KVM: Improve MMIO Coalescing API
Subsystems (3)
v2 cover KVM: Improve MMIO Coalescing API
The current MMIO coalescing design has a few drawbacks which limit its usefulness. Currently all coalesced MMIO zones use the same ring buffer. That means that upon a userspace exit we have to handle potentially unrelated MMIO writes synchronously. And a VM-wide lock needs to be taken in the kernel when an MMIO exit occurs.
Additionally, there is no direct way for userspace to be notified about coalesced MMIO writes. If the next MMIO exit to userspace is when the ring buffer has filled then a substantial (and unbounded) amount of time may have passed since the first coalesced MMIO.
This series adds new ioctls to KVM that allow for greater control by making it possible to associate different MMIO zones with different ring buffers. It also allows userspace to use poll() to check for coalesced writes in order to avoid userspace exits in vCPU threads (see patch 3 for why this can be useful).
The idea of improving the API in this way originally came from Paul Durrant ([email protected]) but the implementation is mine.
The first patch in the series is a bug in the existing code that I discovered while writing a selftest and can be merged independently.
Changelog:
v2:
- Fixed a bug in poll() in patch 3 where a NULL check was omitted.
- Added an explicit include to allow the build to succeed on powerpc.
- Rebased on top kvm/queue resolving a few conflicts.
- Tiny changes on the last 2 patches.
v1: https://lore.kernel.org/kvm/[email protected]/T/
Ilias Stamatis (6):
KVM: Fix coalesced_mmio_has_room()
KVM: Add KVM_CREATE_COALESCED_MMIO_BUFFER ioctl
KVM: Support poll() on coalesced mmio buffer fds
KVM: Add KVM_(UN)REGISTER_COALESCED_MMIO2 ioctls
KVM: Documentation: Document v2 of coalesced MMIO API
KVM: selftests: Add coalesced_mmio_test
Documentation/virt/kvm/api.rst | 91 +++++ include/linux/kvm_host.h | 1 + include/uapi/linux/kvm.h | 18 + tools/testing/selftests/kvm/Makefile | 1 + .../selftests/kvm/coalesced_mmio_test.c | 313 ++++++++++++++++++ virt/kvm/coalesced_mmio.c | 205 +++++++++++- virt/kvm/coalesced_mmio.h | 17 +- virt/kvm/kvm_main.c | 40 ++- 8 files changed, 665 insertions(+), 21 deletions(-) create mode 100644 tools/testing/selftests/kvm/coalesced_mmio_test.c
Signature · 1 line
2.34.1
PATCH v2 1/6 KVM: Fix coalesced_mmio_has_room()
The following calculation used in coalesced_mmio_has_room() to check whether the ring buffer is full is wrong and only allows half the buffer to be used.
avail = (ring->first - last - 1) % KVM_COALESCED_MMIO_MAX; if (avail == 0)
/* full */
The % operator in C is not the modulo operator but the remainder operator. Modulo and remainder operators differ with respect to negative values. But all values are unsigned in this case anyway.
The above might have worked as expected in python for example:
Quoted 1 line
Quoted 1 line
Quoted 1 line
(-86) % 170
84
However it doesn't work the same way in C.
printf("avail: %d\n", (-86) % 170);printf("avail: %u\n", (-86) % 170);printf("avail: %u\n", (-86u) % 170u);
Using gcc-11 these print:
avail: -86
avail: 4294967210
avail: 0
Fix the calculation and allow all but one entries in the buffer to be used as originally intended.
--- virt/kvm/coalesced_mmio.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-)
virt/kvm/coalesced_mmio.c +1 -3
@@ -43,7 +43,6 @@ static int coalesced_mmio_in_range(struct kvm_coalesced_mmio_dev *dev,static int coalesced_mmio_has_room (struct kvm_coalesced_mmio_dev *dev,u32 last) {struct kvm_coalesced_mmio_ring *ring; -unsigned avail;/* Are we able to batch it ? */
@@ -52,8 +51,7 @@ static int coalesced_mmio_has_room(struct kvm_coalesced_mmio_dev *dev, u32 last)* there is always one unused entry in the buffer */ ring = dev->kvm->coalesced_mmio_ring; - avail = (ring->first - last -1 ) %KVM_COALESCED_MMIO_MAX ; -if (avail ==0 ) { +if ((last +1 ) %KVM_COALESCED_MMIO_MAX ==READ_ONCE (ring->first)) {/* full */ return 0 ; }
Signature · 1 line
2.34.1
PATCH v2 2/6 KVM: Add KVM_CREATE_COALESCED_MMIO_BUFFER ioctl
In reply to Ilias StamatisThe current MMIO coalescing design has a few drawbacks which limit its usefulness. Currently all coalesced MMIO zones use the same ring buffer. That means that upon a userspace exit we have to handle potentially unrelated MMIO writes synchronously. And a VM-wide lock needs to be taken in the kernel when an MMIO exit occurs.
Additionally, there is no direct way for userspace to be notified about coalesced MMIO writes. If the next MMIO exit to userspace is when the ring buffer has filled then a substantial (and unbounded) amount of time may have passed since the first coalesced MMIO.
Add a KVM_CREATE_COALESCED_MMIO_BUFFER ioctl to KVM. This ioctl simply returns a file descriptor to the caller but does not allocate a ring buffer. Userspace can then pass this fd to mmap() to actually allocate a buffer and map it to its address space.
Subsequent patches will allow userspace to:
- Associate the fd with a coalescing zone when registering it so that writes to that zone are accumulated in that specific ring buffer rather than the VM-wide one.
- Poll for MMIO writes using this fd.
