Thank you very much. Adding kaslr successfully solved the problem.
Ilya Leoshkevich 于2024年9月24日周二 00:34写道：
> On Mon, 2024-09-23 at 22:36 +0800, jie ren wrote:
> > Hi, I have a question for help
> > I recently Using ubuntu24.4 system qemu-system-aarch64 +gdb-
> > m
Hi, I have a question for help
I recently Using ubuntu24.4 system qemu-system-aarch64 +gdb-multiarch
to debug the kernel, setting breakpoints cannot be stopped.
system information:
ubuntu version: 22.04
Debug kernel version: 5.0.0
qemu version:
· qemu-system-a
0x7fca423e8ecc in kvm_handle_io
28 0x7fca423ecb48 in kvm_cpu_exec
29 0x7fca424279d5 in qemu_kvm_cpu_thread_fn
30 0x7fca423c9480 in qemu_thread_start
31 0x7fca4257ff3b in ?? () from /usr/lib64/libpthread.so.0
32 0x7fca4270b550 in clone () from /usr/lib64/libc.so.6
Signed-off-by: Ji
`section` passed by the caller to determine the
pages that need to be write-protected. Compared to the original method,
this saves time.
Moreover, this makes only pages whose dirty-bits
are cleared write-protected instead of making the whole memory slot
write-protected.
Signed-off-by: Yan-Jie Wang
work properly.
Buglink: https://bugs.launchpad.net/qemu/+bug/1827005
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 62
include/sysemu/hvf_int.h | 14 +
target/arm/hvf/hvf.c | 5
target/i386/hvf/hvf.c| 25
We follow how KVM accel does in its memory listener (kvm-all.c) and add
a lock for the memory related functions.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 18 ++
1 file changed, 18 insertions(+)
diff --git a/accel/hvf/hvf-mem.c b/accel/hvf/hvf-mem.c
index 896e718374
It is no need to dirty-track MMIO regions or other readonly regions.
Before we start or stop to dirty-track a memory region, check the type of
the memory region. The region must be a writeable ram to be dirty-tracked.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 6 ++
1 file
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-accel-ops.c | 220 +
accel/hvf/hvf-mem.c | 252 ++
accel/hvf/meson.build | 1 +
include/sysemu/hvf_int.h | 2 +
4 files changed, 256 insertions(+), 219 deletions
Follow the QEMU coding style. Structured type names are in CamelCase.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 14 +++---
include/sysemu/hvf_int.h | 8
target/i386/hvf/hvf.c| 4 ++--
3 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/accel
Follow the QEMU coding style. Use hwaddr for guest physical address.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 2 +-
include/sysemu/hvf_int.h | 8
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/accel/hvf/hvf-mem.c b/accel/hvf/hvf-mem.c
index 32452696b6
`section` passed by the caller to determine the
pages that need to be write-protected. Compared to the original method,
this saves time.
Moreover, this makes only pages whose dirty-bits
are cleared write-protected instead of making the whole memory slot
write-protected.
Signed-off-by: Yan-Jie Wang
work properly.
Buglink: https://bugs.launchpad.net/qemu/+bug/1827005
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 62
include/sysemu/hvf_int.h | 14 +
target/arm/hvf/hvf.c | 5
target/i386/hvf/hvf.c| 25
es instead of arrays to store
in-kernel KVM memory slots.
5. Add a lock to protect the data structures of HVF memory slots
Patches have been tested on Apple Silicon Macs and Intel Macs.
Yan-Jie Wang (9):
hvf: move memory related functions from hvf-accel-ops.c to hvf-mem.c
hvf: sim
Follow the QEMU coding style. Use hwaddr for guest physical address.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 2 +-
include/sysemu/hvf_int.h | 8
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/accel/hvf/hvf-mem.c b/accel/hvf/hvf-mem.c
index 32452696b6
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-accel-ops.c | 220 +
accel/hvf/hvf-mem.c | 252 ++
accel/hvf/meson.build | 1 +
include/sysemu/hvf_int.h | 2 +
4 files changed, 256 insertions(+), 219 deletions
Currently, there are only 32 memory slots in the fixed size array.
