The function add_mdev_supported_type() was renamed mdev_type_add() in
commit da44c340c4 ("vfio/mdev: simplify mdev_type handling").
Update the comment accordingly.
Note that just as mdev_type_release() now states that its put pairs
with the get in mdev_type_add(), mdev_type_add() already stated that
its get pairs with the put in mdev_type_release().
Signed-off-by: Julia Lawall <[email protected]>
Reviewed-by: Kirti Wankhede <[email protected]>
Reviewed-by: Jason Gunthorpe <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alex Williamson <[email protected]>
This 'struct kobj_type' is not modified. It is only used in
kobject_init_and_add() which takes a 'const struct kobj_type *ktype'
parameter.
Constifying this structure and moving it to a read-only section,
and this can increase over all security.
```
[Before]
text data bss dec hex filename
2372 600 0 2972 b9c drivers/vfio/mdev/mdev_sysfs.o
[After]
text data bss dec hex filename
2436 568 0 3004 bbc drivers/vfio/mdev/mdev_sysfs.o
```
Signed-off-by: Hongbo Li <[email protected]>
Reviewed-by: Jason Gunthorpe <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alex Williamson <[email protected]>
Many of the mdev drivers use a simple counter for keeping track of the
available instances. Move this code to the core code and store the counter
in the mdev_parent. Implement it using correct locking, fixing mdpy.
Drivers just provide the value in the mdev_driver at registration time
and the core code takes care of maintaining it and exposing the value in
sysfs.
[hch: count instances per-parent instead of per-type, use an atomic_t
to avoid taking mdev_list_lock in the show method]
Signed-off-by: Jason Gunthorpe <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Kirti Wankhede <[email protected]>
Reviewed-by: Eric Farman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alex Williamson <[email protected]>
Every driver just emits a string, simply add a method to the mdev_driver
to return it and provide a standard sysfs show function.
Remove the now unused types_attrs field in struct mdev_driver and the
support code for it.
Signed-off-by: Christoph Hellwig <[email protected]>
Reviewed-by: Jason Gunthorpe <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Kirti Wankhede <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alex Williamson <[email protected]>
Instead of abusing struct attribute_group to control initialization of
struct mdev_type, just define the actual attributes in the mdev_driver,
allocate the mdev_type structures in the caller and pass them to
mdev_register_parent.
This allows the caller to use container_of to get at the containing
structure and thus significantly simplify the code.
Signed-off-by: Christoph Hellwig <[email protected]>
Reviewed-by: Jason Gunthorpe <[email protected]>
Reviewed-by: Tony Krowiak <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Kirti Wankhede <[email protected]>
Reviewed-by: Eric Farman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alex Williamson <[email protected]>
Simplify mdev_{un}register_device by requiring the caller to pass in
a structure allocate as part of the parent device structure. This
removes the need for a list of parents and the separate mdev_parent
refcount as we can simplify rely on the reference to the parent device.
Signed-off-by: Christoph Hellwig <[email protected]>
Reviewed-by: Jason Gunthorpe <[email protected]>
Reviewed-by: Tony Krowiak <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Kirti Wankhede <[email protected]>
Reviewed-by: Eric Farman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alex Williamson <[email protected]>
The device creator is supposed to use the dev.groups value to add sysfs
files before device_add is called, not call sysfs_create_files() after
device_add() returns. This creates a race with uevent delivery where the
extra attribute will not be visible.
This was being done because the groups had been co-opted by the mdev
driver, now that prior patches have moved the driver's groups to the
struct device_driver the dev.group is properly free for use here.
Signed-off-by: Jason Gunthorpe <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
Signed-off-by: Zhi Wang <[email protected]>
Link: http://patchwork.freedesktop.org/patch/msgid/[email protected]
Reviewed-by: Kirti Wankhede <[email protected]>
Reviewed-by: Zhi Wang <[email protected]>
This is only used by one sample to print a fixed string that is pointless.
