In acrn_dev_ioctl(), vm_param is not released or passed out on the
error path of "if ((vm_param->reserved0 | vm_param->reserved1) != 0)",
which could lead to a memory leak.
Fix this bug by adding a kfree of vm_param on the error path.
This bug was found by a static analyzer.
Builds with CONFIG_ACRN_GUEST=y, CONFIG_ACRN_HSM=y show no new warnings,
and our static analyzer no longer warns about this code.
Fixes: 9c5137aedd11 (“9c5137aedd11 virt: acrn: Introduce VM management interfaces”)
Signed-off-by: Zhou Qingyang <[email protected]>
---
The analysis employs differential checking to identify inconsistent
security operations (e.g., checks or kfrees) between two code paths
and confirms that the inconsistent operations are not recovered in
the current function or the callers, so they constitute bugs.
Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.
drivers/virt/acrn/hsm.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/virt/acrn/hsm.c b/drivers/virt/acrn/hsm.c
index 5419794fccf1..205f4c637556 100644
--- a/drivers/virt/acrn/hsm.c
+++ b/drivers/virt/acrn/hsm.c
@@ -136,9 +136,11 @@ static long acrn_dev_ioctl(struct file *filp, unsigned int cmd,
if (IS_ERR(vm_param))
return PTR_ERR(vm_param);
- if ((vm_param->reserved0 | vm_param->reserved1) != 0)
- return -EINVAL;
-
+ if ((vm_param->reserved0 | vm_param->reserved1) != 0) {
+ ret = -EINVAL;
+ kfree(vm_param);
+ break;
+ }
vm = acrn_vm_create(vm, vm_param);
if (!vm) {
ret = -EINVAL;
--
2.25.1
On Tue, Jan 04, 2022 at 10:34:39AM +0800, Zhou Qingyang wrote:
> In acrn_dev_ioctl(), vm_param is not released or passed out on the
> error path of "if ((vm_param->reserved0 | vm_param->reserved1) != 0)",
> which could lead to a memory leak.
>
> Fix this bug by adding a kfree of vm_param on the error path.
>
> This bug was found by a static analyzer.
>
> Builds with CONFIG_ACRN_GUEST=y, CONFIG_ACRN_HSM=y show no new warnings,
> and our static analyzer no longer warns about this code.
>
> Fixes: 9c5137aedd11 (“9c5137aedd11 virt: acrn: Introduce VM management interfaces”)
> Signed-off-by: Zhou Qingyang <[email protected]>
> ---
> The analysis employs differential checking to identify inconsistent
> security operations (e.g., checks or kfrees) between two code paths
> and confirms that the inconsistent operations are not recovered in
> the current function or the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>
Hi Qingyang
Thanks a lot to fix this issue. Would you please to help to fix the same issue
in ACRN_IOCTL_SET_VCPU_REGS case ?
> drivers/virt/acrn/hsm.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/virt/acrn/hsm.c b/drivers/virt/acrn/hsm.c
> index 5419794fccf1..205f4c637556 100644
> --- a/drivers/virt/acrn/hsm.c
> +++ b/drivers/virt/acrn/hsm.c
> @@ -136,9 +136,11 @@ static long acrn_dev_ioctl(struct file *filp, unsigned int cmd,
> if (IS_ERR(vm_param))
> return PTR_ERR(vm_param);
>
> - if ((vm_param->reserved0 | vm_param->reserved1) != 0)
> - return -EINVAL;
> -
> + if ((vm_param->reserved0 | vm_param->reserved1) != 0) {
> + ret = -EINVAL;
> + kfree(vm_param);
> + break;
> + }
> vm = acrn_vm_create(vm, vm_param);
> if (!vm) {
> ret = -EINVAL;
> --
> 2.25.1
>
In acrn_dev_ioctl(), vm_param is not released or passed out on the
error path of "if ((vm_param->reserved0 | vm_param->reserved1) != 0)",
which could lead to a memory leak.
Fix this bug by adding a kfree of vm_param on the error path.
This bug was found by a static analyzer.
Builds with CONFIG_ACRN_GUEST=y, CONFIG_ACRN_HSM=y show no new warnings,
and our static analyzer no longer warns about this code.
