Age | Commit message (Collapse) | Author |
|
Fix "Octal permissions are preffered than symbolic ones" issues.
Signed-off-by: Gioh Kim <gi-oh.kim@profitbricks.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
This patch fixes only obvious lines.
There are still more issues.
Signed-off-by: Gioh Kim <gi-oh.kim@profitbricks.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
While the timesync protocol was a great idea, it never ended up getting
implemented by any known hardware devices. It's also a bit
"interesting" in how it ties into the platform controller.
So, just remove it for now. It's not needed, no one uses it, and it's a
stumbling block in getting the greybus core code merged out of the
staging tree. If anyone wants it in the future, reverting this patch is
a great place to start from.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Johan Hovold <johan@kernel.org>
Reviewed-by: Alex Elder <elder@kernel.org>
Acked-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull vfs updates from Al Viro:
- more ->d_init() stuff (work.dcache)
- pathname resolution cleanups (work.namei)
- a few missing iov_iter primitives - copy_from_iter_full() and
friends. Either copy the full requested amount, advance the iterator
and return true, or fail, return false and do _not_ advance the
iterator. Quite a few open-coded callers converted (and became more
readable and harder to fuck up that way) (work.iov_iter)
- several assorted patches, the big one being logfs removal
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
logfs: remove from tree
vfs: fix put_compat_statfs64() does not handle errors
namei: fold should_follow_link() with the step into not-followed link
namei: pass both WALK_GET and WALK_MORE to should_follow_link()
namei: invert WALK_PUT logics
namei: shift interpretation of LOOKUP_FOLLOW inside should_follow_link()
namei: saner calling conventions for mountpoint_last()
namei.c: get rid of user_path_parent()
switch getfrag callbacks to ..._full() primitives
make skb_add_data,{_nocache}() and skb_copy_to_page_nocache() advance only on success
[iov_iter] new primitives - copy_from_iter_full() and friends
don't open-code file_inode()
ceph: switch to use of ->d_init()
ceph: unify dentry_operations instances
lustre: switch to use of ->d_init()
|
|
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
This resolves a merge issue with
drivers/staging/iio/accel/sca3000_core.c and we want the fixes all in
here as well.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Fix to return a negative error code from the es2_arpc_in_enable() error
handling case instead of 0, as done elsewhere in this function.
Fixes: 9d9d3777a9db ("greybus: es2: Add a new bulk in endpoint for APBridgeA RPC")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Fix size field of arpc message request by using the header size and not
the pointer size.
Signed-off-by: Rui Miguel Silva <rmfrfs@gmail.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
This patch replaces kmalloc and memcpy with kmemdup for duplication of
memory.
Signed-off-by: sayli karnik <karniksayli1995@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The last thing remaining in kernel_ver.h was the setting of
CONFIG_PM_RUNTIME, which isn't needed in a in-tree implementation. So
remove the setting of this value, and the .h file entirely as that was
the last thing left in it.
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Properly report which endpoints are being ignored and which ones are
"unknown" to the driver.
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
usb_kill_urb() and usb_free_urb() can be called with NULL pointers, so
no need to check before calling them.
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
We don't need the defines for the number of bulk in or out endpoints
anymore, as the driver just grabs the first ones it finds and runs with
it.
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
We only care about one bulk IN endpoint for cports, and one for ARPC, so
drop the array of bulk IN endpoints to simplify things.
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
We only care about one bulk out endpoint, the first one, so remove
the pretense of keeping an array of these things. Just grab the first
one in the list and run away!
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
We were not really using this structure at all, it was only returning
'0' when asked what cport matched to what pair, so remove it all.
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
These functions were never used, and we are about to remove the
structures it was trying to reference, so let's remove it to get it out
of the way.
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
This matches a corresponding firmware change to declare a single
BULK EP pair for all greybus traffic and 1 BULK IN ep for ARPC.
Thus, avoiding URB submittions for all the other unused BULK IN
endpoints on the HOST side that took considerable amount
of bus time.
Testing Done:
Tested with modified AP<->APB1 loopback test and also with GPBridge
modules to ensure there are no regressions
Signed-off-by: Sandeep Patil <sspatil@google.com>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Remove the now obsolete and redundant cport-reset ARPC, along with the
consequently unused cport_disable callback.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Remove the now obsolete ping and cport_features_enable/disable
callbacks.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Implement the shutdown callback which is used to execute shutdown
operations on offloaded connections.
This adds a new shutdown ARPC.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Implement the flush callback which is called as part of connection tear
down to flush host-device queues and stop any ongoing transfers.
