This fixes the memory corruption case, if nbytes is less than offset
and sizeof(struct channel_header)
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The efi framebuffer is defined within the s-Par video channel
console. Before we get the device create message for the video
console, s-Par has alreaady informed linux about the efi
framebuffer and a memory region is already set up for it. Since
we do not use the video channel in linux, we are just ignoring
the failure of the video channel request_mem_region.
Testing: This patch was tested on top of s-Par and we no longer
leave the partition in a failed state.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fix coding style issue "Missing a blank line after declarations"
detected by checkpatch.pl in visorchipset.c
Signed-off-by: Alessandro Parini <alessandro.parini@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fix coding style issue "braces {} are not necessary for single statement blocks"
detected by checkpatch.pl in visorchipset.c
Signed-off-by: Alessandro Parini <alessandro.parini@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently if a driver is not loaded for a device, we will not
respond to the device create until it is done. This causes
s-Par to not mark the partition running if the driver for the
device is not loaded. Since there are several devices that
could be assigned to a guest that don't have an actual
driver this will cause us to never go running.
If the device driver is loaded, we WILL continue to only
respond to the device PAUSE message when the device driver
has responded that it is done with the device.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This moves phys_info to iochannel.h. It is only
used by iochannel.h and is not needed in
vmcallinterfaces.h.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch remove unneeded variable used to store return value.
Signed-off-by: Abdul Hussain <habdul@visteon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the bus device was created the list_all variables were not
being initialized. When the CONTROLVM_BUS_CONFIGURE message
was being sent, it was failing to find the bus and produced a
panic. Initialize the bus_info->list_all variable by doing a
INIT_LIST_HEAD.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When unloading a module, we need to cleanup the platform registration.
However, unregistering the platform uncovered a couple of quirks, namely
a missing device_release function. Fix things up so module unload works
and allows us to reload the module.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When we combined visorchipset and visorbus into one driver we negelected
to strip out some of the MODULE_ stuff from one of the files. When
building the drivers in, it causes a WARN that we try to create
/proc/modules/visorbus/version when it is already created.
visorchipset.c is the driver entry point, remove the cruft from
visorbus_main.c.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Diagchannel.h is used primarily for the diagnostics channel. The
diagnostics channel is not being used by linux guests currently, so
the majority of the file is not needed. What is left is what is
needed to perform postcode vmcalls. Those postcodes will eventually
end up in the diag channel.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch removes the legacy dev_info struct and instead creates
and passes around a traditional struct device.
This allows us to remove a lot of the various look up code and
removes the doubt if the struct exists or not.
Half of the churn is just the conversion of visorchipset_device_info
to visor_device. Various cleanups include re-arranging the failure
paths to make more sense.
Pay attention to the create_visor_device function. This had a lot of
churn to simplify everything.
Lots of functions disappeared because they are not needed any more.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch removes the legacy bus_info struct and instead creates
and passes around a traditional struct device.
This allows us to remove a lot of the various look up code and
removes the doubt if the struct exists or not.
Half of the churn is just the conversion of visorchipset_bus_info
to visor_device. Various cleanups include re-arranging the failure
paths to make more sense.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
I used 0 as the device id for the bus root, neglecting the fact that
device 0 is a valid id in Unisys's configuration. Modify this to
use UINT_MAX instead as a unique number.
As fallout from this change it was noticed the bus_no and dev_no was not
defined the same way consistently. Fix visorbus.h to use u32 there. Fix
the resulting printk warning too.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The bus driver doesn't work in server mode, just remove the left over
pieces.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Removing stuff that isn't being used. Another prepartion patch to
allow us to use visor_device everywhere without the baggage of
bus/dev_info.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The bus device and regular device were using two different
structs. Let's combine them as they are not entirely different
from one another.
This allows us to move this creation up the stack later and
actually remove bus/dev_info easily.
Most of the churn is just renaming devdata -> dev and 'struct
visorbus_devdata' to 'struct visor_device'.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
API consolidation with coccinelle found:
./drivers/staging/unisys/visorbus/periodic_work.c:196:3-19:
consolidation with schedule_timeout_*() recommended
This is a 1:1 conversion with respect to schedule_timeout() to the
schedule_timeout_interruptible() helper only - so only an API
consolidation to improve readability. The hard coded timeout of 10
jiffies is HZ dependent which it should not be, so it is converted
with msecs_to_jiffies.
Patch was compile tested with x86_64_defconfig + CONFIG_STAGING=y,
CONFIG_UNISYSSPAR=y, CONFIG_UNISYS_VISORBUS=m
Patch is against 4.1-rc5 (localversion-next is -next-20150529)
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In order to remove bus_info, we need to migrate vbus_hdr_info into
the public namespace of visor_device. Because the struct is private,
we use a void * to hide the contents.
As a result, we need to allocate vbus_hdr_info and manage it.
Also work around vbus_valid, as that variable will not be used
in the public namespace.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In order for bus/dev_info structs to become public structs, one
element, pending_msg_hdr, needs to become opaque. This is to keep
all the internals of the controlvm struct private to the bus layer.
So a simple conversion of embedding the pending_msg_hdr struct into
a pointer is done. The rest of the patch is the fallout.
The rules are modified slightly. Instead of relying on the 'id' to be
CONTROLVM_INVALID to indicate invalid, just use the pointer set to NULL.