Fixes: 9c5137aedd11 (“9c5137aedd11 virt: acrn: Introduce VM management interfaces”)
Signed-off-by: Zhou Qingyang <[email protected]>
---
The analysis employs differential checking to identify inconsistent
security operations (e.g., checks or kfrees) between two code paths
and confirms that the inconsistent operations are not recovered in
the current function or the callers, so they constitute bugs.
Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.
Changes in v2:
- Fix the same bug in ACRN_IOCTL_SET_VCPU_REGS.
drivers/virt/acrn/hsm.c | 24 +++++++++++++++++-------
1 file changed, 17 insertions(+), 7 deletions(-)
diff --git a/drivers/virt/acrn/hsm.c b/drivers/virt/acrn/hsm.c
index 5419794fccf1..7a8f722f431c 100644
--- a/drivers/virt/acrn/hsm.c
+++ b/drivers/virt/acrn/hsm.c
@@ -136,9 +136,11 @@ static long acrn_dev_ioctl(struct file *filp, unsigned int cmd,
if (IS_ERR(vm_param))
return PTR_ERR(vm_param);
- if ((vm_param->reserved0 | vm_param->reserved1) != 0)
- return -EINVAL;
-
+ if ((vm_param->reserved0 | vm_param->reserved1) != 0) {
+ ret = -EINVAL;
+ kfree(vm_param);
+ break;
+ }
vm = acrn_vm_create(vm, vm_param);
if (!vm) {
ret = -EINVAL;
@@ -182,21 +184,29 @@ static long acrn_dev_ioctl(struct file *filp, unsigned int cmd,
return PTR_ERR(cpu_regs);
for (i = 0; i < ARRAY_SIZE(cpu_regs->reserved); i++)
- if (cpu_regs->reserved[i])
+ if (cpu_regs->reserved[i]) {
+ kfree(cpu_regs);
return -EINVAL;
+ }
for (i = 0; i < ARRAY_SIZE(cpu_regs->vcpu_regs.reserved_32); i++)
- if (cpu_regs->vcpu_regs.reserved_32[i])
+ if (cpu_regs->vcpu_regs.reserved_32[i]) {
+ kfree(cpu_regs);
return -EINVAL;
+ }
for (i = 0; i < ARRAY_SIZE(cpu_regs->vcpu_regs.reserved_64); i++)
- if (cpu_regs->vcpu_regs.reserved_64[i])
+ if (cpu_regs->vcpu_regs.reserved_64[i]) {
+ kfree(cpu_regs);
return -EINVAL;
+ }
for (i = 0; i < ARRAY_SIZE(cpu_regs->vcpu_regs.gdt.reserved); i++)
if (cpu_regs->vcpu_regs.gdt.reserved[i] |
- cpu_regs->vcpu_regs.idt.reserved[i])
+ cpu_regs->vcpu_regs.idt.reserved[i]) {
+ kfree(cpu_regs);
return -EINVAL;
+ }
ret = hcall_set_vcpu_regs(vm->vmid, virt_to_phys(cpu_regs));
if (ret < 0)
--
2.25.1
On Tue, Jan 04, 2022 at 11:45:01AM +0800, Zhou Qingyang wrote:
> In acrn_dev_ioctl(), vm_param is not released or passed out on the
> error path of "if ((vm_param->reserved0 | vm_param->reserved1) != 0)",
> which could lead to a memory leak.
>
> Fix this bug by adding a kfree of vm_param on the error path.
>
> This bug was found by a static analyzer.
>
> Builds with CONFIG_ACRN_GUEST=y, CONFIG_ACRN_HSM=y show no new warnings,
> and our static analyzer no longer warns about this code.
>
> Fixes: 9c5137aedd11 (“9c5137aedd11 virt: acrn: Introduce VM management interfaces”)
> Signed-off-by: Zhou Qingyang <[email protected]>
> ---
> The analysis employs differential checking to identify inconsistent
> security operations (e.g., checks or kfrees) between two code paths
> and confirms that the inconsistent operations are not recovered in
> the current function or the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>
> Changes in v2:
> - Fix the same bug in ACRN_IOCTL_SET_VCPU_REGS.
>
Signed-off-by: Fei Li <[email protected]>
Thanks a lot.