It is not scalable. Instead of using fixed size array, use GTree
(from glib library) and dynamically-allocated structures to store
memory slots.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 63
We follow how KVM accel does in its memory listener (kvm-all.c) and add
a lock for the memory related functions.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 18 ++
1 file changed, 18 insertions(+)
diff --git a/accel/hvf/hvf-mem.c b/accel/hvf/hvf-mem.c
index 896e718374
Follow the QEMU coding style. Structured type names are in CamelCase.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 14 +++---
include/sysemu/hvf_int.h | 8
target/i386/hvf/hvf.c| 4 ++--
3 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/accel
Currently, there are only 32 memory slots in the fixed size array.
It is not scalable. Instead of using fixed size array, use GTree
(from glib library) and dynamically-allocated structures to store
memory slots.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 63
not overlap with existing sections.
- Try to align memory sections first before giving up sections that are not
aligned to host page size.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-accel-ops.c | 1 -
accel/hvf/hvf-mem.c | 211 +++---
include/sysemu
not overlap with existing sections.
- Try to align memory sections first before giving up sections that are not
aligned to host page size.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-accel-ops.c | 1 -
accel/hvf/hvf-mem.c | 211 +++---
include/sysemu
t the data structures of HVF memory slots
Patches have been tested on Apple Silicon Macs and Intel Macs.
Yan-Jie Wang (9):
hvf: move memory related functions from hvf-accel-ops.c to hvf-mem.c
hvf: simplify data structures and codes of memory related functions
hvf: use correct data types for
It is no need to dirty-track MMIO regions or other readonly regions.
Before we start or stop to dirty-track a memory region, check the type of
the memory region. The region must be a writeable ram to be dirty-tracked.
Signed-off-by: Yan-Jie Wang
---
accel/hvf/hvf-mem.c | 5 +
1 file
Sorry, I made a mistake.
The last line in the function, hvf_find_free_slot, introduced in this
commit "hvf: simplify data structures and codes of memory related
functions" should be "return NULL;"
static hvf_slot *hvf_find_free_slot(void)
{
hvf_slot *slot;
int x;
for (x = 0; x < HVF_NU
For the dirty-tracking part in my patch series, the major difference
between this patch and Alex's patch is that the dirty-tracking logic in
my patch will only mark the page being written dirty instead of marking
the whole memory slot dirty, and will only restore the write permission
to the pag
s add req_has_ua to avoid this kind of thing.
Signed-off-by: suruifeng
Signed-off-by: Jie Wang
---
hw/scsi/scsi-bus.c | 10 --
include/hw/scsi/scsi.h | 1 +
2 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/hw/scsi/scsi-bus.c b/hw/scsi/scsi-bus.c
index 2a0a98cac9..20ec4
if execute qemu_open success, have no branch to free the fd,
so unlink it inadvance, let it free by process exit.
Signed-off-by: Jie Wang
---
util/oslib-posix.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/util/oslib-posix.c b/util/oslib-posix.c
index 36820fec16..fa881f2ee8 100644
--- a
if execute qemu_open success, have no branch to free the fd,
so unlink it inadvance, let it free by process exit.
