vmxnet3 allocates a page per skb fragment. We must account
PAGE_SIZE increments on skb->truesize, not the actual frag length.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Shreyas Bhatewara <sbhatewara@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ftgmac100 allocates a page per skb fragment. We must account
PAGE_SIZE increments on skb->truesize, not the actual frag length.
If frame is under 64 bytes, page is freed, and truesize adjusted.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Po-Yu Chuang <ratbert@faraday-tech.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The w83627ehf driver is improperly reporting thermal diode sensors as
type 2, instead of 3. This caused "sensors" and possibly other
monitoring tools to report these sensors as "transistor" instead of
"thermal diode".
Furthermore, diode subtype selection (CPU vs. external) is only
supported by the original W83627EHF/EHG. All later models only support
CPU diode type, and some (NCT6776F) don't even have the register in
question so we should avoid reading from it.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: stable@kernel.org
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>
sky2 allocates a page per skb fragment. We must account
PAGE_SIZE increments on skb->truesize, not the actual frag length.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
e1000e allocates a page per skb fragment. We must account
PAGE_SIZE increments on skb->truesize, not the actual frag length.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ixgbe allocates half a page per skb fragment. We must account
PAGE_SIZE/2 increments on skb->truesize, not the actual frag length.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
e1000 allocates half a page per skb fragment. We must account
PAGE_SIZE/2 increments on skb->truesize, not the actual frag length.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
e1000 allocates a full page per skb fragment. We must account PAGE_SIZE
increments on skb->truesize, not the actual frag length.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
bnx2 allocates a full page per fragment. We must account PAGE_SIZE
increments on skb->truesize, not the actual frag length.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix skb truesize underestimations of this driver.
Each frag truesize is exactly rx_frag_size bytes. (2048 bytes per
default)
A driver should not use "sizeof(struct sk_buff)" at all.
Signed-off-by: Eric Dumazet <eric.dumazet>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
skb truesize currently accounts for sk_buff struct and part of skb head.
kmalloc() roundings are also ignored.
Considering that skb_shared_info is larger than sk_buff, its time to
take it into account for better memory accounting.
This patch introduces SKB_TRUESIZE(X) macro to centralize various
assumptions into a single place.
At skb alloc phase, we put skb_shared_info struct at the exact end of
skb head, to allow a better use of memory (lowering number of
reallocations), since kmalloc() gives us power-of-two memory blocks.
Unless SLUB/SLUB debug is active, both skb->head and skb_shared_info are
aligned to cache lines, as before.
Note: This patch might trigger performance regressions because of
misconfigured protocol stacks, hitting per socket or global memory
limits that were previously not reached. But its a necessary step for a
more accurate memory accounting.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Andi Kleen <ak@linux.intel.com>
CC: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
gpio_base was set to 0 if no system platform data or open firmware
platform data was provided. This led to conflicts, if any other gpiochip
with a gpiobase of 0 was instantiated already. Setting it to -1 will
automatically use the first one available.
Signed-off-by: Hartmut Knaack <knaack.h@gmx.de>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
With commit f64ad1a0e2, "gpio/omap: cleanup _set_gpio_wakeup(), remove
ifdefs", access to build time conditionally omitted 'suspend_wakeup'
member of the 'gpio_bank' structure has been placed unconditionally in
function _set_gpio_wakeup(), which is always built. This resulted in the
driver compilation broken for certain OMAP1, i.e., non-OMAP16xx,
configurations.
Really required or not in previously excluded cases, define this
structure member unconditionally as a fix.
Tested with a custom OMAP1510 only configuration.
Signed-off-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Acked-by: Kevin Hilman <khilman@ti.com>
Tested-by: Aaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
This driver provides two functions in one configuration:
a mass storage, and a ACM (serial port) link.
Heavily based on multi.c and cdc2.c
Signed-off-by: Klaus Schwarzkopf <schwarzkopf@sensortherm.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
| drivers/usb/gadget/mv_udc_core.c: In function 'handle_setup_packet':
| drivers/usb/gadget/mv_udc_core.c:1556:6: warning: 'status' may be \
used uninitialized in this function
Signed-off-by: Felipe Balbi <balbi@ti.com>
This patch is going to support clock gating when vbus detection is
posible. Clock and phy will be on only when usb gadget is used(vbus valid).
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
use DMA API for status_req's dma address, it is needed by dtd.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
The controller will prime failure sometimes when do the iperf test.
Add delay to wait controller release dtd dma before we free it.
Then the issue is gone.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This patch is going to correct the ep0 state, and the unexpected
ep0 package warning can be removed.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
1: Add parameter check.
2: For controller endpoint, we need to flush in and out directions.