---
v1->v2:
- Rebased on top of kvm/queue resolving the conflict in kvm.h
include/linux/kvm_host.h | 1 + include/uapi/linux/kvm.h | 2 + virt/kvm/coalesced_mmio.c | 142 +++++++++++++++++++++++++++++++++++--- virt/kvm/coalesced_mmio.h | 9 +++ virt/kvm/kvm_main.c | 4 ++ 5 files changed, 150 insertions(+), 8 deletions(-)
include/linux/kvm_host.h +1 -0
@@ -801,6 +801,7 @@ struct kvm {struct kvm_coalesced_mmio_ring *coalesced_mmio_ring;spinlock_t ring_lock;struct list_head coalesced_zones; +struct list_head coalesced_buffers;#endif struct mutex irq_lock;
include/uapi/linux/kvm.h +2 -0
@@ -1573,4 +1573,6 @@ struct kvm_pre_fault_memory { __u64 padding[5]; }; +#define KVM_CREATE_COALESCED_MMIO_BUFFER _IO(KVMIO, 0xd6) +#endif /* __LINUX_KVM_H */
virt/kvm/coalesced_mmio.c +134 -8
210 lines not on this page · raw message
virt/kvm/coalesced_mmio.h +9 -0
@@ -20,6 +20,14 @@ struct kvm_coalesced_mmio_dev {struct kvm_io_device dev;struct kvm *kvm;struct kvm_coalesced_mmio_zone zone; +struct kvm_coalesced_mmio_buffer_dev *buffer_dev; +}; + +struct kvm_coalesced_mmio_buffer_dev { +struct list_head list; +struct kvm *kvm; +spinlock_t ring_lock; +struct kvm_coalesced_mmio_ring *ring; };int kvm_coalesced_mmio_init (struct kvm *kvm);
@@ -28,6 +36,7 @@ int kvm_vm_ioctl_register_coalesced_mmio(struct kvm *kvm, struct kvm_coalesced_mmio_zone *zone);int kvm_vm_ioctl_unregister_coalesced_mmio (struct kvm *kvm,struct kvm_coalesced_mmio_zone *zone); +int kvm_vm_ioctl_create_coalesced_mmio_buffer (struct kvm *kvm);#else
virt/kvm/kvm_main.c +4 -0
@@ -5246,6 +5246,10 @@ static long kvm_vm_ioctl(struct file *filp, r =kvm_vm_ioctl_unregister_coalesced_mmio (kvm, &zone); break; } +case KVM_CREATE_COALESCED_MMIO_BUFFER : { + r =kvm_vm_ioctl_create_coalesced_mmio_buffer (kvm); +break ; + }#endif case KVM_IRQFD : {struct kvm_irqfd data;
Signature · 1 line
2.34.1
Sean Christopherson
reviewerRe: [PATCH v2 2/6] KVM: Add KVM_CREATE_COALESCED_MMIO_BUFFER ioctl
Quoted 20 lines · Ilias Stamatis
virt/kvm/coalesced_mmio.c +1
-4@@ -40,17 +42,14 @@ static int coalesced_mmio_in_range(struct kvm_coalesced_mmio_dev *dev,return 1 ; } -static int coalesced_mmio_has_room (struct kvm_coalesced_mmio_dev *dev,u32 last) +static int coalesced_mmio_has_room (struct kvm_coalesced_mmio_ring *ring,u32 last) { -struct kvm_coalesced_mmio_ring *ring; -/* Are we able to batch it ? */ /* last is the first free entry * check if we don't meet the first used entry * there is always one unused entry in the buffer */ - ring = dev->kvm->coalesced_mmio_ring;if ((last +1 ) %KVM_COALESCED_MMIO_MAX ==READ_ONCE (ring->first)) {/* full */ return 0 ;
virt/kvm/coalesced_mmio.c static int coalesced_mmio_write(struct kvm_vcpu *vcpu, · line 69
Quoted 6 lines
virt/kvm/coalesced_mmio.c +3
-0@@ -65,17 +64,28 @@ static int coalesced_mmio_write(struct kvm_vcpu *vcpu, {struct kvm_coalesced_mmio_dev *dev =to_mmio (this);struct kvm_coalesced_mmio_ring *ring = dev->kvm->coalesced_mmio_ring; +spinlock_t *lock = dev->buffer_dev ? + &dev->buffer_dev->ring_lock : + &dev->kvm->ring_lock;
I'd prefer
spinlock_t *lock = dev->buffer_dev ? &dev->buffer_dev->ring_lock : &dev->kvm->ring_lock;
to make it more obvious which parts the "guts" of the effective if-statement.
virt/kvm/kvm_main.c static long kvm_vm_ioctl(struct file *filp, · line 5249
Quoted 9 lines
virt/kvm/kvm_main.c +1
-0@@ -5246,6 +5246,10 @@ static long kvm_vm_ioctl(struct file *filp, r =kvm_vm_ioctl_unregister_coalesced_mmio (kvm, &zone);break ; } +case KVM_CREATE_COALESCED_MMIO_BUFFER : {
Curly braces aren't needed, the above case has 'em because it declares local variables.
Quoted 9 lines
virt/kvm/kvm_main.c +3
-0+ r =kvm_vm_ioctl_create_coalesced_mmio_buffer (kvm); + break; + }#endif case KVM_IRQFD : {struct kvm_irqfd data;Signature · 1 line
2.34.1
PATCH v2 3/6 KVM: Support poll() on coalesced mmio buffer fds
In reply to Ilias StamatisThere is no direct way for userspace to be notified about coalesced MMIO writes when using KVM_REGISTER_COALESCED_MMIO. If the next MMIO exit is when the ring buffer has filled then a substantial (and unbounded) amount of time may have passed since the first coalesced MMIO.
To improve this, make it possible for userspace to use poll() and select() on the fd returned by the KVM_CREATE_COALESCED_MMIO_BUFFER ioctl. This way a userspace VMM could have dedicated threads that deal with writes to specific MMIO zones.
