linux-stable.git/drivers/virt/acrn, branch v6.5.2

virt: acrn: Use alloc_ordered_workqueue() to create ordered workqueues

2023-05-09T01:33:06+00:00

BACKGROUND
==========

When multiple work items are queued to a workqueue, their execution order
doesn't match the queueing order. They may get executed in any order and
simultaneously. When fully serialized execution - one by one in the queueing
order - is needed, an ordered workqueue should be used which can be created
with alloc_ordered_workqueue().

However, alloc_ordered_workqueue() was a later addition. Before it, an
ordered workqueue could be obtained by creating an UNBOUND workqueue with
@max_active==1. This originally was an implementation side-effect which was
broken by 4c16bd327c74 ("workqueue: restore WQ_UNBOUND/max_active==1 to be
ordered"). Because there were users that depended on the ordered execution,
5c0338c68706 ("workqueue: restore WQ_UNBOUND/max_active==1 to be ordered")
made workqueue allocation path to implicitly promote UNBOUND workqueues w/
@max_active==1 to ordered workqueues.

While this has worked okay, overloading the UNBOUND allocation interface
this way creates other issues. It's difficult to tell whether a given
workqueue actually needs to be ordered and users that legitimately want a
min concurrency level wq unexpectedly gets an ordered one instead. With
planned UNBOUND workqueue updates to improve execution locality and more
prevalence of chiplet designs which can benefit from such improvements, this
isn't a state we wanna be in forever.

This patch series audits all callsites that create an UNBOUND workqueue w/
@max_active==1 and converts them to alloc_ordered_workqueue() as necessary.

WHAT TO LOOK FOR
================

The conversions are from

  alloc_workqueue(WQ_UNBOUND | flags, 1, args..)

to 

  alloc_ordered_workqueue(flags, args...)

which don't cause any functional changes. If you know that fully ordered
execution is not ncessary, please let me know. I'll drop the conversion and
instead add a comment noting the fact to reduce confusion while conversion
is in progress.

If you aren't fully sure, it's completely fine to let the conversion
through. The behavior will stay exactly the same and we can always
reconsider later.

As there are follow-up workqueue core changes, I'd really appreciate if the
patch can be routed through the workqueue tree w/ your acks. Thanks.

Signed-off-by: Tejun Heo 
Acked-by: Greg Kroah-Hartman 
Cc: Fei Li

virt: acrn: using for_each_set_bit to simplify the code

2022-07-08T13:42:10+00:00

It's more cleanly to use for_each_set_bit() instead of opencoding it.

Signed-off-by: Yang Yingliang 
Signed-off-by: Fei Li 
Link: https://lore.kernel.org/r/20220704125044.2192381-1-yangyingliang@huawei.com
Signed-off-by: Greg Kroah-Hartman

virt: acrn: Prefer array_size and struct_size over open coded arithmetic

2022-04-26T15:20:10+00:00

As noted in the "Deprecated Interfaces, Language Features, Attributes,
and Conventions" documentation [1], size calculations (especially
multiplication) should not be performed in memory allocator (or similar)
function arguments due to the risk of them overflowing. This could lead
to values wrapping around and a smaller allocation being made than the
caller was expecting. Using those allocations could lead to linear
overflows of heap memory and other misbehaviors.

So, use the array_size() helper to do the arithmetic instead of the
argument "count * size" in the vzalloc() function.

Also, take the opportunity to add a flexible array member of struct
vm_memory_region_op to the vm_memory_region_batch structure. And then,
change the code accordingly and use the struct_size() helper to do the
arithmetic instead of the argument "size + size * count" in the kzalloc
function.

This code was detected with the help of Coccinelle and audited and fixed
manually.

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments

Acked-by: Fei Li 
Signed-off-by: Len Baker 
Signed-off-by: Gustavo A. R. Silva

virt: acrn: fix a memory leak in acrn_dev_ioctl()

2022-03-18T12:49:59+00:00

The vm_param and cpu_regs need to be freed via kfree()
before return -EINVAL error.

Fixes: 9c5137aedd11 ("virt: acrn: Introduce VM management interfaces")
Fixes: 2ad2aaee1bc9 ("virt: acrn: Introduce an ioctl to set vCPU registers state")
Signed-off-by: Xiaolong Huang 
Signed-off-by: Fei Li 
Link: https://lore.kernel.org/r/20220308092047.1008409-1-butterflyhuangxx@gmail.com
Signed-off-by: Greg Kroah-Hartman

virt: acrn: obtain pa from VMA with PFNMAP flag

2022-03-18T12:49:56+00:00

 acrn_vm_ram_map can't pin the user pages with VM_PFNMAP flag
 by calling get_user_pages_fast(), the PA(physical pages)
 may be mapped by kernel driver and set PFNMAP flag.