Signed-off-by: Jie Wang
---
util/oslib-posix.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/util/oslib-posix.c b/util/oslib-posix.c
index 36820fec16..30bf39bf4f 100644
--- a
On 2021/4/1 20:12, Viresh Kumar wrote:
+
+/* vhost-user-i2c definitions */
+
+#define MAX_I2C_VDEV(1 << 7)
+#define MAX_I2C_ADAPTER 16
Generally speaking, 16 is big enough for most cases. But comparing with
static configuration,
I think it is better if we
On 2021/3/24 15:33, Viresh Kumar wrote:
+static int vi2c_parse(VuI2c *i2c)
+{
+uint16_t client_addr[MAX_I2C_VDEV];
+int32_t n_adapter = 0, n_client;
+int64_t addr, bus;
+const char *cp, *t;
+
+while (device_list) {
+/* Read :[:] entries one by one */
+cp = st
On 2021/3/24 15:33, Viresh Kumar wrote:
+
+/* Definitions from virtio-i2c specifications */
+#define VHOST_USER_I2C_MAX_QUEUES 1
+
+/* Status */
+#define VIRTIO_I2C_MSG_OK 0
+#define VIRTIO_I2C_MSG_ERR 1
+
+/* The bit 0 of the @virtio_i2c_out_hdr.@flags, used to
On 2019/5/2 0:38, Stefan Hajnoczi wrote:
> On Tue, Apr 30, 2019 at 02:10:14PM +0800, Jie Wang wrote:
>> remove the dead code
>>
>> Signed-off-by: Jie Wang
>> ---
>> hw/virtio/vhost.c | 1 -
>> 1 file changed, 1 deletion(-)
> Reviewed-by: Stefan Hajnoczi
On 2019/5/2 0:40, Stefan Hajnoczi wrote:
> On Tue, Apr 30, 2019 at 03:15:00PM +0800, Jie Wang wrote:
>> fix memory leak in vhost_user_scsi_realize
>>
>> Signed-off-by: Jie Wang
>> ---
>> hw/scsi/vhost-user-scsi.c | 3 +++
>> 1 file changed, 3 insertions(+)
> Reviewed-by: Stefan Hajnoczi
wrote:
> On 5/28/19 7:45 PM, Paolo Bonzini wrote:
>> On 28/05/19 15:06, Jie Wang wrote:
>>> if pr-helper been killed and qemu send disconnect event to libvirt
>>> and libvirt started a new pr-helper process, the new pr-heleper
>>> been killed again when qemu is c
-off-by: Jie Wang
---
scsi/pr-manager-helper.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/scsi/pr-manager-helper.c b/scsi/pr-manager-helper.c
index 438380fced..b7341b8f47 100644
--- a/scsi/pr-manager-helper.c
+++ b/scsi/pr-manager-helper.c
@@ -120,6 +120,7 @@ static int
fix memory leak in vhost_user_scsi_realize
Signed-off-by: Jie Wang
---
hw/scsi/vhost-user-scsi.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/hw/scsi/vhost-user-scsi.c b/hw/scsi/vhost-user-scsi.c
index 8b1e6876db..a9fd8ea305 100644
--- a/hw/scsi/vhost-user-scsi.c
+++ b/hw/scsi/vhost
fix incorrect print type in vhost_virtqueue_stop
Signed-off-by: Jie Wang
---
hw/virtio/vhost.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index 7f61018f2a..286bb27c65 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -1081,7
remove the dead code
Signed-off-by: Jie Wang
---
hw/virtio/vhost.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index 7f61018f2a..2303a8c206 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -1650,7 +1650,6 @@ fail_vq
Does qemu-pr-helper have its own log files？
as input may create
corrupted images in macOS especially for large files, and qemu-img
map may also report wrong things. This patch fixes this undesired
behaviors.
Signed-off-by: Yan-Jie Wang
---
block/file-posix.c | 59 ++
1 file changed, 59 insertions
I have done some experiments and find out that
the behavior of lseek with whence set to SEEK_DATA is different from the
behavior of Linux's lseek.
If the supplied offset is in the middle of a data region, it returns the
start of the next data region. There may be many data regions in a big
file
I also met this bug
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1779120
Title:
disk missing in the guest contingently when hotplug several virtio
scsi disks consecutively
Status in QEMU:
New
if laio_init create linux_aio failed and return NULL, NULL pointer
dereference will occur when laio_attach_aio_context dereference
linux_aio in aio_get_linux_aio. Let's avoid it and report error.
Signed-off-by: Jie Wang
---
block/file-posix.c | 19 +--
util/async.c
if laio_init create linux_aio failed and return NULL, NULL pointer
dereference will occur when laio_attach_aio_context dereference
linux_aio in aio_get_linux_aio. Let's avoid it and report error.
Signed-off-by: Jie Wang
---
block/file-posix.c | 19 +--
util/async.c
if laio_init create linux_aio failed and return NULL, NULL pointer
dereference will occur when laio_attach_aio_context dereference
linux_aio in aio_get_linux_aio, so add assert to avoid it.