For example, a common use of MMIO, particularly in the realm of network devices, is as a doorbell. A write to a doorbell register will trigger the device to initiate a DMA transfer.
When a network device is emulated by userspace a write to a doorbell register would typically result in an MMIO exit so that userspace can emulate the DMA transfer in a timely manner. No further processing can be done until userspace performs the necessary emulation and re-invokes KVM_RUN. Even if userspace makes use of another thread to emulate the DMA transfer such MMIO exits are disruptive to the vCPU and they may also be quite frequent if, for example, the vCPU is sending a sequence of short packets to the network device.
By supporting poll() on coalesced buffer fds, userspace can have dedicated threads wait for new doorbell writes and avoid the performance hit of userspace exits on the main vCPU threads.
---
v1->v2:
- Added a NULL-check for dev->ring in coalesced_mmio_buffer_poll()
virt/kvm/coalesced_mmio.c | 22 ++++++++++++++++++++++ virt/kvm/coalesced_mmio.h | 1 + 2 files changed, 23 insertions(+)
virt/kvm/coalesced_mmio.c +22 -0
@@ -16,6 +16,7 @@#include <linux/slab.h> #include <linux/kvm.h> #include <linux/anon_inodes.h> +#include <linux/poll.h> #include "coalesced_mmio.h"
@@ -98,6 +99,10 @@ static int coalesced_mmio_write(struct kvm_vcpu *vcpu,smp_wmb (); ring->last = (insert +1 ) %KVM_COALESCED_MMIO_MAX ;spin_unlock (lock); + +if (dev->buffer_dev) +wake_up_interruptible (&dev->buffer_dev->wait_queue); +return 0 ; }
@@ -224,9 +229,25 @@ static int coalesced_mmio_buffer_release(struct inode *inode, struct file *file)return 0 ; } +static __poll_t coalesced_mmio_buffer_poll (struct file *file,struct poll_table_struct *wait) +{ +struct kvm_coalesced_mmio_buffer_dev *dev = file->private_data; +__poll_t mask =0 ; + +poll_wait (file, &dev->wait_queue, wait); + +spin_lock (&dev->ring_lock); +if (dev->ring && (READ_ONCE (dev->ring->first) !=READ_ONCE (dev->ring->last))) + mask =POLLIN |POLLRDNORM ; +spin_unlock (&dev->ring_lock); + +return mask; +} +static const struct file_operations coalesced_mmio_buffer_ops = { .mmap = coalesced_mmio_buffer_mmap, .release = coalesced_mmio_buffer_release, + .poll = coalesced_mmio_buffer_poll, };int kvm_vm_ioctl_create_coalesced_mmio_buffer (struct kvm *kvm)
@@ -240,6 +261,7 @@ int kvm_vm_ioctl_create_coalesced_mmio_buffer(struct kvm *kvm)return -ENOMEM ; dev->kvm = kvm; +init_waitqueue_head (&dev->wait_queue);spin_lock_init (&dev->ring_lock); ret = anon_inode_getfd("coalesced_mmio_buf", &coalesced_mmio_buffer_ops,
virt/kvm/coalesced_mmio.h +1 -0
@@ -26,6 +26,7 @@ struct kvm_coalesced_mmio_dev {struct kvm_coalesced_mmio_buffer_dev {struct list_head list;struct kvm *kvm; +wait_queue_head_t wait_queue;spinlock_t ring_lock;struct kvm_coalesced_mmio_ring *ring; };
Signature · 1 line
2.34.1
kernel test robot
build botRe: [PATCH v2 3/6] KVM: Support poll() on coalesced mmio buffer fds
kernel test robot noticed the following build warnings:
Hi Ilias,
kernel test robot noticed the following build warnings:
[auto build test WARNING on kvm/queue]
[also build test WARNING on next-20240719]
[cannot apply to mst-vhost/linux-next linus/master kvm/linux-next v6.10] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Ilias-Stamatis/KVM-Fix-coalesced_mmio_has_room/20240719-034316 base: https://git.kernel.org/pub/scm/virt/kvm/kvm.git queue patch link: https://lore.kernel.org/r/20240718193543.624039-4-ilstam%40amazon.com patch subject: [PATCH v2 3/6] KVM: Support poll() on coalesced mmio buffer fds config: x86_64-randconfig-122-20240719 (https://download.01.org/0day-ci/archive/20240720/[email protected]/config) compiler: clang version 18.1.5 (https://github.com/llvm/llvm-project 617a15a9eac96088ae5e9134248d8236e34b91b1) reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240720/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
sparse warnings: (new ones prefixed by >>)
>> arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: sparse: incorrect type in assignment (different base types) @@ expected restricted __poll_t [usertype] mask @@ got int @@
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: expected restricted __poll_t [usertype] mask
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: got int
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c: note: in included file (through include/linux/mmzone.h, include/linux/gfp.h, include/linux/mm.h, ...):
include/linux/page-flags.h:240:46: sparse: sparse: self-comparison always evaluates to false include/linux/page-flags.h:240:46: sparse: sparse: self-comparison always evaluates to false
vim +241 arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c
231
232 static __poll_t coalesced_mmio_buffer_poll(struct file *file, struct poll_table_struct *wait)
233 {
234 struct kvm_coalesced_mmio_buffer_dev *dev = file->private_data;
235 __poll_t mask = 0;
236
237 poll_wait(file, &dev->wait_queue, wait);
238
239 spin_lock(&dev->ring_lock);
240 if (dev->ring && (READ_ONCE(dev->ring->first) != READ_ONCE(dev->ring->last)))
> 241 mask = POLLIN | POLLRDNORM;
242 spin_unlock(&dev->ring_lock);
243
244 return mask;
245 }
246
Signature · 2 lines
0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
Re: [PATCH v2 3/6] KVM: Support poll() on coalesced mmio buffer fds