 This patch fixes logic to setup EPT mapping for PFN mapped RAM region
 by checking the memory attribute before adding EPT mapping for them.

Fixes: 88f537d5e8dd ("virt: acrn: Introduce EPT mapping management")
Signed-off-by: Yonghua Huang 
Signed-off-by: Fei Li 
Link: https://lore.kernel.org/r/20220228022212.419406-1-yonghua.huang@intel.com
Signed-off-by: Greg Kroah-Hartman

virt: acrn: Remove unsued acrn_irqfds_mutex.

2022-03-18T12:49:49+00:00

acrn_irqfds_mutex is not used, never was.

Remove acrn_irqfds_mutex.

Fixes: aa3b483ff1d71 ("virt: acrn: Introduce irqfd")
Cc: Fei Li 
Signed-off-by: Sebastian Andrzej Siewior 
Link: https://lore.kernel.org/r/YidLo57Kw/u/cpA5@linutronix.de
Signed-off-by: Greg Kroah-Hartman

all: replace find_next{,_zero}_bit with find_first{,_zero}_bit where appropriate

2022-01-15T16:47:31+00:00

find_first{,_zero}_bit is a more effective analogue of 'next' version if
start == 0. This patch replaces 'next' with 'first' where things look
trivial.

Signed-off-by: Yury Norov 
Tested-by: Wolfram Sang

virt: acrn: Introduce interfaces for virtual device creating/destroying

2021-10-05T14:14:10+00:00

The ACRN hypervisor can emulate a virtual device within hypervisor for a
Guest VM. The emulated virtual device can work without the ACRN
userspace after creation. The hypervisor do the emulation of that device.

To support the virtual device creating/destroying, HSM provides the
following ioctls:
  - ACRN_IOCTL_CREATE_VDEV
    Pass data struct acrn_vdev from userspace to the hypervisor, and inform
    the hypervisor to create a virtual device for a User VM.
  - ACRN_IOCTL_DESTROY_VDEV
    Pass data struct acrn_vdev from userspace to the hypervisor, and inform
    the hypervisor to destroy a virtual device of a User VM.

These new APIs will be used by user space code vm_add_hv_vdev and
vm_remove_hv_vdev in
https://github.com/projectacrn/acrn-hypervisor/blob/master/devicemodel/core/vmmapi.c

Signed-off-by: Shuo Liu 
Signed-off-by: Fei Li 
Link: https://lore.kernel.org/r/20210923084128.18902-3-fei1.li@intel.com
Signed-off-by: Greg Kroah-Hartman

virt: acrn: Introduce interfaces for MMIO device passthrough

2021-10-05T14:14:10+00:00

MMIO device passthrough enables an OS in a virtual machine to directly
access a MMIO device in the host. It promises almost the native
performance, which is required in performance-critical scenarios of
ACRN.

HSM provides the following ioctls:
  - Assign - ACRN_IOCTL_ASSIGN_MMIODEV
    Pass data struct acrn_mmiodev from userspace to the hypervisor, and
    inform the hypervisor to assign a MMIO device to a User VM.

  - De-assign - ACRN_IOCTL_DEASSIGN_PCIDEV
    Pass data struct acrn_mmiodev from userspace to the hypervisor, and
    inform the hypervisor to de-assign a MMIO device from a User VM.

These new APIs will be used by user space code vm_assign_mmiodev and
vm_deassign_mmiodev in
https://github.com/projectacrn/acrn-hypervisor/blob/master/devicemodel/core/vmmapi.c

Signed-off-by: Shuo Liu 
Signed-off-by: Fei Li 
Link: https://lore.kernel.org/r/20210923084128.18902-2-fei1.li@intel.com
Signed-off-by: Greg Kroah-Hartman

virt: acrn: Do hcall_destroy_vm() before resource release

2021-07-27T14:48:45+00:00

The ACRN hypervisor has scenarios which could run a real-time guest VM.
The real-time guest VM occupies dedicated CPU cores, be assigned with
dedicated PCI devices. It can run without the Service VM after boot up.
hcall_destroy_vm() returns failure when a real-time guest VM refuses.
The clearing of flag ACRN_VM_FLAG_DESTROYED causes some kernel resource
double-freed in a later acrn_vm_destroy().

Do hcall_destroy_vm() before resource release to drop this chance to
destroy the VM if hypercall fails.

Fixes: 9c5137aedd11 ("virt: acrn: Introduce VM management interfaces")
Cc: stable 
Signed-off-by: Shuo Liu 
Signed-off-by: Fei Li 
Link: https://lore.kernel.org/r/20210722062736.15050-1-fei1.li@intel.com
Signed-off-by: Greg Kroah-Hartman