In general, having a device driver attach sysfs attributes to the parent
is horrific. This should never happen, and always leads to some kind of
liftime bug as it become very difficult for the sysfs attribute to go back
to any data owned by the device driver.
Remove the general mechanism to create this abuse.
Signed-off-by: Jason Gunthorpe <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
Signed-off-by: Zhi Wang <[email protected]>
Link: http://patchwork.freedesktop.org/patch/msgid/[email protected]
Reviewed-by: Kirti Wankhede <[email protected]>
Reviewed-by: Zhi Wang <[email protected]>
The driver core standard is to pass in the properly typed object, the
properly typed attribute and the buffer data. It stems from the root
kobject method:
ssize_t (*show)(struct kobject *kobj, struct kobj_attribute *attr,..)
Each subclass of kobject should provide their own function with the same
signature but more specific types, eg struct device uses:
ssize_t (*show)(struct device *dev, struct device_attribute *attr,..)
In this case the existing signature is:
ssize_t (*show)(struct kobject *kobj, struct device *dev,..)
Where kobj is a 'struct mdev_type *' and dev is 'mdev_type->parent->dev'.
Change the mdev_type related sysfs attribute functions to:
ssize_t (*show)(struct mdev_type *mtype, struct mdev_type_attribute *attr,..)
In order to restore type safety and match the driver core standard
There are no current users of 'attr', but if it is ever needed it would be
hard to add in retroactively, so do it now.
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Signed-off-by: Jason Gunthorpe <[email protected]>
Message-Id: <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
This returns the index in the supported_type_groups array that is
associated with the mdev_type attached to the struct mdev_device or its
containing struct kobject.
Each mdev_device can be spawned from exactly one mdev_type, which in turn
originates from exactly one supported_type_group.
Drivers are using weird string calculations to try and get back to this
index, providing a direct access to the index removes a bunch of wonky
driver code.
mdev_type->group can be deleted as the group is obtained using the
type_group_id.
Reviewed-by: Christoph Hellwig <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Jason Gunthorpe <[email protected]>
Message-Id: <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
struct mdev_type holds a pointer to the kref'd object struct mdev_parent,
but doesn't hold the kref. The lifetime of the parent becomes implicit
because parent_remove_sysfs_files() is supposed to remove all the access
before the parent can be freed, but this is very hard to reason about.
Make it obviously correct by adding the missing get.
Reviewed-by: Christoph Hellwig <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Jason Gunthorpe <[email protected]>
Message-Id: <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
The kobj pointer in mdev_device is actually pointing at a struct
mdev_type. Use the proper type so things are understandable.
There are a number of places that are confused and passing both the mdev
and the mtype as function arguments, fix these to derive the mtype
directly from the mdev to remove the redundancy.
Reviewed-by: Christoph Hellwig <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Jason Gunthorpe <[email protected]>
Message-Id: <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
The mdev API should accept and pass a 'struct mdev_device *' in all
places, not pass a 'struct device *' and cast it internally with
to_mdev_device(). Particularly in its struct mdev_driver functions, the
whole point of a bus's struct device_driver wrapper is to provide type
safety compared to the default struct device_driver.
Further, the driver core standard is for bus drivers to expose their
device structure in their public headers that can be used with
container_of() inlines and '&foo->dev' to go between the class levels, and
'&foo->dev' to be used with dev_err/etc driver core helper functions. Move
'struct mdev_device' to mdev.h
Once done this allows moving some one instruction exported functions to
static inlines, which in turns allows removing one of the two grotesque
symbol_get()'s related to mdev in the core code.
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Jason Gunthorpe <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Message-Id: <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
There is a small race where the parent is NULL even though the kobj has
already been made visible in sysfs.
For instance the attribute_group is made visible in sysfs_create_files()
and the mdev_type_attr_show() does:
ret = attr->show(kobj, type->parent->dev, buf);
Which will crash on NULL parent. Move the parent setup to before the type
pointer leaves the stack frame.