Quoted 34 lines · kernel test robot
Hi Ilias,
kernel test robot noticed the following build warnings:
[auto build test WARNING on kvm/queue]
[also build test WARNING on next-20240719]
[cannot apply to mst-vhost/linux-next linus/master kvm/linux-next v6.10] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]url: https://github.com/intel-lab-lkp/linux/commits/Ilias-Stamatis/KVM-Fix-coalesced_mmio_has_room/20240719-034316 base: https://git.kernel.org/pub/scm/virt/kvm/kvm.git queue patch link: https://lore.kernel.org/r/20240718193543.624039-4-ilstam%40amazon.com patch subject: [PATCH v2 3/6] KVM: Support poll() on coalesced mmio buffer fds config: x86_64-randconfig-122-20240719 (https://download.01.org/0day-ci/archive/20240720/[email protected]/config) compiler: clang version 18.1.5 (https://github.com/llvm/llvm-project 617a15a9eac96088ae5e9134248d8236e34b91b1) reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240720/[email protected]/reproduce)If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
Quoted 2 lines
sparse warnings: (new ones prefixed by >>)
Quoted 1 line
Quoted 1 line
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: sparse: incorrect type in assignment (different base types) @@ expected restricted__poll_t[usertype] mask @@ got int @@
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: expected restricted__poll_t[usertype] mask
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: got int
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c: note: in included file (throughinclude/linux/mmzone.h,include/linux/gfp.h,include/linux/mm.h, ...):
include/linux/page-flags.h:240:46: sparse: sparse: self-comparison always evaluates to falseinclude/linux/page-flags.h:240:46: sparse: sparse: self-comparison always evaluates to falsevim +241
arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c
Quoted 11 lines
231 232 static __poll_t coalesced_mmio_buffer_poll(struct file *file, struct poll_table_struct *wait) 233 { 234 struct kvm_coalesced_mmio_buffer_dev *dev = file->private_data; 235 __poll_t mask = 0; 236 237 poll_wait(file, &dev->wait_queue, wait); 238 239 spin_lock(&dev->ring_lock); 240 if (dev->ring && (READ_ONCE(dev->ring->first) != READ_ONCE(dev->ring->last)))Quoted 1 line
241 mask = POLLIN | POLLRDNORM;
This must be EPOLLIN | EPOLLRDNORM, but I will wait for more feedback / reviews before sending another version.
Ilias
kernel test robot
build botRe: [PATCH v2 3/6] KVM: Support poll() on coalesced mmio buffer fds
In reply to Ilias Stamatiskernel test robot noticed the following build warnings:
Hi Ilias,
kernel test robot noticed the following build warnings:
[auto build test WARNING on kvm/queue]
[also build test WARNING on next-20240719]
[cannot apply to mst-vhost/linux-next linus/master kvm/linux-next v6.10] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Ilias-Stamatis/KVM-Fix-coalesced_mmio_has_room/20240719-034316 base: https://git.kernel.org/pub/scm/virt/kvm/kvm.git queue patch link: https://lore.kernel.org/r/20240718193543.624039-4-ilstam%40amazon.com patch subject: [PATCH v2 3/6] KVM: Support poll() on coalesced mmio buffer fds config: riscv-randconfig-r121-20240719 (https://download.01.org/0day-ci/archive/20240721/[email protected]/config) compiler: clang version 14.0.6 (https://github.com/llvm/llvm-project f28c006a5895fc0e329fe15fead81e37457cb1d1) reproduce: (https://download.01.org/0day-ci/archive/20240721/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
sparse warnings: (new ones prefixed by >>)
>> arch/riscv/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: sparse: incorrect type in assignment (different base types) @@ expected restricted __poll_t [usertype] mask @@ got int @@
arch/riscv/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: expected restricted __poll_t [usertype] mask
arch/riscv/kvm/../../../virt/kvm/coalesced_mmio.c:241:22: sparse: got int
vim +241 arch/riscv/kvm/../../../virt/kvm/coalesced_mmio.c
231
232 static __poll_t coalesced_mmio_buffer_poll(struct file *file, struct poll_table_struct *wait)
233 {
234 struct kvm_coalesced_mmio_buffer_dev *dev = file->private_data;
235 __poll_t mask = 0;
236
237 poll_wait(file, &dev->wait_queue, wait);
238
239 spin_lock(&dev->ring_lock);
240 if (dev->ring && (READ_ONCE(dev->ring->first) != READ_ONCE(dev->ring->last)))
> 241 mask = POLLIN | POLLRDNORM;
242 spin_unlock(&dev->ring_lock);
243
244 return mask;
245 }
246
Signature · 2 lines
0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
PATCH v2 4/6 KVM: Add KVM_(UN)REGISTER_COALESCED_MMIO2 ioctls
In reply to Ilias StamatisAdd 2 new ioctls, KVM_REGISTER_COALESCED_MMIO2 and KVM_UNREGISTER_COALESCED_MMIO2. These do the same thing as their v1 equivalents except an fd returned by KVM_CREATE_COALESCED_MMIO_BUFFER needs to be passed as an argument to them.
The fd representing a ring buffer is associated with an MMIO region registered for coalescing and all writes to that region are accumulated there. This is in contrast to the v1 API where all regions have to share the same buffer. Nevertheless, userspace code can still use the same ring buffer for multiple zones if it wishes to do so.
Userspace can check for the availability of the new API by checking if the KVM_CAP_COALESCED_MMIO2 capability is supported.
