[PATCH] drivers/misc/vmw_vmci: restrict too big queue size in qp_host_alloc_queue

[PATCH] drivers/misc/vmw_vmci: restrict too big queue size in qp_host_alloc_queue

[Sabyrzhan Tasbolatov]

syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
when num_pages is 0x100001, giving queue_size + queue_page_size
bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
condition.

queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.


FYI, I've also noticed in vmci_queue_pair.c other SLAB allocations with no
length check that might exceed KMALLOC_MAX_SIZE as well,
but syzbot doesn't have reproduces for them.

in qp_alloc_ppn_set():
	produce_ppns =
	    kmalloc_array(num_produce_pages, sizeof(*produce_ppns),
			  GFP_KERNEL);
[..]
	consume_ppns =
	    kmalloc_array(num_consume_pages, sizeof(*consume_ppns),
			  GFP_KERNEL);
[..]
in qp_alloc_hypercall():
	msg_size = sizeof(*alloc_msg) +
	    (size_t) entry->num_ppns * ppn_size;
	alloc_msg = kmalloc(msg_size, GFP_KERNEL);
[..]
in qp_broker_create():
	entry->local_mem = kcalloc(QPE_NUM_PAGES(entry->qp),
					   PAGE_SIZE, GFP_KERNEL);

[1]
Call Trace:
 alloc_pages include/linux/gfp.h:547 [inline]
 kmalloc_order+0x40/0x130 mm/slab_common.c:837
 kmalloc_order_trace+0x15/0x70 mm/slab_common.c:853
 kmalloc_large include/linux/slab.h:481 [inline]
 __kmalloc+0x257/0x330 mm/slub.c:3959
 kmalloc include/linux/slab.h:557 [inline]
 kzalloc include/linux/slab.h:682 [inline]
 qp_host_alloc_queue drivers/misc/vmw_vmci/vmci_queue_pair.c:540 [inline]
 qp_broker_create drivers/misc/vmw_vmci/vmci_queue_pair.c:1351 [inline]
 qp_broker_alloc+0x936/0x2740 drivers/misc/vmw_vmci/vmci_queue_pair.c:1739

Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
---
 drivers/misc/vmw_vmci/vmci_queue_pair.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
index c49065887e8f..f6af406fda80 100644
--- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
+++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
@@ -537,6 +537,11 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
 
 	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
 
+	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE) {
+		pr_warn("too big queue to allocate\n");
+		return NULL;
+	}
+
 	queue = kzalloc(queue_size + queue_page_size, GFP_KERNEL);
 	if (queue) {
 		queue->q_header = NULL;
-- 
2.25.1

Re: [PATCH] drivers/misc/vmw_vmci: restrict too big queue size in qp_host_alloc_queue

[Greg KH]

On Sat, Feb 06, 2021 at 11:34:09AM +0600, Sabyrzhan Tasbolatov wrote:
> syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
> when num_pages is 0x100001, giving queue_size + queue_page_size
> bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
> condition.
> 
> queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.
> 
> 
> FYI, I've also noticed in vmci_queue_pair.c other SLAB allocations with no
> length check that might exceed KMALLOC_MAX_SIZE as well,
> but syzbot doesn't have reproduces for them.
> 
> in qp_alloc_ppn_set():
> 	produce_ppns =
> 	    kmalloc_array(num_produce_pages, sizeof(*produce_ppns),
> 			  GFP_KERNEL);
> [..]
> 	consume_ppns =
> 	    kmalloc_array(num_consume_pages, sizeof(*consume_ppns),
> 			  GFP_KERNEL);
> [..]
> in qp_alloc_hypercall():
> 	msg_size = sizeof(*alloc_msg) +
> 	    (size_t) entry->num_ppns * ppn_size;
> 	alloc_msg = kmalloc(msg_size, GFP_KERNEL);
> [..]
> in qp_broker_create():
> 	entry->local_mem = kcalloc(QPE_NUM_PAGES(entry->qp),
> 					   PAGE_SIZE, GFP_KERNEL);
> 
> [1]
> Call Trace:
>  alloc_pages include/linux/gfp.h:547 [inline]
>  kmalloc_order+0x40/0x130 mm/slab_common.c:837
>  kmalloc_order_trace+0x15/0x70 mm/slab_common.c:853
>  kmalloc_large include/linux/slab.h:481 [inline]
>  __kmalloc+0x257/0x330 mm/slub.c:3959
>  kmalloc include/linux/slab.h:557 [inline]
>  kzalloc include/linux/slab.h:682 [inline]
>  qp_host_alloc_queue drivers/misc/vmw_vmci/vmci_queue_pair.c:540 [inline]
>  qp_broker_create drivers/misc/vmw_vmci/vmci_queue_pair.c:1351 [inline]
>  qp_broker_alloc+0x936/0x2740 drivers/misc/vmw_vmci/vmci_queue_pair.c:1739
> 
> Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
> Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
> ---
>  drivers/misc/vmw_vmci/vmci_queue_pair.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> index c49065887e8f..f6af406fda80 100644
> --- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
> +++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> @@ -537,6 +537,11 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
>  
>  	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
>  
> +	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE) {
> +		pr_warn("too big queue to allocate\n");