Fixes: 7b96953bc6 ("vfio: Mediated device Core driver")
Reviewed-by: Christoph Hellwig <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Max Gurtovoy <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Jason Gunthorpe <[email protected]>
Message-Id: <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
kobject_init_and_add() takes reference even when it fails.
If this function returns an error, kobject_put() must be called to
properly clean up the memory associated with the object. Thus,
replace kfree() by kobject_put() to fix this issue. Previous
commit "b8eb718348b8" fixed a similar problem.
Fixes: 7b96953bc6 ("vfio: Mediated device Core driver")
Signed-off-by: Qiushi Wu <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Reviewed-by: Kirti Wankhede <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
The create attribute is not exported, so make it
static to avoid the following sparse warning:
drivers/vfio/mdev/mdev_sysfs.c:77:1: warning: symbol 'mdev_type_attr_create' was not declared. Should it be static?
Signed-off-by: Ben Dooks (Codethink) <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
Based on 2 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license version 2 as
published by the free software foundation
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license version 2 as
published by the free software foundation #
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 4122 file(s).
Signed-off-by: Thomas Gleixner <[email protected]>
Reviewed-by: Enrico Weigelt <[email protected]>
Reviewed-by: Kate Stewart <[email protected]>
Reviewed-by: Allison Randal <[email protected]>
Cc: [email protected]
Link: https://lkml.kernel.org/r/[email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>
If device is removal is initiated by two threads as below, mdev core
attempts to create a syfs remove file on stale device.
During this flow, below [1] call trace is observed.
cpu-0 cpu-1
----- -----
mdev_unregister_device()
device_for_each_child
mdev_device_remove_cb
mdev_device_remove
user_syscall
remove_store()
mdev_device_remove()
[..]
unregister device();
/* not found in list or
* active=false.
*/
sysfs_create_file()
..Call trace
Now that mdev core follows correct device removal sequence of the linux
bus model, remove shouldn't fail in normal cases. If it fails, there is
no point of creating a stale file or checking for specific error status.
kernel: WARNING: CPU: 2 PID: 9348 at fs/sysfs/file.c:327
sysfs_create_file_ns+0x7f/0x90
kernel: CPU: 2 PID: 9348 Comm: bash Kdump: loaded Not tainted
5.1.0-rc6-vdevbus+ #6
kernel: Hardware name: Supermicro SYS-6028U-TR4+/X10DRU-i+, BIOS 2.0b
08/09/2016
kernel: RIP: 0010:sysfs_create_file_ns+0x7f/0x90
kernel: Call Trace:
kernel: remove_store+0xdc/0x100 [mdev]
kernel: kernfs_fop_write+0x113/0x1a0
kernel: vfs_write+0xad/0x1b0
kernel: ksys_write+0x5a/0xe0
kernel: do_syscall_64+0x5a/0x210
kernel: entry_SYSCALL_64_after_hwframe+0x49/0xbe
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Parav Pandit <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
This patch addresses below two issues and prepares the code to address
3rd issue listed below.
1. mdev device is placed on the mdev bus before it is created in the
vendor driver. Once a device is placed on the mdev bus without creating
its supporting underlying vendor device, mdev driver's probe() gets
triggered. However there isn't a stable mdev available to work on.
create_store()
mdev_create_device()
device_register()
...
vfio_mdev_probe()
[...]
parent->ops->create()
vfio_ap_mdev_create()
mdev_set_drvdata(mdev, matrix_mdev);
/* Valid pointer set above */
Due to this way of initialization, mdev driver who wants to use the mdev,
doesn't have a valid mdev to work on.
2. Current creation sequence is,
parent->ops_create()
groups_register()
Remove sequence is,
parent->ops->remove()
groups_unregister()
However, remove sequence should be exact mirror of creation sequence.
Once this is achieved, all users of the mdev will be terminated first
before removing underlying vendor device.
(Follow standard linux driver model).
At that point vendor's remove() ops shouldn't fail because taking the
device off the bus should terminate any usage.