Signed-off-by: Jie Wang
---
util/async.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/util/async.c
epoll_available will only be set if epollfd != -1, os we
can swap the two variables in aio_epoll_disable, and
aio_context_destroy can call aio_epoll_disable directly.
Signed-off-by: Jie Wang
---
util/aio-posix.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/util
When we call addIOThread, the epollfd created in aio_context_setup,
but not close it in the process of delIOThread, so the epollfd will leak.
Signed-off-by: Jie Wang
---
include/block/aio.h | 8
util/aio-posix.c| 9 +
util/aio-win32.c| 4
util/async.c| 1
When we call addIOThread, the epollfd created in aio_context_setup,
but not close it in the process of delIOThread, so the epollfd will leak.
Signed-off-by: Jie Wang
---
include/block/aio.h | 8
util/aio-posix.c| 13 ++---
util/aio-win32.c| 4
util/async.c
When we call addIOThread, the epollfd created in aio_context_setup,
but not close it in the process of delIOThread, so the epollfd will leak.
Signed-off-by: Jie Wang
---
include/block/aio.h | 8
util/aio-posix.c| 9 +
util/aio-win32.c| 4
util/async.c| 1
From: w00251574
When we call addIOThread, the epollfd created in aio_context_setup,
but not close it in the process of delIOThread, so the epollfd will leak.
Signed-off-by: Jie Wang
---
include/block/aio.h | 8
util/aio-posix.c| 9 +
util/aio-win32.c| 4
util
From: w00251574
When we call addIOThread, the epollfd created in aio_context_setup,
but not close it in the process of delIOThread, so the epollfd will leak.
Signed-off-by: Jie Wang
---
include/block/aio.h | 8
util/aio-posix.c| 7 +++
util/aio-win32.c| 4
util
When we call addIOThread, the epollfd created in aio_context_setup,
but not close it in the process of delIOThread, so the epollfd will leak.
Signed-off-by: Jie Wang
---
iothread.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/iothread.c b/iothread.c
index aff1281..23ac0a8 100644
--- a
Public bug reported:
Hi all, we systematically tested the QEMU implementation for emulating
arm user mode programs. We found that QEMU incorrectly emulate the FPSCR
register. The following the proof of code:
/*** Beginning of the bug: arm.c **/
int printf(const char *format, ...)
Public bug reported:
Hi all, we found 28 differently encoded illegal SSE3 instructions
reporting on the most recent x86_64 user mode linux qemu (version
2.7.0). We believe these reporting should be incorrect because the same
code can be executed on a real machine. The instructions are the
followin
Signed-off-by: Jie Wang
---
tests/Makefile.include| 3 ++
tests/drive-mirror-test.c | 96 +++
2 files changed, 99 insertions(+)
create mode 100755 tests/drive-mirror-test.c
diff --git a/tests/Makefile.include b/tests/Makefile.include
index
Signed-off-by: Jie Wang
---
tests/Makefile.include| 3 ++
tests/drive-mirror-test.c | 96 +++
2 files changed, 99 insertions(+)
create mode 100755 tests/drive-mirror-test.c
diff --git a/tests/Makefile.include b/tests/Makefile.include
index
ping
On 2016/8/15 10:53, Zhou Jie wrote:
ping
On 2016/8/2 11:57, Zhou Jie wrote:
During aer err occurs and resume do following to
protect device from being accessed.
1. Make config space read only.
2. Disable INTx/MSI Interrupt.
3. Do nothing for bar regions.
Signed-off-by: Zhou Jie
---
v2
ping
On 2016/8/2 11:57, Zhou Jie wrote:
During aer err occurs and resume do following to
protect device from being accessed.
1. Make config space read only.
2. Disable INTx/MSI Interrupt.
3. Do nothing for bar regions.
Signed-off-by: Zhou Jie
---
v2-v3:
1. Call init_completion() in
During aer err occurs and resume do following to
protect device from being accessed.