As this is controllable by userspace, you just provided a way to flood
the kernel logs.

Please make this a dev_dbg() call instead, if you really want to see it.
Otherwise just return NULL, no need to report anything, right?

thanks,

greg k-h

[PATCH v2] drivers/misc/vmw_vmci: restrict too big queue size in

[Sabyrzhan Tasbolatov]

syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
when num_pages is 0x100001, giving queue_size + queue_page_size
bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
condition.

queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.

Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
---
>As this is controllable by userspace, you just provided a way to flood
>the kernel logs.
>
>Please make this a dev_dbg() call instead, if you really want to see it.
>Otherwise just return NULL, no need to report anything, right?

Thanks, removed pr_warn().

v2: Removed pr_warn() to avoid flood from user-space
---
 drivers/misc/vmw_vmci/vmci_queue_pair.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
index f6af406fda80..ea16df73cde0 100644
--- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
+++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
@@ -538,9 +538,7 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
 	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
 
 	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE) {
-		pr_warn("too big queue to allocate\n");
 		return NULL;
-	}
 
 	queue = kzalloc(queue_size + queue_page_size, GFP_KERNEL);
 	if (queue) {
-- 
2.25.1

Re: [PATCH v2] drivers/misc/vmw_vmci: restrict too big queue size in

[Alex Dewar]

On 09/02/2021 09:31, Sabyrzhan Tasbolatov wrote:
> syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
> when num_pages is 0x100001, giving queue_size + queue_page_size
> bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
> condition.
>
> queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.
>
> Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
> Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
> ---
>> As this is controllable by userspace, you just provided a way to flood
>> the kernel logs.
>>
>> Please make this a dev_dbg() call instead, if you really want to see it.
>> Otherwise just return NULL, no need to report anything, right?
> Thanks, removed pr_warn().
Looks like you forgot to take out the opening brace.

>
> v2: Removed pr_warn() to avoid flood from user-space
> ---
>   drivers/misc/vmw_vmci/vmci_queue_pair.c | 2 --
>   1 file changed, 2 deletions(-)
>
> diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> index f6af406fda80..ea16df73cde0 100644
> --- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
> +++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> @@ -538,9 +538,7 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
>   	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
>   
>   	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE) {
> -		pr_warn("too big queue to allocate\n");
>   		return NULL;
> -	}
>   
>   	queue = kzalloc(queue_size + queue_page_size, GFP_KERNEL);
>   	if (queue) {

[PATCH v3] drivers/misc/vmw_vmci: restrict too big queue size in

[Sabyrzhan Tasbolatov]

> syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
> when num_pages is 0x100001, giving queue_size + queue_page_size
> bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
> condition.
>
> queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.
>
Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
> ---
>>> As this is controllable by userspace, you just provided a way to flood
>>> the kernel logs.
>>>
>>> Please make this a dev_dbg() call instead, if you really want to see it.
>>> Otherwise just return NULL, no need to report anything, right?
>> Thanks, removed pr_warn().

>Looks like you forgot to take out the opening brace.

Cringe moment. Sorry, should've checked it properly first.

v3: Removed opening brace.
---
 drivers/misc/vmw_vmci/vmci_queue_pair.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
index ea16df73cde0..024dcdbd9d01 100644
--- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
+++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
@@ -537,7 +537,7 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
 
 	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
 
-	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE) {
+	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE)
 		return NULL;
 
 	queue = kzalloc(queue_size + queue_page_size, GFP_KERNEL);
-- 
2.25.1

Re: [PATCH v3] drivers/misc/vmw_vmci: restrict too big queue size in

[Greg KH]

On Tue, Feb 09, 2021 at 03:45:25PM +0600, Sabyrzhan Tasbolatov wrote:
> > syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
> > when num_pages is 0x100001, giving queue_size + queue_page_size
> > bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
> > condition.
> >
> > queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.
> >

Why is this "quoted"?

> Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
> Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
> > ---
> >>> As this is controllable by userspace, you just provided a way to flood
> >>> the kernel logs.
> >>>
> >>> Please make this a dev_dbg() call instead, if you really want to see it.
> >>> Otherwise just return NULL, no need to report anything, right?
> >> Thanks, removed pr_warn().
> 
> >Looks like you forgot to take out the opening brace.
> 
> Cringe moment. Sorry, should've checked it properly first.
> 
> v3: Removed opening brace.
> ---
>  drivers/misc/vmw_vmci/vmci_queue_pair.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> index ea16df73cde0..024dcdbd9d01 100644
> --- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
> +++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> @@ -537,7 +537,7 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
>  
>  	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
>  
> -	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE) {
> +	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE)
>  		return NULL;

This patch does not apply to the tree...

thanks,

greg k-h

[PATCH v4] drivers/misc/vmw_vmci: restrict too big queue size in qp_host_alloc_queue

[Sabyrzhan Tasbolatov]

syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
when num_pages is 0x100001, giving queue_size + queue_page_size
bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
condition.

queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.

[1]
Call Trace:
 alloc_pages include/linux/gfp.h:547 [inline]
 kmalloc_order+0x40/0x130 mm/slab_common.c:837
 kmalloc_order_trace+0x15/0x70 mm/slab_common.c:853
 kmalloc_large include/linux/slab.h:481 [inline]
 __kmalloc+0x257/0x330 mm/slub.c:3959
 kmalloc include/linux/slab.h:557 [inline]
 kzalloc include/linux/slab.h:682 [inline]
 qp_host_alloc_queue drivers/misc/vmw_vmci/vmci_queue_pair.c:540 [inline]
 qp_broker_create drivers/misc/vmw_vmci/vmci_queue_pair.c:1351 [inline]
 qp_broker_alloc+0x936/0x2740 drivers/misc/vmw_vmci/vmci_queue_pair.c:1739

Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
---
> This patch does not apply to the tree...
Apologies, it was so stupid from my side.
Tested locally and via syzbot.

v4: made a patch based on commit 65f0d2414b("Merge tag 'sound-5.11-rc4'
of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound")
---
 drivers/misc/vmw_vmci/vmci_queue_pair.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
index c49065887e8f..024dcdbd9d01 100644
--- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
+++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
@@ -537,6 +537,9 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
 
 	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
 
+	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE)
+		return NULL;
+
 	queue = kzalloc(queue_size + queue_page_size, GFP_KERNEL);
 	if (queue) {
 		queue->q_header = NULL;
-- 
2.25.1

Re: [PATCH v2] drivers/misc/vmw_vmci: restrict too big queue size in

[Greg KH]

On Tue, Feb 09, 2021 at 03:31:01PM +0600, Sabyrzhan Tasbolatov wrote:
> syzbot found WARNING in qp_broker_alloc[1] in qp_host_alloc_queue()
> when num_pages is 0x100001, giving queue_size + queue_page_size
> bigger than KMALLOC_MAX_SIZE for kzalloc(), resulting order >= MAX_ORDER
> condition.
> 
> queue_size + queue_page_size=0x8000d8, where KMALLOC_MAX_SIZE=0x400000.
> 
> Reported-by: syzbot+15ec7391f3d6a1a7cc7d@syzkaller.appspotmail.com
> Signed-off-by: Sabyrzhan Tasbolatov <snovitoll@gmail.com>
> ---
> >As this is controllable by userspace, you just provided a way to flood
> >the kernel logs.
> >
> >Please make this a dev_dbg() call instead, if you really want to see it.
> >Otherwise just return NULL, no need to report anything, right?
> 
> Thanks, removed pr_warn().
> 
> v2: Removed pr_warn() to avoid flood from user-space
> ---
>  drivers/misc/vmw_vmci/vmci_queue_pair.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> index f6af406fda80..ea16df73cde0 100644
> --- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
> +++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
> @@ -538,9 +538,7 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
>  	queue_page_size = num_pages * sizeof(*queue->kernel_if->u.h.page);
>  
>  	if (queue_size + queue_page_size > KMALLOC_MAX_SIZE) {
> -		pr_warn("too big queue to allocate\n");
>  		return NULL;
> -	}
>  

This patch doesn't apply to the tree, you can't send me a patch that is
relative to your previous one, that doesn't work at all.

And also, test-build your patches always first, to not do so is a bit
rude to maintainers...

thanks,

greg k-h