gadget methods should be called without
spinlocks held.
Reported-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of tcm, ep->driver_data was used only for endpoint claiming so
we can simplify code by reducing it. We also remove give_back_ep()
function which is not needed after all - when error code is returned
from bind() function, composite will release all endpoints anyway.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of dbgp, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of u_serial ep->driver_data stores pointer to struct gs_port,
which is referenced in many places in code. Code using ep->driver_data
to mark endpoint as enabled/disabled has been removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of u_ether we only need to store in ep->driver_data pointer to
struct eth_dev, as it's used in rx_complete() and tx_complete() callbacks.
All other uses of ep->driver_data are now meaningless and can be safely
removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_uvc, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_uac2, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_uac1, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_subset, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_sourcesink we only need to store in ep->driver_data pointer
to struct f_sourcesink, as it's used in source_sink_complete() callback.
All other uses of ep->driver_data are now meaningless and can be safely
removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_serial, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_rndis, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_printer we only need to store in ep->driver_data pointer to
struct printer_dev, as it's used in rx_complete() and tx_complete()
callbacks. All other uses of ep->driver_data are now meaningless and can
be safely removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_phonet we only need to store in ep->driver_data pointer to
struct f_phonet, as it's used in pn_tx_complete() and pn_rx_complete()
callbacks. All other uses of ep->driver_data are now meaningless and can
be safely removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_obex, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_ncm, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_midi we only need to store in ep->driver_data pointer to
struct f_midi, as it's used in f_midi_complete() callback and related
functions. All other uses of ep->driver_data are now meaningless and
can be safely removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_mass_storage we only need to store in ep->driver_data
pointer to struct fsg_common, which is used in bulk_in_complete() and
bulk_out_complete() callbacks. All other uses of ep->driver_data are now
meaningless and can be safely removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_hid we only need to store in ep->driver_data pointer to
struct f_loopback, as it's used in loopback_complete() callback. All
other uses of ep->driver_data are now meaningless and can be safely
removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_hid we only need to store in ep->driver_data pointer to
struct f_hidg, as it's used in f_hidg_req_complete() callback. All
other uses of ep->driver_data are now meaningless and can be safely
removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_ecm, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_acm we only need to store in ep->driver_data pointer to
struct f_acm, as it's used in acm_complete_set_line_coding() callback.
All other uses of ep->driver_data are now meaningless and can be safely
removed.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.
In case of f_ecm, ep->driver_data was used only for endpoint claiming
and marking endpoints as enabled, so we can simplify code by reducing
it.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This patch introduces usb_ep_autoconfig_release() function which allows
to release endpoint previously obtained from usb_ep_autoconfig() during
USB function bind.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
The 'driver_data' field in ep0 is never set to pointer to cdev, so we
have to obtain it from another source as in this context ep->driver_data
contains invalid data.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Fix comments in code to make them up to date.
composite: claiming endpoint is now done by setting ep->claimed flag,
not ep->driver_data.
epautoconf: usb_ep_autoconfig() and usb_ep_autoconfig_ss() return
claimed endpoint with ep->claimed flag already set.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
The millisecond of the last second will be normal if tv_sec is
overflowed. But for y2038 consistency and demonstration purpose,
and avoiding further risks, we need to remove 'timeval' in this
driver, to avoid similair problems.
Signed-off-by: Pingbo Wen <pingbo.wen@linaro.org>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Since the host and gadget can't agree with transfer length before
each transfer, but they agree with max packet size for each
endpoint, we use max packet size to format data pattern.
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
USB network adapters support Jumbo frames. The only thing blocking
that feature is the code in the gadget driver that disposes of
packets larger than 1518 bytes, and the limit on the ioctl to set
the mtu.
This patch relaxes these limits, and allows up to 15k frames sizes.
The 15k value was chosen because 16k does not work on all platforms,
and usingclose to 16k will result in allocating 5 or 8 4k pages to
store the skb, wasting pages at no measurable performance gain.
On a topic-miami board (Zynq-7000), iperf3 performance reports:
MTU= 1500, PC-to-gadget: 139 Mbps, Gadget-to-PC: 116 Mbps
MTU=15000, PC-to-gadget: 239 Mbps, Gadget-to-PC: 361 Mbps
On boards with slower CPUs the performance improvement will be
relatively much larger, e.g. an OMAP-L138 increased from 40 to
220 Mbps using a similar patch on an 2.6.37 kernel.
Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
Signed-off-by: Felipe Balbi <balbi@ti.com>
struct at91_udc_data is now only used inside the driver, move it to its
include.
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
f_midi is not checking whether there is an error on usb_ep_queue
request, ignoring potential problems, such as memory leaks.
Signed-off-by: Felipe F. Tonello <eu@felipetonello.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Instead of allowing a range of 2 to 4 requests,
let's allow the user choose up to 32 requests
as that will give us a better chance of keeping
controller busy.
We still maintain default of 2 so users shouldn't
be affected.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Remove unneeded NULL test.