In addition, because bus/dev_info can be NULL and we still need to send
a
response, pass pending_msg_hdr to all 'responders' instead of bus/
That change causes some fallout in the success case. Instead of setting
state
bits and clearing info in the responders, do all that magic in the
responder
wrappers.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Instead of creating a channel struct to temporarily hold the channel
info and passing it through multiple functions until the device is
created, just create the channel from the start.
This allows us to remove the channel_info struct.
I noticed 'chan_info.addr_type' was not being used, so I just deleted it.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add static declarations to statisfy sparse warnings in:
drivers/staging/unisys/visorbus/visorbus_main.c
warning: symbol 'visorbus_debug' was not declared
warning: symbol 'visorbus_forcematch' was not declared
warning: symbol 'visorbus_forcenomatch' was not declared
warning: symbol 'visorbus_devicetest' was not declared
warning: symbol 'visorbus_debugref' was not declared
warning: symbol 'visorbus_bus_groups' was not declared
warning: symbol 'devmajorminor_create_file' was not declared
warning: symbol 'devmajorminor_remove_file' was not declared
warning: symbol 'devmajorminor_remove_all_files' was not declared
warning: symbol 'unregister_devmajorminor_attributes' was not declared
Signed-off-by: Drew Fustini <pdp7pdp7@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Most device functions pass bus_no and dev_no around and then do a lookup
inside each function to find the dev_info struct. Instead just pass the
pointer.
This prepares us for a later conversion to using visor device.
No real technical changes. Just function header changes and little
cleanups as a result.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Most bus functions pass bus_no around and then do a lookup inside each
function to find the bus_info struct. Instead just pass the pointer.
This prepares us for a later conversion to using visor_device.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Convert the device attribute files to properly use visor_device.
This removes a whole bunch of checks and assumptions and simplifies
the code. Everything is straightforward.
No testing down as I can't mimic channel info correctl.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The conversion to visor_device caused some compile issues.The main
problem was the new fields in 'struct visor_device' were not public.
Remove one that wasn't being used for now.
struct irq_info intr
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There was a bunch of channel creation checks before the
visorchannel_create function was called, moving some of those
checks inside. This keeps the outside code cleaner and handles
the situation where a caller forgets to make these checks.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch is an attempt to help hide the channel info behind
accessory functions. I was trying to keep visorchannel as private
as possible.
The only function missing that seemed to be needed for now was
the ability to set the clientpartition. So I expose that here.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If we are going to remove the bus_info structs than we need a way
to find the devices when the *_create/destroy cmds are sent over
the vmchannel.
This function crudely impements what pci has. It takes a bus_no
and dev_no and finds the matching 'struct visor_device'.
This function can/should be optimzed later once we get our heads
wrapped around its needs. For now, I am using dev_no=0 to mean
the visorbus itself.
The function is limited to chipset.c only because it is only needed
to do the lookups upon receiving a vmchannel command. Future patches
will make sure the resulting 'struct visor_device' is used every
where else.
Also allow visorbus_type to be more visible for use.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The visorbus driver has three _info structs lying around
(device, bus, channel) that store subsets of info from the
bigger structs.
Having these structs around make resource handling very difficult
and more complicated than it needs to be. Use the device
infrastructure and instead pass 'struct visor_device' all
over the place.
In order to do that 'struct visor_device' needs to get smarter.
This patch adds the pieces to prep for it. The new elements
will be used in later patches.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Properly hook into the struct device groups element. This allows the
core infrastructure to manage the files instead of the bus layer. And
makes the code easier to read.
I didn't clean up the _show functions just modified them a bit to handle
the different args for now and to prevent a build warning.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove stale code.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Simplify things by moving the version file handling into the core.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Mimicing what other drivers do, this seems appropriate. Yeah, it
is a bus, but it is a bus _device_. This makes things work better
and smoother. Now the sysfs looks like
[root@dhcp-17-174 visorbus]# ls -l /sys/bus/visorbus/devices/
total 0
lrwxrwxrwx. 1 root root 0 Apr 17 16:09 vbus1:dev2 ->
../../../devices/visorbus1/vbus1:dev2
lrwxrwxrwx. 1 root root 0 Apr 17 16:09 visorbus1 ->
../../../devices/visorbus1
Which looks correct. All the attributes are still correct too, based on my
very minimal testing of 'ls -lR'. :-)
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch moves the attributes to underneath the bus device correctly.
This will help remove a bunch of cruft from the code and let the kernel
infrastructure manage the sysfs files instead of the driver.
After the move remove all the leftover code.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Handling the sysfs attributes become easier to deal with when you can just
run container_of(dev) to get devdata.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Now that the header file movement is complete, remove the temporary ccflags
include
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
remove uisqueue.h, uisthread.h, and uisutils.h
replace HOSTADDRESS with u64
remove "uisutils.h" from header list in visorchipset.c
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Makefiles still had common-spar listed in ccflags.
This gets rid of them.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move last three files out of common-spar
iochannel.h --> include (will be used by visorhba and visornic)
version.h --> moved to include
controlvmcompletionstatus.h --> moved to visorbus, part of
controlvmchannel.h
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Channel.h is used by all channels, it needs to be in include.
Controlvm Channel is only used by visorbus, it needs to just be there.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move hypervisor calls into visorbus and move vbusdeviceinfo.h into
visorbus.
Drivers will call into that to update clientInfo field.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To prep for the moving of include files, temporarily add the
visorbus/ as a ccflags -I. Once the header files are all
transitioned, we can remove this.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This preps for the possible build breakage in the next few
patches.
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
[updated changelog to reflect new patch order; instead
of 'fixes' the patch 'prevents' - dcz]
Signed-off-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Jes Sorensen