From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:48831) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1UWN4w-0005Df-Iq for qemu-devel@nongnu.org; Sun, 28 Apr 2013 04:35:30 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1UWN4t-0007fu-W2 for qemu-devel@nongnu.org; Sun, 28 Apr 2013 04:35:26 -0400 Received: from mx1.redhat.com ([209.132.183.28]:11576) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1UWN4t-0007fm-OP for qemu-devel@nongnu.org; Sun, 28 Apr 2013 04:35:23 -0400 Date: Sun, 28 Apr 2013 11:35:14 +0300 From: "Michael S. Tsirkin" Message-ID: <20130428083514.GE7106@redhat.com> References: <1366965244-20542-1-git-send-email-jasowang@redhat.com> <20130427192642.GA30188@redhat.com> <517CD5AC.2080708@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <517CD5AC.2080708@redhat.com> Subject: Re: [Qemu-devel] [PATCH 1/3] virtio-pci: properly validate address before accessing config List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jason Wang Cc: aliguori@us.ibm.com, pmatouse@redhat.com, qemu-devel@nongnu.org On Sun, Apr 28, 2013 at 03:54:20PM +0800, Jason Wang wrote: > On 04/28/2013 03:26 AM, Michael S. Tsirkin wrote: > > On Fri, Apr 26, 2013 at 04:34:02PM +0800, Jason Wang wrote: > >> There are several several issues in the current checking: > >> > >> - The check was based on the minus of unsigned values which can overflow > >> - It was done after .{set|get}_config() which can lead crash when config_len is > >> zero since vdev->config is NULL > >> > >> Fix this by: > >> > >> - Validate the address in virtio_pci_config_{read|write}() before > >> .{set|get}_config > >> - Use addition instead minus to do the validation > >> > >> Cc: Michael S. Tsirkin > >> Cc: Petr Matousek > >> Signed-off-by: Jason Wang > > Why do this in virtio-pci and not in virtio.c? > > If instead we correct the checks in virtio.c we > > get less code, and all transports will benefit > > automatically. > > I wish I could but looks like vitio_config_read{b|w|l} were only used by > virtio-pci. Other transport such as ccw and s390-virtio-bus have their > own implementation. Okay but still, the bug is in checks in virtio.c, why not fix it there instead of making it assume caller does the checks? > > > >> --- > >> hw/virtio/virtio-pci.c | 9 +++++++++ > >> hw/virtio/virtio.c | 18 ------------------ > >> 2 files changed, 9 insertions(+), 18 deletions(-) > >> > >> diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c > >> index a1f15a8..7f6c7d1 100644 > >> --- a/hw/virtio/virtio-pci.c > >> +++ b/hw/virtio/virtio-pci.c > >> @@ -400,6 +400,10 @@ static uint64_t virtio_pci_config_read(void *opaque, hwaddr addr, > >> } > >> addr -= config; > >> > >> + if (addr + size > proxy->vdev->config_len) { > >> + return (uint32_t)-1; > >> + } > >> + > >> switch (size) { > >> case 1: > >> val = virtio_config_readb(proxy->vdev, addr); > >> @@ -430,6 +434,11 @@ static void virtio_pci_config_write(void *opaque, hwaddr addr, > >> return; > >> } > >> addr -= config; > >> + > >> + if (addr + size > proxy->vdev->config_len) { > >> + return; > >> + } > >> + > >> /* > >> * Virtio-PCI is odd. Ioports are LE but config space is target native > >> * endian. > >> diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c > >> index 1c2282c..3397b5e 100644 > >> --- a/hw/virtio/virtio.c > >> +++ b/hw/virtio/virtio.c > >> @@ -563,9 +563,6 @@ uint32_t virtio_config_readb(VirtIODevice *vdev, uint32_t addr) > >> > >> vdev->get_config(vdev, vdev->config); > >> > >> - if (addr > (vdev->config_len - sizeof(val))) > >> - return (uint32_t)-1; > >> - > >> val = ldub_p(vdev->config + addr); > >> return val; > >> } > >> @@ -576,9 +573,6 @@ uint32_t virtio_config_readw(VirtIODevice *vdev, uint32_t addr) > >> > >> vdev->get_config(vdev, vdev->config); > >> > >> - if (addr > (vdev->config_len - sizeof(val))) > >> - return (uint32_t)-1; > >> - > >> val = lduw_p(vdev->config + addr); > >> return val; > >> } > >> @@ -589,9 +583,6 @@ uint32_t virtio_config_readl(VirtIODevice *vdev, uint32_t addr) > >> > >> vdev->get_config(vdev, vdev->config); > >> > >> - if (addr > (vdev->config_len - sizeof(val))) > >> - return (uint32_t)-1; > >> - > >> val = ldl_p(vdev->config + addr); > >> return val; > >> } > >> @@ -600,9 +591,6 @@ void virtio_config_writeb(VirtIODevice *vdev, uint32_t addr, uint32_t data) > >> { > >> uint8_t val = data; > >> > >> - if (addr > (vdev->config_len - sizeof(val))) > >> - return; > >> - > >> stb_p(vdev->config + addr, val); > >> > >> if (vdev->set_config) > >> @@ -613,9 +601,6 @@ void virtio_config_writew(VirtIODevice *vdev, uint32_t addr, uint32_t data) > >> { > >> uint16_t val = data; > >> > >> - if (addr > (vdev->config_len - sizeof(val))) > >> - return; > >> - > >> stw_p(vdev->config + addr, val); > >> > >> if (vdev->set_config) > >> @@ -626,9 +611,6 @@ void virtio_config_writel(VirtIODevice *vdev, uint32_t addr, uint32_t data) > >> { > >> uint32_t val = data; > >> > >> - if (addr > (vdev->config_len - sizeof(val))) > >> - return; > >> - > >> stl_p(vdev->config + addr, val); > >> > >> if (vdev->set_config) > >> -- > >> 1.7.1