> drivers/virt/acrn/hsm.c | 24 +++++++++++++++++-------
> 1 file changed, 17 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/virt/acrn/hsm.c b/drivers/virt/acrn/hsm.c
> index 5419794fccf1..7a8f722f431c 100644
> --- a/drivers/virt/acrn/hsm.c
> +++ b/drivers/virt/acrn/hsm.c
> @@ -136,9 +136,11 @@ static long acrn_dev_ioctl(struct file *filp, unsigned int cmd,
> if (IS_ERR(vm_param))
> return PTR_ERR(vm_param);
>
> - if ((vm_param->reserved0 | vm_param->reserved1) != 0)
> - return -EINVAL;
> -
> + if ((vm_param->reserved0 | vm_param->reserved1) != 0) {
> + ret = -EINVAL;
> + kfree(vm_param);
> + break;
> + }
> vm = acrn_vm_create(vm, vm_param);
> if (!vm) {
> ret = -EINVAL;
> @@ -182,21 +184,29 @@ static long acrn_dev_ioctl(struct file *filp, unsigned int cmd,
> return PTR_ERR(cpu_regs);
>
> for (i = 0; i < ARRAY_SIZE(cpu_regs->reserved); i++)
> - if (cpu_regs->reserved[i])
> + if (cpu_regs->reserved[i]) {
> + kfree(cpu_regs);
> return -EINVAL;
> + }
>
> for (i = 0; i < ARRAY_SIZE(cpu_regs->vcpu_regs.reserved_32); i++)
> - if (cpu_regs->vcpu_regs.reserved_32[i])
> + if (cpu_regs->vcpu_regs.reserved_32[i]) {
> + kfree(cpu_regs);
> return -EINVAL;
> + }
>
> for (i = 0; i < ARRAY_SIZE(cpu_regs->vcpu_regs.reserved_64); i++)
> - if (cpu_regs->vcpu_regs.reserved_64[i])
> + if (cpu_regs->vcpu_regs.reserved_64[i]) {
> + kfree(cpu_regs);
> return -EINVAL;
> + }
>
> for (i = 0; i < ARRAY_SIZE(cpu_regs->vcpu_regs.gdt.reserved); i++)
> if (cpu_regs->vcpu_regs.gdt.reserved[i] |
> - cpu_regs->vcpu_regs.idt.reserved[i])
> + cpu_regs->vcpu_regs.idt.reserved[i]) {
> + kfree(cpu_regs);
> return -EINVAL;
> + }
>
> ret = hcall_set_vcpu_regs(vm->vmid, virt_to_phys(cpu_regs));
> if (ret < 0)
> --
> 2.25.1
>
On Tue, Jan 04, 2022 at 11:50:33AM +0800, Li Fei1 wrote:
> On Tue, Jan 04, 2022 at 11:45:01AM +0800, Zhou Qingyang wrote:
> > In acrn_dev_ioctl(), vm_param is not released or passed out on the
> > error path of "if ((vm_param->reserved0 | vm_param->reserved1) != 0)",
> > which could lead to a memory leak.
> >
> > Fix this bug by adding a kfree of vm_param on the error path.
> >
> > This bug was found by a static analyzer.
> >
> > Builds with CONFIG_ACRN_GUEST=y, CONFIG_ACRN_HSM=y show no new warnings,
> > and our static analyzer no longer warns about this code.
> >
> > Fixes: 9c5137aedd11 (“9c5137aedd11 virt: acrn: Introduce VM management interfaces”)
> > Signed-off-by: Zhou Qingyang <[email protected]>
> > ---
> > The analysis employs differential checking to identify inconsistent
> > security operations (e.g., checks or kfrees) between two code paths
> > and confirms that the inconsistent operations are not recovered in
> > the current function or the callers, so they constitute bugs.
> >
> > Note that, as a bug found by static analysis, it can be a false
> > positive or hard to trigger. Multiple researchers have cross-reviewed
> > the bug.
> >
> > Changes in v2:
> > - Fix the same bug in ACRN_IOCTL_SET_VCPU_REGS.
> >
>
> Signed-off-by: Fei Li <[email protected]>
For the multiple obvious reasons (not the least being you didn't even
run it through our normal testing scripts), I am not going to take this
change, sorry.