---
v1->v2:
- Rebased on top of kvm/queue resolving the conflict in kvm.h
include/uapi/linux/kvm.h | 16 ++++++++++++++++ virt/kvm/coalesced_mmio.c | 37 +++++++++++++++++++++++++++++++------ virt/kvm/coalesced_mmio.h | 7 ++++--- virt/kvm/kvm_main.c | 36 +++++++++++++++++++++++++++++++++++- 4 files changed, 86 insertions(+), 10 deletions(-)
include/uapi/linux/kvm.h +16 -0
@@ -480,6 +480,16 @@ struct kvm_coalesced_mmio_zone { }; }; +struct kvm_coalesced_mmio_zone2 { +__u64 addr; +__u32 size; +union { +__u32 pad; +__u32 pio; + }; +int buffer_fd; +}; +struct kvm_coalesced_mmio {__u64 phys_addr;__u32 len;
@@ -933,6 +943,7 @@ struct kvm_enable_cap {#define KVM_CAP_PRE_FAULT_MEMORY 236#define KVM_CAP_X86_APIC_BUS_CYCLES_NS 237#define KVM_CAP_X86_GUEST_MODE 238 +#define KVM_CAP_COALESCED_MMIO2 239struct kvm_irq_routing_irqchip {__u32 irqchip;
@@ -1573,6 +1584,11 @@ struct kvm_pre_fault_memory { __u64 padding[5]; }; +/* Available with KVM_CAP_COALESCED_MMIO2 */ #define KVM_CREATE_COALESCED_MMIO_BUFFER _IO(KVMIO, 0xd6) +#define KVM_REGISTER_COALESCED_MMIO2 \ + _IOW(KVMIO, 0xd7, struct kvm_coalesced_mmio_zone2) +#define KVM_UNREGISTER_COALESCED_MMIO2 \ + _IOW(KVMIO, 0xd8, struct kvm_coalesced_mmio_zone2)#endif /* __LINUX_KVM_H */
virt/kvm/coalesced_mmio.c +31 -6
@@ -17,6 +17,7 @@#include <linux/kvm.h> #include <linux/anon_inodes.h> #include <linux/poll.h> +#include <linux/file.h> #include "coalesced_mmio.h"
@@ -279,19 +280,40 @@ int kvm_vm_ioctl_create_coalesced_mmio_buffer(struct kvm *kvm) }int kvm_vm_ioctl_register_coalesced_mmio (struct kvm *kvm, -struct kvm_coalesced_mmio_zone *zone) +struct kvm_coalesced_mmio_zone2 *zone, +bool use_buffer_fd) { -int ret; +int ret =0 ; +struct file *file;struct kvm_coalesced_mmio_dev *dev;struct kvm_coalesced_mmio_buffer_dev *buffer_dev =NULL ;if (zone->pio !=1 && zone->pio !=0 )return -EINVAL ; +if (use_buffer_fd) { + file =fget (zone->buffer_fd); +if (!file) +return -EBADF ; + +if (file->f_op != &coalesced_mmio_buffer_ops) { +fput (file); +return -EINVAL ; + } + + buffer_dev = file->private_data; +if (!buffer_dev->ring) { +fput (file); +return -ENOBUFS ; + } + } + dev =kzalloc (sizeof (struct kvm_coalesced_mmio_dev ),GFP_KERNEL_ACCOUNT ); -if (!dev) -return -ENOMEM ; +if (!dev) { + ret = -ENOMEM ; + goto out_free_file; + }kvm_iodevice_init (&dev->dev, &coalesced_mmio_ops); dev->kvm = kvm;
@@ -307,17 +329,20 @@ int kvm_vm_ioctl_register_coalesced_mmio(struct kvm *kvm,list_add_tail (&dev->list, &kvm->coalesced_zones);mutex_unlock (&kvm->slots_lock); -return 0 ; + goto out_free_file; out_free_dev:mutex_unlock (&kvm->slots_lock);kfree (dev); +out_free_file: +if (use_buffer_fd) +fput (file);return ret; }int kvm_vm_ioctl_unregister_coalesced_mmio (struct kvm *kvm, -struct kvm_coalesced_mmio_zone *zone) +struct kvm_coalesced_mmio_zone2 *zone) {struct kvm_coalesced_mmio_dev *dev, *tmp;int r;
virt/kvm/coalesced_mmio.h +4 -3
@@ -19,7 +19,7 @@ struct kvm_coalesced_mmio_dev {struct list_head list;struct kvm_io_device dev;struct kvm *kvm; -struct kvm_coalesced_mmio_zone zone; +struct kvm_coalesced_mmio_zone2 zone;struct kvm_coalesced_mmio_buffer_dev *buffer_dev; };
@@ -34,9 +34,10 @@ struct kvm_coalesced_mmio_buffer_dev {int kvm_coalesced_mmio_init (struct kvm *kvm);void kvm_coalesced_mmio_free (struct kvm *kvm);int kvm_vm_ioctl_register_coalesced_mmio (struct kvm *kvm, -struct kvm_coalesced_mmio_zone *zone); +struct kvm_coalesced_mmio_zone2 *zone, +bool use_buffer_fd);int kvm_vm_ioctl_unregister_coalesced_mmio (struct kvm *kvm, -struct kvm_coalesced_mmio_zone *zone); +struct kvm_coalesced_mmio_zone2 *zone);int kvm_vm_ioctl_create_coalesced_mmio_buffer (struct kvm *kvm);#else
virt/kvm/kvm_main.c +35 -1
@@ -4892,6 +4892,7 @@ static int kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)#ifdef CONFIG_KVM_MMIO case KVM_CAP_COALESCED_MMIO :return KVM_COALESCED_MMIO_PAGE_OFFSET ; +case KVM_CAP_COALESCED_MMIO2 :case KVM_CAP_COALESCED_PIO :return 1 ;#endif
