lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20210118112756.ekfebcbyqwz4dd4b@steredhat>
Date:   Mon, 18 Jan 2021 12:27:56 +0100
From:   Stefano Garzarella <sgarzare@...hat.com>
To:     Jorgen Hansen <jhansen@...are.com>
Cc:     linux-kernel@...r.kernel.org,
        virtualization@...ts.linux-foundation.org, pv-drivers@...are.com,
        gregkh@...uxfoundation.org, Norbert Slusarek <nslusarek@....net>
Subject: Re: [PATCH] VMCI: Enforce queuepair max size for
 IOCTL_VMCI_QUEUEPAIR_ALLOC

+Cc: Norbert Slusarek <nslusarek@....net>

On Mon, Jan 11, 2021 at 04:18:53AM -0800, Jorgen Hansen wrote:
>When create the VMCI queue pair tracking data structures on the host
>side, the IOCTL for creating the VMCI queue pair didn't validate
>the queue pair size parameters. This change adds checks for this.
>
>This avoids a memory allocation issue in qp_host_alloc_queue, as
>reported by nslusarek@....net. The check in qp_host_alloc_queue
>has also been updated to enforce the maximum queue pair size
>as defined by VMCI_MAX_GUEST_QP_MEMORY.
>
>The fix has been verified using sample code supplied by
>nslusarek@....net.
>
>Reported-by: nslusarek@....net
>Reviewed-by: Vishnu Dasa <vdasa@...are.com>
>Signed-off-by: Jorgen Hansen <jhansen@...are.com>
>---
> drivers/misc/vmw_vmci/vmci_queue_pair.c | 12 ++++++++----
> include/linux/vmw_vmci_defs.h           |  4 ++--
> 2 files changed, 10 insertions(+), 6 deletions(-)
>
>diff --git a/drivers/misc/vmw_vmci/vmci_queue_pair.c b/drivers/misc/vmw_vmci/vmci_queue_pair.c
>index 525ef96..39d2f191 100644
>--- a/drivers/misc/vmw_vmci/vmci_queue_pair.c
>+++ b/drivers/misc/vmw_vmci/vmci_queue_pair.c
>@@ -237,7 +237,9 @@ static struct qp_list qp_guest_endpoints = {
> #define QPE_NUM_PAGES(_QPE) ((u32) \
> 			     (DIV_ROUND_UP(_QPE.produce_size, PAGE_SIZE) + \
> 			      DIV_ROUND_UP(_QPE.consume_size, PAGE_SIZE) + 2))
>-
>+#define QP_SIZES_ARE_VALID(_prod_qsize, _cons_qsize) \
>+	((_prod_qsize) + (_cons_qsize) >= max(_prod_qsize, _cons_qsize) && \
>+	 (_prod_qsize) + (_cons_qsize) <= VMCI_MAX_GUEST_QP_MEMORY)
>
> /*
>  * Frees kernel VA space for a given queue and its queue header, and
>@@ -528,7 +530,7 @@ static struct vmci_queue *qp_host_alloc_queue(u64 size)
> 	u64 num_pages;
> 	const size_t queue_size = sizeof(*queue) + sizeof(*(queue->kernel_if));
>
>-	if (size > SIZE_MAX - PAGE_SIZE)
>+	if (size > min(VMCI_MAX_GUEST_QP_MEMORY, SIZE_MAX - PAGE_SIZE))
> 		return NULL;
> 	num_pages = DIV_ROUND_UP(size, PAGE_SIZE) + 1;
> 	if (num_pages > (SIZE_MAX - queue_size) /
>@@ -1929,6 +1931,9 @@ int vmci_qp_broker_alloc(struct vmci_handle handle,
> 			 struct vmci_qp_page_store *page_store,
> 			 struct vmci_ctx *context)
> {
>+	if (!QP_SIZES_ARE_VALID(produce_size, consume_size))
>+		return VMCI_ERROR_NO_RESOURCES;
>+
> 	return qp_broker_alloc(handle, peer, flags, priv_flags,
> 			       produce_size, consume_size,
> 			       page_store, context, NULL, NULL, NULL, NULL);
>@@ -2685,8 +2690,7 @@ int vmci_qpair_alloc(struct vmci_qp **qpair,
> 	 * used by the device is NO_RESOURCES, so use that here too.
> 	 */
>
>-	if (produce_qsize + consume_qsize < max(produce_qsize, consume_qsize) ||
>-	    produce_qsize + consume_qsize > VMCI_MAX_GUEST_QP_MEMORY)
>+	if (!QP_SIZES_ARE_VALID(produce_qsize, consume_qsize))
> 		return VMCI_ERROR_NO_RESOURCES;
>
> 	retval = vmci_route(&src, &dst, false, &route);
>diff --git a/include/linux/vmw_vmci_defs.h b/include/linux/vmw_vmci_defs.h
>index be0afe6..e36cb11 100644
>--- a/include/linux/vmw_vmci_defs.h
>+++ b/include/linux/vmw_vmci_defs.h
>@@ -66,7 +66,7 @@ enum {
>  * consists of at least two pages, the memory limit also dictates the
>  * number of queue pairs a guest can create.
>  */
>-#define VMCI_MAX_GUEST_QP_MEMORY (128 * 1024 * 1024)
>+#define VMCI_MAX_GUEST_QP_MEMORY ((size_t)(128 * 1024 * 1024))
> #define VMCI_MAX_GUEST_QP_COUNT  (VMCI_MAX_GUEST_QP_MEMORY / PAGE_SIZE / 2)
>
> /*
>@@ -80,7 +80,7 @@ enum {
>  * too much kernel memory (especially on vmkernel).  We limit a queuepair to
>  * 32 KB, or 16 KB per queue for symmetrical pairs.
>  */
>-#define VMCI_MAX_PINNED_QP_MEMORY (32 * 1024)
>+#define VMCI_MAX_PINNED_QP_MEMORY ((size_t)(32 * 1024))
>
> /*
>  * We have a fixed set of resource IDs available in the VMX.
>-- 
>2.6.2
>
>_______________________________________________
>Virtualization mailing list
>Virtualization@...ts.linux-foundation.org
>https://lists.linuxfoundation.org/mailman/listinfo/virtualization
>

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