1. Make config space read only.
2. Disable INTx/MSI Interrupt.
3. Do nothing for bar regions.
Signed-off-by: Zhou Jie
---
v2-v3:
1. Call init_completion() in vfio_pci_probe.
2. Call reinit_completion() in
Hi, Alex
Clearly this has only been tested for a single instance of an AER error
event and resume per device. Are the things you're intending to block
actually blocked for subsequent events? Note how complete_all() fills
the done field to let all current and future waiters go through and
nowhe
Hi, Alex
On 2016/7/30 1:12, Alex Williamson wrote:
On Tue, 19 Jul 2016 15:32:43 +0800
Zhou Jie wrote:
From: Chen Fan
During aer err occurs and resume do following to
protect device from being accessed.
1. Make config space read only.
2. Disable INTx/MSI Interrupt.
3. Do nothing for bar
Hi, Alex
On 2016/7/30 1:11, Alex Williamson wrote:
On Tue, 19 Jul 2016 15:52:45 +0800
Zhou Jie wrote:
From: Chen Fan
An empty commit log is unacceptable for all but the most trivial
patches.
There's also no sign-off on this patch.
Sorry. I should note it.
I also don't know w
ping
On 2016/7/19 16:13, Zhou Jie wrote:
From: Chen Fan
v1-v2:
1. Add aer process to vfio driver.
Chen Fan (2):
vfio : add aer process
vfio : resume notifier
drivers/vfio/pci/vfio_pci.c | 58 -
drivers/vfio/pci/vfio_pci_intrs.c | 18
From: Chen Fan
v1-v2:
1. Add aer process to vfio driver.
Chen Fan (2):
vfio : add aer process
vfio : resume notifier
drivers/vfio/pci/vfio_pci.c | 58 -
drivers/vfio/pci/vfio_pci_intrs.c | 18
drivers/vfio/pci/vfio_pci_private.h
From: Chen Fan
---
drivers/vfio/pci/vfio_pci.c | 28 +++-
drivers/vfio/pci/vfio_pci_intrs.c | 18 ++
drivers/vfio/pci/vfio_pci_private.h | 1 +
include/uapi/linux/vfio.h | 1 +
4 files changed, 47 insertions(+), 1 deletion(-)
diff --
prevent it from further
interacting with the VM.
4. Reset bus.
Signed-off-by: Chen Fan
Signed-off-by: Zhou Jie
---
hw/vfio/pci.c | 51 +-
hw/vfio/pci.h | 1 +
linux-headers/linux/vfio.h | 4
3 files changed, 55
From: Chen Fan
Signed-off-by: Chen Fan
---
hw/vfio/pci.c | 21 +
1 file changed, 17 insertions(+), 4 deletions(-)
diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
index 11c895c..21fd801 100644
--- a/hw/vfio/pci.c
+++ b/hw/vfio/pci.c
@@ -2060,14 +2060,27 @@ static void vfio_pci_po
From: Chen Fan
when the vfio device encounters an uncorrectable error in host,
the vfio_pci driver will signal the eventfd registered by this
vfio device, resulting in the qemu eventfd handler getting
invoked.
this patch is to pass the error to guest and let the guest driver
recover from the err
From: Chen Fan
add 'aer' property to let user able to decide whether expose
the aer capability. by default we should disable aer feature,
because it needs configuration restrictions.
Signed-off-by: Chen Fan
---
hw/vfio/pci.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/hw/vfio/pci.c b
From: Chen Fan
Calling pcie_aer_init to initilize aer related registers for
vfio device, then reload physical related registers to expose
device capability.
Signed-off-by: Chen Fan
---
hw/vfio/pci.c | 75 ++-
hw/vfio/pci.h | 3 +++
2 fil
From: Chen Fan
When assigning a vfio device with AER enabled, we must check whether
the device supports a host bus reset (ie. hot reset) as this may be
used by the guest OS in order to recover the device from an AER
error. QEMU must therefore have the ability to perform a physical
host bus reset
From: Chen Fan
Due to all devices assigned to VM on the same way as host if enable
aer, so we can easily do the hot reset by selecting the function #0
to do the hot reset.