Fei, please be more careful in the future.
greg k-h
On Tue, Jan 04, 2022 at 08:37:27AM +0100, Greg KH wrote:
> On Tue, Jan 04, 2022 at 11:50:33AM +0800, Li Fei1 wrote:
> > On Tue, Jan 04, 2022 at 11:45:01AM +0800, Zhou Qingyang wrote:
> > > In acrn_dev_ioctl(), vm_param is not released or passed out on the
> > > error path of "if ((vm_param->reserved0 | vm_param->reserved1) != 0)",
> > > which could lead to a memory leak.
> > >
> > > Fix this bug by adding a kfree of vm_param on the error path.
> > >
> > > This bug was found by a static analyzer.
> > >
> > > Builds with CONFIG_ACRN_GUEST=y, CONFIG_ACRN_HSM=y show no new warnings,
> > > and our static analyzer no longer warns about this code.
> > >
> > > Fixes: 9c5137aedd11 (“9c5137aedd11 virt: acrn: Introduce VM management interfaces”)
> > > Signed-off-by: Zhou Qingyang <[email protected]>
> > > ---
> > > The analysis employs differential checking to identify inconsistent
> > > security operations (e.g., checks or kfrees) between two code paths
> > > and confirms that the inconsistent operations are not recovered in
> > > the current function or the callers, so they constitute bugs.
> > >
> > > Note that, as a bug found by static analysis, it can be a false
> > > positive or hard to trigger. Multiple researchers have cross-reviewed
> > > the bug.
> > >
> > > Changes in v2:
> > > - Fix the same bug in ACRN_IOCTL_SET_VCPU_REGS.
> > >
> >
> > Signed-off-by: Fei Li <[email protected]>
>
> For the multiple obvious reasons (not the least being you didn't even
> run it through our normal testing scripts), I am not going to take this
> change, sorry.
>
> Fei, please be more careful in the future.
>
Hi Greg
Thanks for your kind reminder. I will.
> greg k-h
On Tue, Jan 04, 2022 at 08:37:27AM +0100, Greg KH wrote:
> On Tue, Jan 04, 2022 at 11:50:33AM +0800, Li Fei1 wrote:
> > On Tue, Jan 04, 2022 at 11:45:01AM +0800, Zhou Qingyang wrote:
> > > In acrn_dev_ioctl(), vm_param is not released or passed out on the
> > > error path of "if ((vm_param->reserved0 | vm_param->reserved1) != 0)",
> > > which could lead to a memory leak.
> > >
> > > Fix this bug by adding a kfree of vm_param on the error path.
> > >
> > > This bug was found by a static analyzer.
> > >
> > > Builds with CONFIG_ACRN_GUEST=y, CONFIG_ACRN_HSM=y show no new warnings,
> > > and our static analyzer no longer warns about this code.
> > >
> > > Fixes: 9c5137aedd11 (“9c5137aedd11 virt: acrn: Introduce VM management interfaces”)
> > > Signed-off-by: Zhou Qingyang <[email protected]>
> > > ---
> > > The analysis employs differential checking to identify inconsistent
> > > security operations (e.g., checks or kfrees) between two code paths
> > > and confirms that the inconsistent operations are not recovered in
> > > the current function or the callers, so they constitute bugs.
> > >
> > > Note that, as a bug found by static analysis, it can be a false
> > > positive or hard to trigger. Multiple researchers have cross-reviewed
> > > the bug.
> > >
> > > Changes in v2:
> > > - Fix the same bug in ACRN_IOCTL_SET_VCPU_REGS.
Hi Qingyang
I think the "Changes in v2" was not fixes the code in commit "9c5137aedd11 virt: acrn: Introduce VM management interfaces", but in commit 2ad2aaee1bc9 "virt: acrn: Introduce an ioctl to set vCPU registers state".
Would you please to split this patch into two and please replace “...” with "..." in your commit message.
Thanks.
> > >
> >
> > Signed-off-by: Fei Li <[email protected]>
Please ignore my "Signed-off" here. Sorry.
>
> For the multiple obvious reasons (not the least being you didn't even
> run it through our normal testing scripts), I am not going to take this
> change, sorry.
>
> Fei, please be more careful in the future.
>
> greg k-h