From: Rusty Russell <rusty@rustcorp.com.au>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: virtualization@lists.linux-foundation.org,
"H. Peter Anvin" <hpa@zytor.com>
Subject: Re: [PATCH 16/22] virtio_pci: use separate notification offsets for each vq.
Date: Fri, 05 Apr 2013 11:55:44 +1030 [thread overview]
Message-ID: <874nfln4yv.fsf@rustcorp.com.au> (raw)
In-Reply-To: <20130404082559.GA10369@redhat.com>
"Michael S. Tsirkin" <mst@redhat.com> writes:
> By the way, Gleb pointed out that on older hosts MMIO will
> always be slower since we need to do a shadow page walk to
> translate virtual to physical address.
> Hopefully not a big concern, and after all we are still
> keeping PIO around for use by BIOS ...
Yeah, slow hosts will be slow :)
>> + /* get offset of notification byte for this virtqueue */
>> + off = ioread16(&vp_dev->common->queue_notify_off);
>> + if (off * vp_dev->notify_offset_multiplier > vp_dev->notify_len) {
>
> maybe check there's no overflow too?
> if (UINT32_MAX / vp_dev->notify_offset_multiplier > off)
> return -EINVAL;
I think it's clearer to just do:
if ((u64)off * vp_dev->notify_offset_multiplier > vp_dev->notify_len) {
Since off is 16 bits and notify_offset_multiplier is 32, this catches
overflow.
>> + dev_warn(&vp_dev->pci_dev->dev,
>> + "bad notification offset %u for queue %u (> %u)",
>> + off, index, vp_dev->notify_len);
>> + err = -EINVAL;
>> + goto out_info;
>> + }
>
> I don't know if you want to limit this to "0 or power of two",
> if yes you'd do
> if (vp_dev->notify_offset_multiplier & (vp_dev->notify_offset_multiplier - 1))
> return -EINVAL;
>
>> + info->notify = vp_dev->notify_base + off * vp_dev->notify_offset_multiplier;
>>
>> info->queue = alloc_virtqueue_pages(&num);
>> if (info->queue == NULL) {
I don't see a reason to restrict it. It's not like you'll be
calculating this value more than once even in a micro implementation...
Also, let's add 2, since we write a word in there...
Cheers,
Rusty.
diff --git a/drivers/virtio/virtio_pci.c b/drivers/virtio/virtio_pci.c
index 3d0318d..1c3591a 100644
--- a/drivers/virtio/virtio_pci.c
+++ b/drivers/virtio/virtio_pci.c
@@ -502,12 +502,14 @@ static struct virtqueue *setup_vq(struct virtio_device *vdev, unsigned index,
info->msix_vector = msix_vec;
- /* get offset of notification byte for this virtqueue */
+ /* get offset of notification word for this vq (shouldn't wrap) */
off = ioread16(&vp_dev->common->queue_notify_off);
- if (off * vp_dev->notify_offset_multiplier + 2 > vp_dev->notify_len) {
+ if ((u64)off * vp_dev->notify_offset_multiplier + 2
+ > vp_dev->notify_len) {
dev_warn(&vp_dev->pci_dev->dev,
- "bad notification offset %u for queue %u (> %u)",
- off, index, vp_dev->notify_len);
+ "bad notification offset %u (x %u) for queue %u > %u",
+ off, vp_dev->notify_offset_multiplier,
+ index, vp_dev->notify_len);
err = -EINVAL;
goto out_info;
}
next prev parent reply other threads:[~2013-04-05 1:25 UTC|newest]
Thread overview: 94+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-03-21 8:29 [PATCH 00/22] New virtio PCI layout Rusty Russell
2013-03-21 8:29 ` [PATCH 01/22] virtio_config: introduce size-based accessors Rusty Russell