Signed-off-by: Chen Fan
---
hw/vfio/pci.c | 14 ++
hw/vfio/pci.h | 1 +
2 files changed, 15 insertions(+)
di
From: Chen Fan
when function 0 is hot-added, we can check the vfio device
whether support hot bus reset.
Signed-off-by: Chen Fan
---
hw/vfio/pci.c | 14 ++
1 file changed, 14 insertions(+)
diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
index 242c1e4..8bcb26b 100644
--- a/hw/vfio/pci.c
From: Chen Fan
v8-v9:
1. Don't use resume notification.
Host vfio driver will process aer.
Poll the vfio_device_info to assure host reset completely.
v7-v8:
1. Use bitmap to record error and resume notification.
v6-v7:
1. Stall any access to the device until resume is signa
From: Chen Fan
Squeeze out vfio_pci_do_hot_reset to do host bus reset when AER recovery.
Signed-off-by: Chen Fan
---
hw/vfio/pci.c | 75 +++
1 file changed, 44 insertions(+), 31 deletions(-)
diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
ind
From: Chen Fan
PCI hotplug requires that function 0 is added last to close the
slot. Since vfio supporting AER, we require that the VM bus
contains the same set of devices as the host bus to support AER,
we can perform an AER validation test whenever a function 0 in
the VM is hot-added.
Signed-
From: Chen Fan
The function is used to get affected devices by bus reset.
So here extract it, and can used for aer soon.
Signed-off-by: Chen Fan
---
hw/vfio/pci.c | 66 +++
1 file changed, 48 insertions(+), 18 deletions(-)
diff --git a/h
From: Chen Fan
During aer err occurs and resume do following to
protect device from being accessed.
1. Make config space read only.
2. Disable INTx/MSI Interrupt.
3. Do nothing for bar regions.
Signed-off-by: Zhou Jie
---
drivers/vfio/pci/vfio_pci.c | 30
From: root
---
drivers/vfio/pci/vfio_pci.c | 28 +++-
drivers/vfio/pci/vfio_pci_intrs.c | 18 ++
drivers/vfio/pci/vfio_pci_private.h | 1 +
include/uapi/linux/vfio.h | 1 +
4 files changed, 47 insertions(+), 1 deletion(-)
diff --git
From: Chen Fan
v1-v2:
1. Add aer process to vfio driver.
Chen Fan (1):
vfio : add aer process
root (1):
vfio : resume notifier
drivers/vfio/pci/vfio_pci.c | 58 -
drivers/vfio/pci/vfio_pci_intrs.c | 18
drivers/vfio/pci/vfio_pc
e to true in "struct vfio_pci_device"
6. vfio_fops
Add poll function
I would still suggest that the name "workable_state" is quite vague.
Something like aer_error_in_progress is much more specific. Thanks,
OK, I will alter the name.
Sincerely
Zhou Jie
to false in "struct vfio_pci_device"
teardown the interrupt
5. vfio_pci_aer_resume
Set workable_state to true in "struct vfio_pci_device"
6. vfio_fops
Add poll function
Sincerely
Zhou Jie
_slot_reset,
.resume = ixgbevf_io_resume,
};
User tear-down the interrupts in ixgbevf_io_error_detected function.
And up the interrupts in ixgbevf_io_resume.
Guest OS driver will do both tear-down the interrupts
and re-establish them.
Because it don't know what host vfio driver has done.
I disable the interrupts to pretend them interfere with device reset.
Sincerely
Zhou Jie
Hi Alex,
The variable clearly isn't visible to the user, so the user can know
whether the kernel supports this feature, but not whether the feature
is currently active. Perhaps there's no way to avoid races completely,
but don't you expect that if we define that certain operations are
blocked a
he guest will do.
What I want to do in vfio driver is as following.
1. aer occurs
2. Disable INTx and MSI
3. aer driver reset the device
4. Restore INTx and MSI
5. user process the aer event
Sincerely
Zhou Jie
user
5. user process the aer event
Maybe reset the device and reinitialization
What I do is make sure the following points.