@@ -5230,15 +5231,48 @@ static long kvm_vm_ioctl(struct file *filp,#ifdef CONFIG_KVM_MMIO case KVM_REGISTER_COALESCED_MMIO : {struct kvm_coalesced_mmio_zone zone; +struct kvm_coalesced_mmio_zone2 zone2; r = -EFAULT ;if (copy_from_user (&zone, argp,sizeof (zone))) goto out; - r =kvm_vm_ioctl_register_coalesced_mmio (kvm, &zone); + + zone2.addr = zone.addr; + zone2.size = zone.size; + zone2.pio = zone.pio; + zone2.buffer_fd =-1 ; + + r =kvm_vm_ioctl_register_coalesced_mmio (kvm, &zone2, false); +break ; + } +case KVM_REGISTER_COALESCED_MMIO2 : { +struct kvm_coalesced_mmio_zone2 zone; + + r = -EFAULT ; +if (copy_from_user (&zone, argp,sizeof (zone))) + goto out; + + r =kvm_vm_ioctl_register_coalesced_mmio (kvm, &zone, true);break ; }case KVM_UNREGISTER_COALESCED_MMIO : {struct kvm_coalesced_mmio_zone zone; +struct kvm_coalesced_mmio_zone2 zone2; + + r = -EFAULT ; +if (copy_from_user (&zone, argp,sizeof (zone))) + goto out; + + zone2.addr = zone.addr; + zone2.size = zone.size; + zone2.pio = zone.pio; + zone2.buffer_fd =-1 ; + + r =kvm_vm_ioctl_unregister_coalesced_mmio (kvm, &zone2); +break ; + } +case KVM_UNREGISTER_COALESCED_MMIO2 : { +struct kvm_coalesced_mmio_zone2 zone; r = -EFAULT ;if (copy_from_user (&zone, argp,sizeof (zone)))
Signature · 1 line
2.34.1
Sean Christopherson
reviewerRe: [PATCH v2 4/6] KVM: Add KVM_(UN)REGISTER_COALESCED_MMIO2 ioctls
Quoted 33 lines · Ilias Stamatis
Add 2 new ioctls,
KVM_REGISTER_COALESCED_MMIO2andKVM_UNREGISTER_COALESCED_MMIO2. These do the same thing as their v1 equivalents except an fd returned byKVM_CREATE_COALESCED_MMIO_BUFFERneeds to be passed as an argument to them.The fd representing a ring buffer is associated with an MMIO region registered for coalescing and all writes to that region are accumulated there. This is in contrast to the v1 API where all regions have to share the same buffer. Nevertheless, userspace code can still use the same ring buffer for multiple zones if it wishes to do so.
Userspace can check for the availability of the new API by checking if the
KVM_CAP_COALESCED_MMIO2capability is supported.Signed-off-by Ilias Stamatis <[email protected]>Reviewed-by Paul Durrant <[email protected]>---
v1->v2:
- Rebased on top of kvm/queue resolving the conflict inkvm.hinclude/uapi/linux/kvm.h | 16 ++++++++++++++++ virt/kvm/coalesced_mmio.c | 37 +++++++++++++++++++++++++++++++------ virt/kvm/coalesced_mmio.h | 7 ++++--- virt/kvm/kvm_main.c | 36 +++++++++++++++++++++++++++++++++++- 4 files changed, 86 insertions(+), 10 deletions(-)include/uapi/linux/kvm.h +0
-0@@ -480,6 +480,16 @@ struct kvm_coalesced_mmio_zone { };
include/uapi/linux/kvm.h struct kvm_coalesced_mmio_zone { · line 490
Quoted 10 lines
include/uapi/linux/kvm.h +8
-0}; +struct kvm_coalesced_mmio_zone2 { +__u64 addr; +__u32 size; +union { +__u32 pad; +__u32 pio; + }; +int buffer_fd;
Dunno if it matters, but KVM typically uses __u32 for file descriptors and then does the verificaton on the backend (which needs to be there regardless).
Quoted 13 lines
virt/kvm/coalesced_mmio.c +1
-0@@ -17,6 +17,7 @@#include <linux/kvm.h> #include <linux/anon_inodes.h> #include <linux/poll.h> +#include <linux/file.h> #include "coalesced_mmio.h"
virt/kvm/coalesced_mmio.c int kvm_vm_ioctl_create_coalesced_mmio_buffer(struct kvm *kvm) · line 286
Quoted 9 lines
virt/kvm/coalesced_mmio.c +3
-2@@ -279,19 +280,40 @@ int kvm_vm_ioctl_create_coalesced_mmio_buffer(struct kvm *kvm) }int kvm_vm_ioctl_register_coalesced_mmio (struct kvm *kvm, -struct kvm_coalesced_mmio_zone *zone) +struct kvm_coalesced_mmio_zone2 *zone, +bool use_buffer_fd) { -int ret; +int ret =0 ;
Heh, init ret to zero, and then return a pile of error before ret is ever touched. I assume ret can be set to '0' after mutex_unlock(&kvm->slots_lock); in the happy path, which would also show that it is indeed the happy path.