The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@@ expression x; @@
-if (x != NULL)
\(kmem_cache_destroy\|mempool_destroy\|dma_pool_destroy\)(x);
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Felipe Balbi <balbi@ti.com>
dummy_timer uses transfer() to update transfer limit. However,
limit passed to dummy_timer changes depending on transfer type,
so the actual limit is overwritten.
This can cause unpredictably slow / fast bulk transfers when
coupled with control / interrupt transfers.
Fix by returning actual amount of data sent in transfer() and
substracting from total.
Signed-off-by: Igor Kotrasinski <i.kotrasinsk@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
transfer() schedules a rescan for transfers larger than
maxpacket, which is wrong for transfers that are multiples
of maxpacket.
Rewrite to fix and clarify packet multiple / remainder
transfer logic.
Signed-off-by: Igor Kotrasinski <i.kotrasinsk@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
We already know at this point that to_host is false.
Signed-off-by: Igor Kotrasinski <i.kotrasinsk@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
currently, when a zlp flag is set and an urb/usb_request
buffer is filled without a short packet, transfer() leaves
its status at -EINPROGRESS and does not rescan for short
packet.
In a scenario where ep.maxpacket bytes are copied,
URB_ZERO_PACKET is set, urb buffer is filled and usb_request
buffer is not, transfer() returns with an urb with
-EINPROGRESS status, which dummy_hcd treats as incomplete
transfer.
Check for zlp and rescan appropriately.
Signed-off-by: Igor Kotrasinski <i.kotrasinsk@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
The recently added endpoint capabilities flags verification breaks Atmel
USBA because the endpoint configuration was only added when the driver
is bound using the legacy pdata interface.
Convert endpoint configuration to new capabilities model when driver is
bound to a device tree as well.
Signed-off-by: Sylvain Rochet <sylvain.rochet@finsecur.com>
Fixes: 47bef38651 ("usb: gadget: atmel_usba_udc: add ep capabilities support")
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
If a failure happens early in udc_pci_probe(), error handling code
just kfree(dev) and returns. The patch adds proper resource
deallocations in udc_pci_probe() itself,
since udc_pci_remove() is not suitabe to be called so early
in initialization process.
By the way, iounmap(dev->regs) is replaced by iounmap(dev->virt_addr)
in udc_pci_remove() for clarity.
Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This patch fixes possible regression introduced by
patch reworking endpoint claiming mechanism. It restores
setring ep->driver_data to NULL in usb_ep_autoconfig_reset(),
which was removed by patch commit cc476b42a3.
Fixes: cc476b42a3 ("usb: gadget: encapsulate endpoint
claiming mechanism")
Reported-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Definitions from linux/platform_data/atmel.h are not used, remove the
include.
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
New support for Allwinne SoC on the MUSB driver has been added to the list of
glue layers. MUSB also got support for building all DMA engines in one binary;
this will be great for distros.
DWC3 now has no trace of dev_dbg()/dev_vdbg() usage. We will rely solely on
tracing to debug DWC3. There was also a fix for memory corruption with EP0 when
maxpacket size transfers are > 512 bytes.
Robert's EP capabilities flags is making EP selection a lot simpler. UDCs are
now required to set these flags up when adding endpoints to the framework.
Other than these, we have the usual set of miscelaneous cleanups and minor
fixes.
Signed-off-by: Felipe Balbi <balbi@ti.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=WuIB
-----END PGP SIGNATURE-----
Merge tag 'usb-for-v4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb into usb-next
Felipe writes:
usb: patches for v4.3 merge window
New support for Allwinne SoC on the MUSB driver has been added to the list of
glue layers. MUSB also got support for building all DMA engines in one binary;
this will be great for distros.
DWC3 now has no trace of dev_dbg()/dev_vdbg() usage. We will rely solely on
tracing to debug DWC3. There was also a fix for memory corruption with EP0 when
maxpacket size transfers are > 512 bytes.
Robert's EP capabilities flags is making EP selection a lot simpler. UDCs are
now required to set these flags up when adding endpoints to the framework.
Other than these, we have the usual set of miscelaneous cleanups and minor
fixes.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Looks like the word "contiguous" is often mistyped.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
<linux/uaccess.h> was originally being pulled
indirectly through some other header, however
it's not anymore, so we need to include it
directly
Reported-by: Jim Davis <jim.epost@gmail.com>
Suggested-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Felipe Balbi <balbi@ti.com>
This header file contains helpers for quirks based on UDC controller name.
Since we have generic quirk bitfields in usb_gadget structure for all of
these quirks we don't need to have this header any longer.
This patch removes gadget_chips.h file and makes sure that it's no longer
included anywhere in kernel sources.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Add 'match_ep' callback to utilize chip-specific knowledge in endpoint matching
process. Function does the same that was done by chip-specific code inside
of epautoconf. Now this code can be removed from there to separate generic code
from platform specific logic.
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Add 'match_ep' callback to utilize chip-specific knowledge in endpoint matching
process. Function does the same that was done by chip-specific code inside
of epautoconf. Now this code can be removed from there to separate generic code
from platform specific logic.
[ balbi@ti.com : fix build breakage ]
Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