1. Host can reset the device between step 2 and 4.
2. The user settings is the same at step 1 and 5.
Sincerely
Zhou Jie
ping
On 2016/7/3 12:00, Zhou Jie wrote:
Hi Alex,
On 2016/6/30 9:45, Zhou Jie wrote:
Hi Alex,
On 2016/6/30 2:22, Alex Williamson wrote:
On Wed, 29 Jun 2016 16:54:05 +0800
Zhou Jie wrote:
Hi Alex,
And yet we have struct pci_dev.broken_intx_masking and we test for
working DisINTx via
Hi Alex,
On 2016/6/30 9:45, Zhou Jie wrote:
Hi Alex,
On 2016/6/30 2:22, Alex Williamson wrote:
On Wed, 29 Jun 2016 16:54:05 +0800
Zhou Jie wrote:
Hi Alex,
And yet we have struct pci_dev.broken_intx_masking and we test for
working DisINTx via pci_intx_mask_supported() rather than simply
Hi Alex,
On 2016/6/30 2:22, Alex Williamson wrote:
On Wed, 29 Jun 2016 16:54:05 +0800
Zhou Jie wrote:
Hi Alex,
And yet we have struct pci_dev.broken_intx_masking and we test for
working DisINTx via pci_intx_mask_supported() rather than simply
looking for a PCIe device. Some devices are
cause write configure space maybe happened in interrupt handler.
I think block is not a good choice.
Sincerely
Zhou Jie
Hi Alex,
On 2016/6/28 11:58, Alex Williamson wrote:
On Tue, 28 Jun 2016 11:26:33 +0800
Zhou Jie wrote:
Hi Alex,
The INTx/MSI part needs further definition for the user. Are we
actually completely tearing down interrupts with the expectation that
the user will re-enable them or are we just
ms like a reasonable approach, but I can't guarantee we
won't find new issues along the way. For instance we'll need to test
how -EAGAIN returns interact with existing QEMU and maybe decided
whether there are cases that are better handled by doing an
interruptible wait. Thanks,
I will dig into it.
Sincerely
Zhou Jie
Hi Alex,
We should never depend on the guest driver to behave in a certain way,
but we need to prioritize what that actually means. vfio in the kernel
has a responsibility first and foremost to the host kernel. User owned
devices cannot be allowed to exploit or interfere with the host
regardle
Hi Alex,
On 2016/6/22 13:45, Zhou Jie wrote:
Hi Alex,
In vfio I have some questions.
1. How can I disable the access by mmap?
We can disable all access to vfio fd by returning a EAGAIN error
if user try to access it during the reset period until the host
reset finished.
But
Hi Alex,
In vfio I have some questions.
1. How can I disable the access by mmap?
We can disable all access to vfio fd by returning a EAGAIN error
if user try to access it during the reset period until the host
reset finished.
But about the bar region which is maped by vfio_pci_m
Hi Alex,
Hi Alex,
on kernel side, I think if we don't trust the user behaviors, we
should
disable the access of vfio-pci interface once vfio-pci driver got the
error_detected,
we should disable all access to vfio fd regardless whether the vfio-pci
was assigned to a VM, we also can re
only make sure the disabled device
will not interact with the VM.
Sincerely
Zhou jie
Alex
.
--
周潔
Dept 1
No. 6 Wenzhu Road,
Nanjing, 210012, China
TEL：+86+25-86630566-8557
FUJITSU INTERNAL：7998-8557
E-Mail：zhoujie2...@cn.fujitsu.com
ping
On 2016/6/12 10:38, Zhou Jie wrote:
Hi, Alex
It seems like we have a number of questions open in the thread with MST
from the previous version, particularly whether we should actually drop
the resume notifier and block the reset in the kernel. The concern
being that it's not very
Hi, Alex
On 2016/6/9 23:39, Alexander Duyck wrote:
On Thu, Jun 9, 2016 at 3:14 AM, Zhou Jie wrote:
TO Alex
TO Michael
In your solution you add a emulate PCI bridge to act as
a bridge between direct assigned devices and the host bridge.
Do you mean put all direct assigned devices to
ing resume notification?
Sincerely
Zhou Jie
1 - 100 of 143 matches
Mail list logo