Quoted 22 lines
virt/kvm/kvm_main.c +2
-1@@ -4892,6 +4892,7 @@ static int kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)#ifdef CONFIG_KVM_MMIO case KVM_CAP_COALESCED_MMIO :return KVM_COALESCED_MMIO_PAGE_OFFSET ; +case KVM_CAP_COALESCED_MMIO2 :case KVM_CAP_COALESCED_PIO :return 1 ;#endif @@ -5230,15 +5231,48 @@ static long kvm_vm_ioctl(struct file *filp,#ifdef CONFIG_KVM_MMIO case KVM_REGISTER_COALESCED_MMIO : {struct kvm_coalesced_mmio_zone zone; +struct kvm_coalesced_mmio_zone2 zone2; r = -EFAULT ;if (copy_from_user (&zone, argp,sizeof (zone))) goto out; - r =kvm_vm_ioctl_register_coalesced_mmio (kvm, &zone);
virt/kvm/kvm_main.c static long kvm_vm_ioctl(struct file *filp, · line 5248
Quoted 10 lines
virt/kvm/kvm_main.c +10
-0+ + zone2.addr = zone.addr; + zone2.size = zone.size; + zone2.pio = zone.pio; + zone2.buffer_fd =-1 ; + + r =kvm_vm_ioctl_register_coalesced_mmio (kvm, &zone2, false); +break ; + } +case KVM_REGISTER_COALESCED_MMIO2 : {
Ah, the curly braces were just early :-)
Quoted 7 lines
virt/kvm/kvm_main.c +7
-0+struct kvm_coalesced_mmio_zone2 zone; + + r = -EFAULT ; +if (copy_from_user (&zone, argp,sizeof (zone))) + goto out; + + r =kvm_vm_ioctl_register_coalesced_mmio (kvm, &zone, true);
PATCH v2 5/6 KVM: Documentation: Document v2 of coalesced MMIO API
In reply to Ilias StamatisDocument the KVM_CREATE_COALESCED_MMIO_BUFFER andKVM_REGISTER_COALESCED_MMIO2 ioctls.
---
v1->v2:
- Added a sentence explaining that closing the fd results in deregistering the associated regions.
- Rebased on top of kvm/queue and updated the section numbers.
Documentation/virt/kvm/api.rst | 91 ++++++++++++++++++++++++++++++++++ 1 file changed, 91 insertions(+)
Documentation/virt/kvm/api.rst +91 -0
@@ -4919,6 +4919,8 @@ For the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. :Parameters: struct kvm_coalesced_mmio_zone :Returns: 0 on success, < 0 on error +KVM_(UN)REGISTER_COALESCED_MMIO2 can be used instead if available. + Coalesced I/O is a performance optimization that defers hardware register write emulation so that userspace exits are avoided. It is typically used to reduce the overhead of emulating frequently accessed
@@ -6418,6 +6420,95 @@ the capability to be present. `flags` must currently be zero. +4.144 KVM_CREATE_COALESCED_MMIO_BUFFER +------------------------------------- + +:Capability: KVM_CAP_COALESCED_MMIO2 +:Architectures: all +:Type: vm ioctl +:Parameters: none +:Returns: An fd on success, < 0 on error + +Returns an fd, but does not allocate a buffer. Also see +KVM_REGISTER_COALESCED_MMIO2. + +The fd must be first passed to mmap() to allocate a page to be used as a ring +buffer that is shared between kernel and userspace. The page must be +interpreted as a struct kvm_coalesced_mmio_ring. + +:: + + struct kvm_coalesced_mmio_ring { + __u32 first, last; + struct kvm_coalesced_mmio coalesced_mmio[]; + }; + +The kernel will increment the last index and userspace is expected to do the +same with the first index after consuming entries. The upper bound of the +coalesced_mmio array is defined as KVM_COALESCED_MMIO_MAX. + +:: + + struct kvm_coalesced_mmio { + __u64 phys_addr; + __u32 len; + union { + __u32 pad; + __u32 pio; + }; + __u8 data[8]; + }; + +After allocating a buffer with mmap(), the fd must be passed as an argument to +KVM_REGISTER_COALESCED_MMIO2 to associate an I/O region to which writes are +coalesced with the ring buffer. Multiple I/O regions can be associated with the +same ring buffer. Closing the fd after unmapping it automatically deregisters +all I/O regions associated with it. + +poll() is also supported on the fd so that userspace can be notified of I/O +writes without having to wait until the next exit to userspace. + +4.145 KVM_(UN)REGISTER_COALESCED_MMIO2 +------------------------------------- + +:Capability: KVM_CAP_COALESCED_MMIO2 +:Architectures: all +:Type: vm ioctl +:Parameters: struct kvm_coalesced_mmio_zone2 +:Returns: 0 on success, < 0 on error + +Coalesced I/O is a performance optimization that defers hardware register write +emulation so that userspace exits are avoided. It is typically used to reduce +the overhead of emulating frequently accessed hardware registers. + +When a hardware register is configured for coalesced I/O, write accesses do not +exit to userspace and their value is recorded in a ring buffer that is shared +between kernel and userspace. + +:: + + struct kvm_coalesced_mmio_zone2 { + __u64 addr; + __u32 size; + union { + __u32 pad; + __u32 pio; + }; + int buffer_fd; + }; + +KVM_CREATE_COALESCED_MMIO_BUFFER must be used to allocate a buffer fd which +must be first mmaped before passed to KVM_REGISTER_COALESCED_MMIO2, otherwise +the ioctl will fail. + +Coalesced I/O is used if one or more write accesses to a hardware register can +be deferred until a read or a write to another hardware register on the same +device. This last access will cause a vmexit and userspace will process +accesses from the ring buffer before emulating it. That will avoid exiting to +userspace on repeated writes. + +Alternatively, userspace can call poll() on the buffer fd if it wishes to be +notified of new I/O writes that way. 5. The kvm_run structure ========================
Signature · 1 line
2.34.1
PATCH v2 6/6 KVM: selftests: Add coalesced_mmio_test
In reply to Ilias StamatisTest the KVM_CREATE_COALESCED_MMIO_BUFFER, KVM_REGISTER_COALESCED_MMIO2 and KVM_UNREGISTER_COALESCED_MMIO2 ioctls.
