOMIC is defined seems to have broken the
condition to use mips_alloc_from_contiguous() in
mips_dma_alloc_coherent().
I couldn't bottom out the exact change but I think it's this commit d0164adc89f6 ("mm, page_alloc: distinguish between being unable to
sleep, unwilling to sleep and avoiding waking kswapd").
GFP_ATOMIC has multiple bits set and the check for !(gfp & GFP_ATOMIC)
isn't enough.
The reason behind this condition is to check whether we can potentially
do a sleeping memory allocation. Use gfpflags_allow_blocking() instead
which should be more robust.
Dmitry V. Levin [Fri, 11 Dec 2015 21:41:06 +0000 (13:41 -0800)]
sh64: fix __NR_fgetxattr
According to arch/sh/kernel/syscalls_64.S and common sense, __NR_fgetxattr
has to be defined to 259, but it doesn't. Instead, it's defined to 269,
which is of course used by another syscall, __NR_sched_setaffinity in this
case.
This bug was found by strace test suite.
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org> Acked-by: Geert Uytterhoeven <geert+renesas@glider.be> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Junxiao Bi [Fri, 11 Dec 2015 21:41:03 +0000 (13:41 -0800)]
ocfs2: fix SGID not inherited issue
Commit 8f1eb48758aa ("ocfs2: fix umask ignored issue") introduced an
issue, SGID of sub dir was not inherited from its parents dir. It is
because SGID is set into "inode->i_mode" in ocfs2_get_init_inode(), but
is overwritten by "mode" which don't have SGID set later.
Fixes: 8f1eb48758aa ("ocfs2: fix umask ignored issue") Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com> Cc: Mark Fasheh <mfasheh@suse.de> Cc: Joel Becker <jlbec@evilplan.org> Acked-by: Srinivas Eeda <srinivas.eeda@oracle.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Chen Jie [Fri, 11 Dec 2015 21:41:00 +0000 (13:41 -0800)]
mm/oom_kill.c: avoid attempting to kill init sharing same memory
It's possible that an oom killed victim shares an ->mm with the init
process and thus oom_kill_process() would end up trying to kill init as
well.
This has been shown in practice:
Out of memory: Kill process 9134 (init) score 3 or sacrifice child
Killed process 9134 (init) total-vm:1868kB, anon-rss:84kB, file-rss:572kB
Kill process 1 (init) sharing same memory
...
Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000009
And this will result in a kernel panic.
If a process is forked by init and selected for oom kill while still
sharing init_mm, then it's likely this system is in a recoverable state.
However, it's better not to try to kill init and allow the machine to
panic due to unkillable processes.
[rientjes@google.com: rewrote changelog]
[akpm@linux-foundation.org: fix inverted test, per Ben] Signed-off-by: Chen Jie <chenjie6@huawei.com> Signed-off-by: David Rientjes <rientjes@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Hillf Danton <hillf.zj@alibaba-inc.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Li Zefan <lizefan@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Seth Jennings [Fri, 11 Dec 2015 21:40:57 +0000 (13:40 -0800)]
drivers/base/memory.c: prohibit offlining of memory blocks with missing sections
Commit bdee237c0343 ("x86: mm: Use 2GB memory block size on large-memory
x86-64 systems") and 982792c782ef ("x86, mm: probe memory block size for
generic x86 64bit") introduced large block sizes for x86. This made it
possible to have multiple sections per memory block where previously,
there was a only every one section per block.
Since blocks consist of contiguous ranges of section, there can be holes
in the blocks where sections are not present. If one attempts to
offline such a block, a crash occurs since the code is not designed to
deal with this.
This patch is a quick fix to gaurd against the crash by not allowing
blocks with non-present sections to be offlined.
Hugh Dickins [Fri, 11 Dec 2015 21:40:55 +0000 (13:40 -0800)]
tmpfs: fix shmem_evict_inode() warnings on i_blocks
Dmitry Vyukov provides a little program, autogenerated by syzkaller,
which races a fault on a mapping of a sparse memfd object, against
truncation of that object below the fault address: run repeatedly for a
few minutes, it reliably generates shmem_evict_inode()'s
WARN_ON(inode->i_blocks).
(But there's nothing specific to memfd here, nor to the fstat which it
happened to use to generate the fault: though that looked suspicious,
since a shmem_recalc_inode() had been added there recently. The same
problem can be reproduced with open+unlink in place of memfd_create, and
with fstatfs in place of fstat.)
v3.7 commit 0f3c42f522dc ("tmpfs: change final i_blocks BUG to WARNING")
explains one cause of such a warning (a race with shmem_writepage to
swap), and possible solutions; but we never took it further, and this
syzkaller incident turns out to have a different cause.
shmem_getpage_gfp()'s error recovery, when a freshly allocated page is
then found to be beyond eof, looks plausible - decrementing the alloced
count that was just before incremented - but in fact can go wrong, if a
racing thread (the truncator, for example) gets its shmem_recalc_inode()
in just after our delete_from_page_cache(). delete_from_page_cache()
decrements nrpages, that shmem_recalc_inode() will balance the books by
decrementing alloced itself, then our decrement of alloced take it one
too low: leading to the WARNING when the object is finally evicted.
Once the new page has been exposed in the page cache,
shmem_getpage_gfp() must leave it to shmem_recalc_inode() itself to get
the accounting right in all cases (and not fall through from "trunc:" to
"decused:"). Adjust that error recovery block; and the reinitialization
of info and sbinfo can be removed too.
While we're here, fix shmem_writepage() to avoid the original issue: it
will be safe against a racing shmem_recalc_inode(), if it merely
increments swapped before the shmem_delete_from_page_cache() which
decrements nrpages (but it must then do its own shmem_recalc_inode()
before that, while still in balance, instead of after). (Aside: why do
we shmem_recalc_inode() here in the swap path? Because its raison d'etre
is to cope with clean sparse shmem pages being reclaimed behind our
back: so here when swapping is a good place to look for that case.) But
I've not now managed to reproduce this bug, even without the patch.
I don't see why I didn't do that earlier: perhaps inhibited by the
preference to eliminate shmem_recalc_inode() altogether. Driven by this
incident, I do now have a patch to do so at last; but still want to sit
on it for a bit, there's a couple of questions yet to be resolved.
projects / karo-tx-linux.git / log