From: "Michael S. Tsirkin" <mst@redhat.com>
To: arei.gonglei@huawei.com
Cc: weidong.huangwei@huawei.com, pbonzini@redhat.com,
alex.williamson@redhat.com, qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH 2/2] pci-assign: Fix memory out of bound when MSI-X table not fit in a single page
Date: Tue, 8 Apr 2014 18:32:12 +0300 [thread overview]
Message-ID: <20140408153212.GA8087@redhat.com> (raw)
In-Reply-To: <1396502304-7456-2-git-send-email-arei.gonglei@huawei.com>
On Thu, Apr 03, 2014 at 01:18:24PM +0800, arei.gonglei@huawei.com wrote:
> From: Gonglei <arei.gonglei@huawei.com>
>
> QEMU only mmap MSIX_PAGE_SIZE memory for all pci devices in
> assigned_dev_register_msix_mmio(), meanwhile the set the one
> page memmory to zero, so the rest memory will be random value
> (maybe etnry.data is not 0). In the assigned_dev_update_msix_mmio()
> maybe occur the issue of entry_nr > 256, and the kmod reports
> the EINVAL error.
>
> Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Okay so this kind of works because guest does not try
to use so many vectors.
But I think it's better not to give guest more entries
than we can actually support.
How about tweaking MSIX capability exposed to guest to limit table size?
> ---
> hw/i386/kvm/pci-assign.c | 15 ++++++++++++---
> 1 file changed, 12 insertions(+), 3 deletions(-)
>
> diff --git a/hw/i386/kvm/pci-assign.c b/hw/i386/kvm/pci-assign.c
> index 570333f..d25a19e 100644
> --- a/hw/i386/kvm/pci-assign.c
> +++ b/hw/i386/kvm/pci-assign.c
> @@ -37,6 +37,7 @@
> #include "hw/pci/pci.h"
> #include "hw/pci/msi.h"
> #include "kvm_i386.h"
> +#include "qemu/osdep.h"
>
> #define MSIX_PAGE_SIZE 0x1000
>
> @@ -59,6 +60,9 @@
> #define DEBUG(fmt, ...)
> #endif
>
> +/* the msix-table size readed from pci device config */
> +static int msix_table_size;
> +
> typedef struct PCIRegion {
> int type; /* Memory or port I/O */
> int valid;
> @@ -1604,7 +1608,12 @@ static void assigned_dev_msix_reset(AssignedDevice *dev)
>
> static int assigned_dev_register_msix_mmio(AssignedDevice *dev)
> {
> - dev->msix_table = mmap(NULL, MSIX_PAGE_SIZE, PROT_READ|PROT_WRITE,
> + msix_table_size = ROUND_UP(dev->msix_max * sizeof(struct MSIXTableEntry),
> + MSIX_PAGE_SIZE);
> +
> + DEBUG("msix_table_size: 0x%x\n", msix_table_size);
> +
> + dev->msix_table = mmap(NULL, msix_table_size, PROT_READ|PROT_WRITE,
> MAP_ANONYMOUS|MAP_PRIVATE, 0, 0);
> if (dev->msix_table == MAP_FAILED) {
> error_report("fail allocate msix_table! %s", strerror(errno));
> @@ -1615,7 +1624,7 @@ static int assigned_dev_register_msix_mmio(AssignedDevice *dev)
> assigned_dev_msix_reset(dev);
>
> memory_region_init_io(&dev->mmio, OBJECT(dev), &assigned_dev_msix_mmio_ops,
> - dev, "assigned-dev-msix", MSIX_PAGE_SIZE);
> + dev, "assigned-dev-msix", msix_table_size);
> return 0;
> }
>
> @@ -1627,7 +1636,7 @@ static void assigned_dev_unregister_msix_mmio(AssignedDevice *dev)
>
> memory_region_destroy(&dev->mmio);
>
> - if (munmap(dev->msix_table, MSIX_PAGE_SIZE) == -1) {
> + if (munmap(dev->msix_table, msix_table_size) == -1) {
> error_report("error unmapping msix_table! %s", strerror(errno));
> }
> dev->msix_table = NULL;
> --
> 1.7.12.4
>
>
next prev parent reply other threads:[~2014-04-08 15:33 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-04-03 5:18 [Qemu-devel] [PATCH 1/2] pci-assign: Fix a bug when map MSI-X table memory failed arei.gonglei
2014-04-03 5:18 ` [Qemu-devel] [PATCH 2/2] pci-assign: Fix memory out of bound when MSI-X table not fit in a single page arei.gonglei
2014-04-08 15:32 ` Michael S. Tsirkin [this message]
2014-04-09 10:56 ` Gonglei (Arei)
2014-04-09 13:52 ` Michael S. Tsirkin
2014-04-10 2:34 ` Gonglei (Arei)
2014-04-09 14:12 ` Laszlo Ersek
2014-04-10 2:05 ` Gonglei (Arei)
2014-04-08 14:02 ` [Qemu-devel] [PATCH 1/2] pci-assign: Fix a bug when map MSI-X table memory failed Gonglei (Arei)
2014-04-08 15:32 ` Michael S. Tsirkin
2014-04-09 14:21 ` Michael S. Tsirkin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20140408153212.GA8087@redhat.com \
--to=mst@redhat.com \
--cc=alex.williamson@redhat.com \
--cc=arei.gonglei@huawei.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=weidong.huangwei@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).