--- tools/testing/selftests/kvm/Makefile | 1 + .../selftests/kvm/coalesced_mmio_test.c | 313 ++++++++++++++++++ 2 files changed, 314 insertions(+) create mode 100644 tools/testing/selftests/kvm/coalesced_mmio_test.c
tools/testing/selftests/kvm/Makefile +1 -0
@@ -147,6 +147,7 @@ TEST_GEN_PROGS_x86_64 += steal_time TEST_GEN_PROGS_x86_64 += kvm_binary_stats_test TEST_GEN_PROGS_x86_64 += system_counter_offset_test TEST_GEN_PROGS_x86_64 += pre_fault_memory_test +TEST_GEN_PROGS_x86_64 += coalesced_mmio_test # Compiled outputs used by test targets TEST_GEN_PROGS_EXTENDED_x86_64 += x86_64/nx_huge_pages_test
tools/testing/selftests/kvm/coalesced_mmio_test.c new file +313 -0
313 lines not on this page · raw message
Signature · 1 line
2.34.1
Sean Christopherson
reviewerQuoted 17 lines · Ilias Stamatis
The current MMIO coalescing design has a few drawbacks which limit its usefulness. Currently all coalesced MMIO zones use the same ring buffer. That means that upon a userspace exit we have to handle potentially unrelated MMIO writes synchronously. And a VM-wide lock needs to be taken in the kernel when an MMIO exit occurs.
Additionally, there is no direct way for userspace to be notified about coalesced MMIO writes. If the next MMIO exit to userspace is when the ring buffer has filled then a substantial (and unbounded) amount of time may have passed since the first coalesced MMIO.
This series adds new ioctls to KVM that allow for greater control by making it possible to associate different MMIO zones with different ring buffers. It also allows userspace to use
poll()to check for coalesced writes in order to avoid userspace exits in vCPU threads (see patch 3 for why this can be useful).
Quoted 5 lines
The idea of improving the API in this way originally came from Paul Durrant ([email protected]) but the implementation is mine.
The first patch in the series is a bug in the existing code that I discovered while writing a selftest and can be merged independently.
Ya, I'll grab it, maybe for 6.11? Doesn't seem urgent though.
Anyways, I gave this a *very* cursory review. I'd go ahead and send v3 though, you're going to need Paolo's eyeballs on this, and he's offline until September.
Sean Christopherson
reviewerQuoted 6 lines · Ilias Stamatis
The current MMIO coalescing design has a few drawbacks which limit its usefulness. Currently all coalesced MMIO zones use the same ring buffer. That means that upon a userspace exit we have to handle potentially unrelated MMIO writes synchronously. And a VM-wide lock needs to be taken in the kernel when an MMIO exit occurs.
Quoted 6 lines
Additionally, there is no direct way for userspace to be notified about coalesced MMIO writes. If the next MMIO exit to userspace is when the ring buffer has filled then a substantial (and unbounded) amount of time may have passed since the first coalesced MMIO.
[...]
Applied patch 1 to kvm-x86 generic. I deliberately didn't put this in fixes or
Cc: it for stable, as the bug has been around for sooo long without anyone noticing that there's basically zero chance that the bug is actively causing issues.
I also reworked and expanded the changelog significantly to make it more clear why things break, what the fallout is (KVM can _sometimes_ use the full ring), and to call out that the lockless scheme that the buggy commit was preparing for never seems to have landed.
Please take a gander at the changelog and holler if I messed anything up.
[1/6] KVM: Fix coalesced_mmio_has_room() to avoid premature userspace exit
https://github.com/kvm-x86/linux/commit/92f6d4130497
Quoted 24 lines · Sean Christopherson
Quoted 6 lines · Ilias Stamatis
The current MMIO coalescing design has a few drawbacks which limit its usefulness. Currently all coalesced MMIO zones use the same ring buffer. That means that upon a userspace exit we have to handle potentially unrelated MMIO writes synchronously. And a VM-wide lock needs to be taken in the kernel when an MMIO exit occurs.
Quoted 6 lines
Additionally, there is no direct way for userspace to be notified about coalesced MMIO writes. If the next MMIO exit to userspace is when the ring buffer has filled then a substantial (and unbounded) amount of time may have passed since the first coalesced MMIO.
[...]
Applied patch 1 to kvm-x86 generic. I deliberately didn't put this in fixes or
Cc: it for stable, as the bug has been around for sooo long without anyone noticing that there's basically zero chance that the bug is actively causing issues.I also reworked and expanded the changelog significantly to make it more clear why things break, what the fallout is (KVM can _sometimes_ use the full ring), and to call out that the lockless scheme that the buggy commit was preparing for never seems to have landed.
Quoted 7 lines
Please take a gander at the changelog and holler if I messed anything up.
[1/6] KVM: Fix
coalesced_mmio_has_room()to avoid premature userspace exithttps://github.com/kvm-x86/linux/commit/92f6d4130497
It looks good, thanks.
Quoted 2 lines
Doh, I applied v2 instead of v3. Though unless mine eyes deceive me, they're the same.
They are the same indeed.
I'm not entirely sure what's the recommended thing to do when updating only some of the patches of a series, whether to send new versions for the revised patches only or resend everything. I opted to do the latter and then I called out the changes between v2 and v3 inside the patches that were actually modified.
Thanks,
Ilias
Sean Christopherson
reviewerQuoted 9 lines · Ilias Stamatis
Quoted 2 lines · Sean Christopherson
Doh, I applied v2 instead of v3. Though unless mine eyes deceive me, they're the same.
They are the same indeed.
I'm not entirely sure what's the recommended thing to do when updating only some of the patches of a series, whether to send new versions for the revised patches only or resend everything. I opted to do the latter
Resend everything, as you did. I simply messed up and grabbed the wrong version.
Quoted 2 lines
and then I called out the changes between v2 and v3 inside the patches that were actually modified.
FWIW, I strongly prefer the delta be captured in the cover letter, though others may obviously have different preferences.
https://lore.kernel.org/all/ZQnAN9TC6b8mSJ%[email protected]