3. When remove operation fails, mdev sysfs removal attempts to add the
file back on already removed device. Following call trace [1] is observed.
[1] call trace:
kernel: WARNING: CPU: 2 PID: 9348 at fs/sysfs/file.c:327 sysfs_create_file_ns+0x7f/0x90
kernel: CPU: 2 PID: 9348 Comm: bash Kdump: loaded Not tainted 5.1.0-rc6-vdevbus+ #6
kernel: Hardware name: Supermicro SYS-6028U-TR4+/X10DRU-i+, BIOS 2.0b 08/09/2016
kernel: RIP: 0010:sysfs_create_file_ns+0x7f/0x90
kernel: Call Trace:
kernel: remove_store+0xdc/0x100 [mdev]
kernel: kernfs_fop_write+0x113/0x1a0
kernel: vfs_write+0xad/0x1b0
kernel: ksys_write+0x5a/0xe0
kernel: do_syscall_64+0x5a/0x210
kernel: entry_SYSCALL_64_after_hwframe+0x49/0xbe
Therefore, mdev core is improved in following ways.
1. Split the device registration/deregistration sequence so that some
things can be done between initialization of the device and hooking it
up to the bus respectively after deregistering it from the bus but
before giving up our final reference.
In particular, this means invoking the ->create() and ->remove()
callbacks in those new windows. This gives the vendor driver an
initialized mdev device to work with during creation.
At the same time, a bus driver who wish to bind to mdev driver also
gets initialized mdev device.
This follows standard Linux kernel bus and device model.
2. During remove flow, first remove the device from the bus. This
ensures that any bus specific devices are removed.
Once device is taken off the mdev bus, invoke remove() of mdev
from the vendor driver.
3. The driver core device model provides way to register and auto
unregister the device sysfs attribute groups at dev->groups.
Make use of dev->groups to let core create the groups and eliminate
code to avoid explicit groups creation and removal.
To ensure, that new sequence is solid, a below stack dump of a
process is taken who attempts to remove the device while device is in
use by vfio driver and user application.
This stack dump validates that vfio driver guards against such device
removal when device is in use.
cat /proc/21962/stack
[<0>] vfio_del_group_dev+0x216/0x3c0 [vfio]
[<0>] mdev_remove+0x21/0x40 [mdev]
[<0>] device_release_driver_internal+0xe8/0x1b0
[<0>] bus_remove_device+0xf9/0x170
[<0>] device_del+0x168/0x350
[<0>] mdev_device_remove_common+0x1d/0x50 [mdev]
[<0>] mdev_device_remove+0x8c/0xd0 [mdev]
[<0>] remove_store+0x71/0x90 [mdev]
[<0>] kernfs_fop_write+0x113/0x1a0
[<0>] vfs_write+0xad/0x1b0
[<0>] ksys_write+0x5a/0xe0
[<0>] do_syscall_64+0x5a/0x210
[<0>] entry_SYSCALL_64_after_hwframe+0x49/0xbe
[<0>] 0xffffffffffffffff
This prepares the code to eliminate calling device_create_file() in
subsequent patch.
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Parav Pandit <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
mdev_remove_sysfs_files() should follow exact mirror sequence of a
create, similar to what is followed in error unwinding path of
mdev_create_sysfs_files().
Fixes: 6a62c1dfb5 ("vfio/mdev: Re-order sysfs attribute creation")
Reviewed-by: Cornelia Huck <[email protected]>
Reviewed-by: Maxim Levitsky <[email protected]>
Signed-off-by: Parav Pandit <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
There are new types and helpers that are supposed to be used in new code.
As a preparation to get rid of legacy types and API functions do
the conversion here.
Cc: Kirti Wankhede <[email protected]>
Cc: Alex Williamson <[email protected]>
Signed-off-by: Andy Shevchenko <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
Set add_mdev_supported_type as static since it is only used within
mdev_sysfs.c. This fixes -Wmissing-prototypes gcc warning.