Note that this should be considered an optimisation of sort since if
the CPort is stuck waiting for credit, the CPort is likely still stuck
when we try to send the cport_shutdown request over it after the
callback returns.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
In case a bulk-in transfer-buffer allocation failed during probe, we'd
currently leak the corresponding URB.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Add a host-device cport_clear callback, which will be called as part
of the new connection tear-down sequence to reset the CPort state.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Acked-by: Sandeep Patil <sspatil@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Add a host-device cport_quiesce callback, which will be called as part
of the new connection tear-down sequence to disable flow control after
first making sure that enough peer buffer space is available for the
next messages about to be sent.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Acked-by: Sandeep Patil <sspatil@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Add a host-device cport_connected callback, which will be called after a
connection has been created and that can be used by the host-device
driver to make sure its internal state is updated to match the CPort
attributes set by the SVC.
This callback will eventually replace the cport_features_enable
callback.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Acked-by: Sandeep Patil <sspatil@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Move the ARPC definitions to their own header.
ARPC is not part of greybus, but is rather an implementation-specific
means of communicating with a certain class of host-device hardware.
Note that the same is true for the APBA USB vendor requests, but we keep
them in the greybus header for the time being.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Make sure to use the common GB_APB prefix for all APBA USB vendor
requests.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Give the USB vendor-request timeout define a more descriptive name.
Also drop the since-long obsolete comments about allowing "the SVC to do
something" and "SVC messages go down our control pipe".
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Add dedicated define for ARPC CPort requests instead of using the
default timeout for USB vendor requests.
We still allow responses to take 500 ms to arrive, but note that this
adds on top of the 500ms already allowed for a requests to be
acknowledged. This should probably be tightened up at some point.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Allow user of ARPC to pass a NULL pointer for the ARPC result.
For simple requests there may only be one error code for remote errors,
or the caller may simply not care to differentiate them.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Make sure to always set the result value for ARPC instead of forcing
every caller to do it in order to avoid compiler warnings.
The ARPC result should still be ignored unless arpc_sync returns
-EREMOTEIO.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Add a _req suffix to request message structures, and a _TYPE_ infix to
request type defines.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Fix ARPC active-list corruption due to incorrect list-interface usage.
The corruption manifested itself as ARPC timeouts due to expected
responses not being recognised whenever more than one ARPC was active.
This could be seen for example when two interfaces were being runtime
suspended in parallel:
[ 165.739896] usb 1-1.1: invalid arpc response id received: 13
[ 165.794743] greybus 1-5.5: gb_interface_refclk_set - 0
[ 166.241202] usb 1-1.1: failed to execute ARPC: -110
Fortunately the impact of this bug has so far been limited to such
timeouts and error messages due to ARPC currently only being used for
CPort reset in the connection tear-down path.
Reported-by: Mark Greer <mgreer@animalcreek.com>
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Add missing le16_to_cpu() to an ARPC error message in the receive path,
and also use %u to print the unsigned id.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
The wrong pointer was checked for allocation failures when allocating
the ARPC response buffer, something which would lead to allocation
failures going undetected.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
There is a dangling u32 in es2/greybus_protocols.h that is only obvious
when you try to compile up gbsim. We need to do a cpu_to_le32 and declare
the 32-bit int in the header as __le32. This patch does that and splits out
the assignment of the req->flags parameter to a temporary variable to allow
for later printing.
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Suggested-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
The compiler has no way of knowing whether a called function will
actually assign something to the object whose address is passed as
an argument. So it must assume it won't happen, and this leads to a
compile warning. Fix this.
Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Fix arpc response lookups that were truncating the 16-bit response id to
8-bit, something which would have lead to all arpc calls timing out
after the 256th request.
Testing done: Enumerated and ejected a module on EVT2.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Alex Elder <elder@linaro.org>
|
|
This is what I got:
greybus/es2.c:1130:29: warning: restricted __le16 degrades to integer
greybus/es2.c:1140:22: warning: incorrect type in assignment (different base types)
greybus/es2.c:1140:22: expected restricted __le16 [usertype] id
greybus/es2.c:1140:22: got unsigned short [unsigned] [usertype] <noident>
greybus/es2.c:1162:52: warning: incorrect type in argument 8 (different base types)
greybus/es2.c:1162:52: expected unsigned short [unsigned] [usertype] size
greybus/es2.c:1162:52: got restricted __le16 [usertype] size
greybus/es2.c:1164:31: warning: restricted __le16 degrades to integer
greybus/es2.c:1253:34: warning: incorrect type in argument 2 (different base types)
greybus/es2.c:1253:34: expected unsigned char [unsigned] [usertype] id
greybus/es2.c:1253:34: got restricted __le16 [usertype] id
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Alex Elder <elder@linaro.org>
|
|
ARPC should return 0 on success, but instead was returning the number of
jiffies left until the timeout.
This caused cport_reset() to report an error and an incorrect error
message to be printed when disabling a connection.
