pskb_may_pull maybe change skb->data and make eth pointer oboslete,
so eth needs to reload
Fixes: 91269e390d ("vxlan: using pskb_may_pull as early as possible")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
pskb_may_pull() called by arphdr_ok can change skb->data, so put the arp
setting after arphdr_ok to avoid the use the freed memory
Fixes: 0714812134 ("openvswitch: Eliminate memset() from flow_extract.")
Cc: Jesse Gross <jesse@nicira.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ip_setup_cork() called inside ip_append_data() steals dst entry from rt to cork
and in case errors in __ip_append_data() nobody frees stolen dst entry
Fixes: 2e77d89b2f ("net: avoid a pair of dst_hold()/dst_release() in ip_append_data()")
Signed-off-by: Vasily Averin <vvs@parallels.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Drivers currently call spi_nor_match_id() and then spi_nor_scan().
This adds a dependency on struct spi_device_id which we want to
avoid. Make spi_nor_scan() do it for them.
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
This simplifies the way we use spi_nor framework and will allow us to
drop spi_nor_match_id.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Currently, if a USB urb completes with an error, that error status
is not transferred back to the gbuf that it's associated with. For
inbound data there's not a lot we can do about an error, but for
outbound data, this means there is no notification to the submitter
that something went wrong.
For outbound data copy the urb status directly back to the gbuf as
its status. Follow USB's lead and set the status to -EINPROGRESS
while a gbuf is "in flight." Assign a gbuf an initial status value
of -EBADR to help identify use of never-set status values.
When an inbound urb fails (SVC or CPort), currently the urb is just
leaked, more or less (i.e., we lose an urb posted to receive
incoming data). Change that so such an error is reported, but
then re-submitted.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
It is expected that i2c writes may fail, and in that case the driver
simply retries some number of times before actually treating it as a
failure. Define a GB_OP_RETRY status, which is interpreted by the
i2c driver as an indication a retry is in order. We just translate
that into an EAGAIN error passed back to the i2c core.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
This patch adds the i2c driver, based on the use of Greybus operations
over Greybus connections. It basically replaces almost all of what
was previously found in "i2c-gb.c".
When gb_connection_device_init(connection) is called, any connection
that talks the GREYBUS_PROTOCOL_I2C is passed to gb_i2c_device_init()
to be initialized.
Initialization involves verifying the code is able to support the
version of the protocol. For I2C, we then query the functionality
mask, and set the retry count and timeout to default values.
After that, we set up the i2c device and associate it with the
connection. The i2c_algorithm methods are then implemented
by translating them into Greybus operations.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
The original CPort message handlers are not needed. All incoming
data is passed to handlers based on the protocol used over the
connection over which the data was transferred. So get rid of the
old CPort handler code.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
At this point all incoming messages are handled by the operation
code, so this obviates the need for the gbuf workqueue.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Set up the infrastructure for initializing connections based on
their protocol.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Create a work queue to do the bulk of processing of received
operation request or response messages.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Give the operation layer a chance to examine incoming data so that
it can handle it appropriately.
Treat the data as an operation message header. If it's a response,
look up the operation it's associated with. If it's not, create a
new operation. Copy the incoming data into the request or response
buffer. The next patch adds a work queue to pick up handling
the request or response from there.
Get rid of gb_operation_submit(). Instead, we have two functions,
one for sending an operation's request message, the other for
sending an operation's response message.
Not fully functional yet, still just filling things in...
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Add a red-black tree indexed by operation id to a connection to
allow pending operations (whose requests are in-flight) to be
found when their matching response is recieved.
Assign the id at the time an operation is inserted, and update
the operation's message header(s) to include it.
Rename gb_connection_op_id() to be more consistent with the
naming conventions being used elsewhere.
(Noting now that this may switch to a simple list implementation
based on Greg's assertion that lists are faster than red-black trees
for up to a few hundred entries.)
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
We need to track both request messages and response messages in
operations. So add another gbuf (and payload pointer) field to
the operation structure, and rename them to indicate which one
is which. Allow the creator specify the size of the response
buffer; just leave it a null pointer if the size is 0.
Define a new helper function gb_operation_gbuf_create() to
encapsulate creating either a request or a response buffer.
Any buffer associated with a connection will (eventually) have been
created as part of an operation. So stash the operation pointer in
the gbuf as the context pointer. Whether a buffer is for the
request or the response can be determined by pointer comparison.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Upcoming patches are going to set up devices based on what is
discovered in the module manifest. Get rid of the hard-coded
initialization done by gb_init_subdevs(), along with other related
code.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Every gbuf is associated with a connection when it is created. And
a connection contains a pointer to the host device that will carry
messages. So there's no need for the submit_gbuf() method to have
the host device pointer passed to it, the function can get it from
the gbuf's connection.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Don't assume the buffer data area will all be overwritten. Zero all
buffer space, to avoid sending crap over the wire.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Don't assume that input buffers have any particular content. The
only thing the gbuf layer needs to be concerned with is the presence
of the cport_id byte at the beginning of a transfer.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
When free_hd() is called, hd_mutex is held. It is the
responsibility of free_hd() to drop that mutex.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
To drop a reference on a gbuf, greybus_free_gbuf() is called. That
uses kref_put_mutex() to drop the refernce under protection of
gbuf_mutex. However the release routine, free_gbuf(), never
releases the mutex as it should. Fix that.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
Most of the disconnect routines for the "subdevs" of a module
blindly assume that initialization of the subdev was successful.