Signed-off-by: Paolo Cretaro <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
There exists a gap at the end of mdev_device_create() where the device
is visible to userspace, but we're not yet ready to handle removal, as
triggered through the 'remove' attribute. We handle this properly in
mdev_device_remove() with an -EAGAIN return, but we can marginally
reduce this gap by adding this attribute as a final step of our sysfs
setup.
Reviewed-by: Kirti Wankhede <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Acked-by: Halil Pasic <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
Design for Mediated Device Driver:
Main purpose of this driver is to provide a common interface for mediated
device management that can be used by different drivers of different
devices.
This module provides a generic interface to create the device, add it to
mediated bus, add device to IOMMU group and then add it to vfio group.
Below is the high Level block diagram, with Nvidia, Intel and IBM devices
as example, since these are the devices which are going to actively use
this module as of now.
+---------------+
| |
| +-----------+ | mdev_register_driver() +--------------+
| | | +<------------------------+ __init() |
| | mdev | | | |
| | bus | +------------------------>+ |<-> VFIO user
| | driver | | probe()/remove() | vfio_mdev.ko | APIs
| | | | | |
| +-----------+ | +--------------+
| |
| MDEV CORE |
| MODULE |
| mdev.ko |
| +-----------+ | mdev_register_device() +--------------+
| | | +<------------------------+ |
| | | | | nvidia.ko |<-> physical
| | | +------------------------>+ | device
| | | | callback +--------------+
| | Physical | |
| | device | | mdev_register_device() +--------------+
| | interface | |<------------------------+ |
| | | | | i915.ko |<-> physical
| | | +------------------------>+ | device
| | | | callback +--------------+
| | | |
| | | | mdev_register_device() +--------------+
| | | +<------------------------+ |
| | | | | ccw_device.ko|<-> physical
| | | +------------------------>+ | device
| | | | callback +--------------+
| +-----------+ |
+---------------+
Core driver provides two types of registration interfaces:
1. Registration interface for mediated bus driver:
/**
* struct mdev_driver - Mediated device's driver
* @name: driver name
* @probe: called when new device created
* @remove:called when device removed
* @driver:device driver structure
*
**/
struct mdev_driver {
const char *name;
int (*probe) (struct device *dev);
void (*remove) (struct device *dev);
struct device_driver driver;
};
Mediated bus driver for mdev device should use this interface to register
and unregister with core driver respectively:
int mdev_register_driver(struct mdev_driver *drv, struct module *owner);
void mdev_unregister_driver(struct mdev_driver *drv);
Mediated bus driver is responsible to add/delete mediated devices to/from
VFIO group when devices are bound and unbound to the driver.
2. Physical device driver interface
This interface provides vendor driver the set APIs to manage physical
device related work in its driver. APIs are :
* dev_attr_groups: attributes of the parent device.
* mdev_attr_groups: attributes of the mediated device.
* supported_type_groups: attributes to define supported type. This is
mandatory field.
* create: to allocate basic resources in vendor driver for a mediated
device. This is mandatory to be provided by vendor driver.
* remove: to free resources in vendor driver when mediated device is
destroyed. This is mandatory to be provided by vendor driver.
* open: open callback of mediated device
* release: release callback of mediated device
* read : read emulation callback.
* write: write emulation callback.
* ioctl: ioctl callback.
* mmap: mmap emulation callback.
Drivers should use these interfaces to register and unregister device to
mdev core driver respectively:
extern int mdev_register_device(struct device *dev,
const struct parent_ops *ops);
extern void mdev_unregister_device(struct device *dev);
There are no locks to serialize above callbacks in mdev driver and
vfio_mdev driver. If required, vendor driver can have locks to serialize
above APIs in their driver.
Signed-off-by: Kirti Wankhede <[email protected]>
Signed-off-by: Neo Jia <[email protected]>
Reviewed-by: Jike Song <[email protected]>
Reviewed-by: Dong Jia Shi <[email protected]>
Signed-off-by: Alex Williamson <[email protected]>
Julia Lawall