Reported-by: Alex Elder <elder@linaro.org>
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Tested-by: Alex Elder <elder@linaro.org>
Signed-off-by: Alex Elder <elder@linaro.org>
|
|
Implement ARPC. In first time, we are going to use it to
implement new request but the goal is to update all existing vendor
request to use ARPC.
In addition, Convert the current USB Vendor request for CPort Reset
to ARPC so that we can be sure that the port has been fully reset
by the time the request completes.
Testing Done:
AP can reset APBA Cports by using the ARPC command.
In addition, tested with a hacked firmware that cause error during
the Cport reset, and Greybus printed the error "failed to reset cport".
Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Alex Elder <elder@linaro.org>
|
|
Add a new bulk in endpoint in order to get RPC status
from APbridgeA without to have to poll on control endpoint.
So the new endpoint layout is:
EP0: control EP
EP1 and EP2: muxed endpoints (bulk in and out)
EP3 to EP14: direct muxed endpoints (bulk in and out)
EP15: ARPC bulk in endpoint
Note:
This patch is allocating ARPC URBs but does nothing with them.
The following patch will use them.
Testing Done:
Tested with an APBridgeA enumerating 16 endpoints.
Kernel doesn't print Not enough endpoints found in device, aborting!
Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
Reviewed-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Alex Elder <elder@linaro.org>
|
|
This reverts commit b44c3b5b0307788750eb4c462ed5982236876a8b.
Someone decided that all use of spin_lock_irq was to be considered a bug
and went on a search-and-replace type "bug-fixing" spree last week.
This is however just plain wrong. Using spin_lock_irq is perfectly fine
in paths were interrupts have not been disabled, and this is in fact
even preferred over the lazy approach of always using spin_lock_irqsave
instead of understanding the code that is being written or modified.
All current uses of spin_lock_irq have already been vetted in this
respect. Also note that it is only used in functions that may sleep,
that is, in functions that must not be called with interrupts disabled
in the first place.
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
Add support for Greybus CPort flags that are sent to the bridge through
a new USB vendor request when enabling a CPort as part of connection
establishment.
Currently the Greybus control and high-priority connection flags are
recognised and forwarded to APBA.
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Johan Hovold <johan@hovoldconsulting.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
spin_[un]lock_irq() routines should be used carefully as they things can
go wrong, if they are mixed with spin_lock_irqsave() or other variants.
The main problem is that spin_[un]lock_irq() routines doesn't check if
the IRQs are already disabled/enabled on the local CPU and so
spin_unlock_irq() will forcefully enable interrupts for example.
This may not work well, if some other code was relying on interrupts
being disabled.
Use spin_lock_irqsave() and spin_unlock_restore() instead.
This patch doesn't claim that it fixes the JIRA completely, but
the issue was harder to reproduce for some iterations after this, which
was quite easy to reproduce earlier on.
Tested on EVT 2.0 with lots of debug patches to kernel and greybus.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
This reverts commit 469fbe5da0229edcb42aa08bef8e10feaa37e6d7.
It isn't correct in places.
Reported-by: Gjorgji Rosikopulos <rosikopulos_gjorgji@projectara.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
spin_[un]lock_irq() routines should be used carefully as they things can
go wrong, if they are mixed with spin_lock_irqsave() or other variants.
The main problem is that spin_[un]lock_irq() routines doesn't check if
the IRQs are already disabled/enabled on the local CPU and so
spin_unlock_irq() will forcefully enable interrupts for example.
This may not work well, if some other code was relying on interrupts
being disabled.
Use spin_lock_irqsave() and spin_unlock_restore() instead.
This patch doesn't claim that it fixes the JIRA completely, but
the issue was harder to reproduce for some iterations after this, which
was quite easy to reproduce earlier on.
Tested on EVT 2.0 with lots of debug patches to kernel and greybus.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Jeffrey Carlyle <jcarlyle@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
This reverts commit 9b891f4fda8dfd6c1d8dc16479c5f6d418a9ccc7.
We discussed this over the other thread,
[PATCH 0/2] Improve watchdog's implementation a bit,
and decided that we shouldn't be trying to hide the watchdog reboot
problem by using such patches, rather we should make sure they occur
consistently so that the real problem can be fixed.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|
|
It is believed that excessive serial messages from greybus
on suspend/resume is leading to watchdog bite.
There is still discussion going on whether ratelimiting
prints would really fix anything, except it may reduce
traffic on serial console and probably bring out real issues
in the front.
So in order to meet the alpha requirement, we all decided to
get ratelimit change "as a TEMP fix" and decide later whether
we should revert back once we get proper suspend/resume implementation.
Please follow the discussion on Jira card SW-6261.
Testing Done: Build tested against ara/main branch.
Signed-off-by: Vaibhav Hiremath <vaibhav.hiremath@linaro.org>
Reviewed-by: Konstantin Buhchev <buhchev_konstantin@projectara.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
|