Fix this by checking for null pointers.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
We can retrieve opt from skb, no need to pass it as a parameter.
And opt should always be non-NULL, no need to check.
Cc: Krzysztof Kolasa <kkolasa@winsoft.pl>
Cc: Eric Dumazet <edumazet@google.com>
Tested-by: Krzysztof Kolasa <kkolasa@winsoft.pl>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
cookie_v4_check() allocates ip_options_rcu in the same way
with tcp_v4_save_options(), we can just make it a helper function.
Cc: Krzysztof Kolasa <kkolasa@winsoft.pl>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 971f10eca1 ("tcp: better TCP_SKB_CB layout to reduce cache line misses")
missed that cookie_v4_check() still calls ip_options_echo() which uses
IPCB(). It should use TCPCB() at TCP layer, so call __ip_options_echo()
instead.
Fixes: commit 971f10eca1 ("tcp: better TCP_SKB_CB layout to reduce cache line misses")
Cc: Krzysztof Kolasa <kkolasa@winsoft.pl>
Cc: Eric Dumazet <edumazet@google.com>
Reported-by: Krzysztof Kolasa <kkolasa@winsoft.pl>
Tested-by: Krzysztof Kolasa <kkolasa@winsoft.pl>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A future commit changes the way various vblank calls behave, which
causes drivers to break. Converting to use the drm_crtc_vblank_on/off
calls avoids this breakage.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
A refcounting leak was found of the original frame buffer attached to
the CRTC when using the page_flip ioctl, resulting in the frame buffer
never being freed.
This was not obvious initially, as if the page flip subsequently
re-attaches the original frame buffer, the refcounts will be balanced.
However, if the original frame buffer is freed, then it will be leaked.
Fix this by ensuring that we take a reference on the incoming fb, but
rely on the queued work to drop that ref count.
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This simplifies the lanai.c driver by using
the module_pci_driver() macro, at the expense
of losing only debugging messages.
Signed-off-by: Michael Opdenacker <michael.opdenacker@free-electrons.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Recently, in commit 13aa93c70e71 ("random: add and use memzero_explicit()
for clearing data"), we have found that GCC may optimize some memset()
cases away when it detects a stack variable is not being used anymore
and going out of scope. This can happen, for example, in cases when we
are clearing out sensitive information such as keying material or any
e.g. intermediate results from crypto computations, etc.
With the help of Coccinelle, we can figure out and fix such occurences
in the crypto subsytem as well. Julia Lawall provided the following
Coccinelle program:
@@
type T;
identifier x;
@@
T x;
... when exists
when any
-memset
+memzero_explicit
(&x,
-0,
...)
... when != x
when strict
@@
type T;
identifier x;
@@
T x[...];
... when exists
when any
-memset
+memzero_explicit
(x,
-0,
...)
... when != x
when strict
Therefore, make use of the drop-in replacement memzero_explicit() for
exactly such cases instead of using memset().
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Julia Lawall <julia.lawall@lip6.fr>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
zatimend has reported that in his environment (3.16/gcc4.8.3/corei7)
memset() calls which clear out sensitive data in extract_{buf,entropy,
entropy_user}() in random driver are being optimized away by gcc.
Add a helper memzero_explicit() (similarly as explicit_bzero() variants)
that can be used in such cases where a variable with sensitive data is
being cleared out in the end. Other use cases might also be in crypto
code. [ I have put this into lib/string.c though, as it's always built-in
and doesn't need any dependencies then. ]
Fixes kernel bugzilla: 82041
Reported-by: zatimend@hotmail.co.uk
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org
The only thing we need is a forward declaration for 'struct cgroup_sel',
that is inside 'struct perf_evsel'.
Include cgroup.h instead on the tools that support cgroups.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-b7kuymbgf0zxi5viyjjtu5hk@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
It was being found, by chance, because evsel.h needlessly includes
util/cgroup.h, which will be sorted out in a following patch.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-xsvxr747wkkpg1ay9dramorr@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
So that when an evsel is embedded into other struct it can free up
resources calling perf_evsel__exit().
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-n1w68pfe9m2vkhm4sqs8y1en@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
If kprobes is disabled uprobes will not compile.
Fix this by including the correct header files.
Signed-off-by: Jan Willeke <willeke@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
On the Haswell platform, a split BAR option to allow creation of 2
32bit BARs (4 and 5) from the 64bit BAR 4. Adding support for this
new option.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Instead of using a module parameter, we should detect the errata via
PCI DID and then set an appropriate flag. This will be used for additional
errata later on.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
To simplify some of the platform detection code. Move the platform detection
to a function to be called earlier.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Move the platform detection function to separate functions to allow
easier maintenence.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Create a debugfs entry for the NTB device to log the basic device info,
as well as display the error count on a number of registers.
Signed-off-by: Jon Mason <jon.mason@intel.com>