3: delete redundant code, make it more readable.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
According to the comment right above the code, we should use
USB_ENDPOINT_XFER_BULK instead.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
The ep enable / disable functions can be called from interrupt
context, and they are not race safe on SMP systems. The critical
data can be modified in more than one routing.
Make them race safe by using IRQ-safe spinlock functions.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
For the code doesn't restrict controller ep must be ep0, so we will go
through all the eps and check if there is a setup package received.
And also we just need to acknowledge the corresponding bit in
ENDPTSETUPSTAT register.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Set next dtd ptr to EP_QUEUE_HEAD_NEXT_TERMINATE for dqh when init ep0.
It means the dQH is empty.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
According to ChipIdea datasheet, there is no toggle flag for ep0.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
The max size of data payload is in bit0 - bit10, so we need use 0x7ff
as the bitmask to fetch from usb_endpoint_descriptor.wMaxPacketSize.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Some platforms will use usb to download images, the controller may not
be stopped correctly when start kernel. In some cases, it may have some
pending interrupts, and they will be triggered immediately when we finish
requesting irq in function probe. But we haven't finished the device
initialization at this time. So let's stop udc here to avoid this case
occurred.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Tag the probe function as __devinit.
Tag the remove function as __devexit.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This patch do the following things:
1. Add header and Copyright for marvell usb driver.
2. Add mv_usb.h in include/linux/platform_data, make the driver
fits all the marvell platform using the same ChipIdea usb ip.
3. Some SOC may has mutiple clock sources, so let me define it
in mv_usb_platform_data and give two helper functions named
udc_clock_enable/udc_clock_disable to deal with the clocks.
4. Different SOCs will have some difference in PHY initialization,
so we will remove file mv_udc_phy.c and add two funtions in
mv_usb_platform_data, let the platform relative driver to realize it.
5. Rewrite probe function according to the modification list above. Find
it will kernel panic when probe failed. The root cause is as follows:
When probe failed, the error handle may call device_unregister()
which in return will call gadget_release.In current code,
gadget_release have two issues:
1: the_controller is a NULL pointer.
2: if we free udc here, then the following code in probe
will access NULL pointer.
Signed-off-by: Neil Zhang <zhangwm@marvell.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
| drivers/usb/renesas_usbhs/fifo.c: In function ‘usbhsf_dma_prepare_push’:
| drivers/usb/renesas_usbhs/fifo.c:823:7: warning: cast from pointer \
to integer of different size [-Wpointer-to-int-cast]
| drivers/usb/renesas_usbhs/fifo.c: In function ‘usbhsf_dma_try_pop’:
| drivers/usb/renesas_usbhs/fifo.c:900:7: warning: cast from pointer \
to integer of different size [-Wpointer-to-int-cast]
Signed-off-by: Felipe Balbi <balbi@ti.com>
In current renesas-usbhs,
there was inconsistency about the style of kernel module or built-in.
This patch solve it.
[ balbi@ti.com : fix compile issue when building modules ]
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
renesas_usbhs has default callback functions and settings.
And it tried overwrite to platform private data
if platform doesn't have them.
So, if renesas_usbhs was compiled as module,
it will be hung-up on 2nd insmod.
This patch fixup it.
Special thanks to Bastian
Reported-by: Bastian Hecht <hechtb@googlemail.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This is mod_host prototype support for renesas_usbhs driver.
It doesn't support USB-Hub, and USB-DMAC for now.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
CCPL setting is needed on only mod_gadget.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
It is enabled to set SOF packet output bit when USBRST bit was set.
And USBRST bit should be set 0 when SOF packet was output.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
some renesas_usbhs device is supporting OTG external device interface.
In that device, it is necessary to control PWEN/EXTLP on DVSTCTR.
This patch support it.
But renesas_usbhs driver doesn't have OTG support for now.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
When data read interrupt happened, the pipe is BUF which means "enable".
then, next un-necessary interrupt/token might be
issued again when all data were popped from fifo.
It will cause un-understandable bug.
This patch decides pipe disable on top of read interrupt.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
renesas_usbhs can manually set DATA0/DATA1.
This patch is prepare for mod_host support
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Automatically packet start by usbhs_pkt_push() was useful.
But the pushed packet will be called twice
if new packet was pushed on usbhs_pkt :: done callback.
(1st is called by usbhs_pkt_push(), 2nd is called by usbhsf_pkt_handler())
This patch disables automatic packet start,
and clarified packet start timing.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
transfer done function was registered in struct struct usbhs_pipe_info.
It was good for mod_gadget, but not good for mod_host.
This function move it to struct usbhs_pkt.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
pipe name is usefull function for mod_xxx
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
device select method will be used on mod_host
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
There was no method to get struct usbhs_priv when
packet transfer done function was called.
This patch allow that callback function receive it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