2013-03-21 8:29 ` [PATCH 02/22] virtio_config: use " Rusty Russell
2013-03-21 8:29 ` [PATCH 03/22] virtio_config: make transports implement accessors Rusty Russell
2013-03-21 9:09 ` Cornelia Huck
2013-03-22 0:31 ` Rusty Russell
2013-03-22 9:13 ` Cornelia Huck
2013-03-22 14:43 ` Sjur Brændeland
2013-03-24 4:24 ` Rusty Russell
2013-04-03 15:58 ` Sjur Brændeland
2013-04-02 17:16 ` Pawel Moll
2013-03-21 8:29 ` [PATCH 04/22] virtio: use u32, not bitmap for struct virtio_device's features Rusty Russell
2013-03-21 10:00 ` Cornelia Huck
2013-03-22 0:48 ` Rusty Russell
2013-03-21 8:29 ` [PATCH 05/22] virtio: add support for 64 bit features Rusty Russell
2013-03-21 10:06 ` Cornelia Huck
2013-03-22 0:50 ` Rusty Russell
2013-03-22 9:15 ` Cornelia Huck
2013-03-22 14:50 ` Sjur Brændeland
2013-03-22 20:12 ` Ohad Ben-Cohen
2013-03-25 8:30 ` Rusty Russell
2013-04-02 17:09 ` Pawel Moll
2013-03-21 8:29 ` [PATCH 06/22] virtio: move vring structure into struct virtqueue Rusty Russell
2013-03-21 8:29 ` [PATCH 07/22] pci: add pci_iomap_range Rusty Russell
2013-03-21 8:29 ` [PATCH 08/22] virtio-pci: define layout for virtio vendor-specific capabilities Rusty Russell
2013-03-21 8:29 ` [PATCH 09/22] virtio_pci: move old defines to legacy, introduce new structure Rusty Russell
2013-03-21 8:29 ` [PATCH 10/22] virtio_pci: use _LEGACY_ defines in virtio_pci_legacy.c Rusty Russell
2013-03-21 8:29 ` [PATCH 11/22] virtio_pci: don't use the legacy driver if we find the new PCI capabilities Rusty Russell
2013-03-21 8:29 ` [PATCH 12/22] virtio_pci: allow duplicate capabilities Rusty Russell
2013-03-21 10:28 ` Michael S. Tsirkin
2013-03-21 14:26 ` H. Peter Anvin
2013-03-21 14:43 ` Michael S. Tsirkin
2013-03-21 14:45 ` H. Peter Anvin
2013-03-21 15:19 ` Michael S. Tsirkin
2013-03-21 15:26 ` H. Peter Anvin
2013-03-21 15:58 ` Michael S. Tsirkin
2013-03-21 16:04 ` H. Peter Anvin
2013-03-21 16:11 ` Michael S. Tsirkin
2013-03-21 16:15 ` H. Peter Anvin
2013-03-21 16:26 ` Michael S. Tsirkin
2013-03-21 16:32 ` H. Peter Anvin
2013-03-21 17:07 ` Michael S. Tsirkin
2013-03-21 17:09 ` H. Peter Anvin
2013-03-21 17:13 ` Michael S. Tsirkin
2013-03-21 17:49 ` Michael S. Tsirkin
2013-03-21 17:54 ` H. Peter Anvin
2013-03-21 18:01 ` Michael S. Tsirkin
2013-03-22 0:57 ` Rusty Russell
2013-03-22 3:17 ` H. Peter Anvin
2013-03-24 13:14 ` Michael S. Tsirkin
2013-03-24 23:23 ` H. Peter Anvin
2013-03-25 6:53 ` Michael S. Tsirkin
2013-03-25 6:54 ` H. Peter Anvin
2013-03-25 10:03 ` Rusty Russell
2013-03-21 8:29 ` [PATCH 13/22] virtio_pci: new, capability-aware driver Rusty Russell
2013-03-21 10:24 ` Michael S. Tsirkin
2013-03-22 1:02 ` Rusty Russell
2013-03-24 13:08 ` Michael S. Tsirkin
2013-03-21 8:29 ` [PATCH 14/22] virtio_pci: layout changes as per hpa's suggestions Rusty Russell
2013-03-21 8:29 ` [PATCH 15/22] virtio_pci: use little endian for config space Rusty Russell
2013-03-21 8:29 ` [PATCH 16/22] virtio_pci: use separate notification offsets for each vq Rusty Russell
2013-03-21 10:13 ` Michael S. Tsirkin
2013-03-21 10:35 ` Michael S. Tsirkin
2013-03-22 2:52 ` Rusty Russell
2013-03-24 14:38 ` Michael S. Tsirkin
2013-03-24 20:19 ` Michael S. Tsirkin
2013-03-24 23:27 ` H. Peter Anvin
2013-03-25 7:05 ` Michael S. Tsirkin
2013-03-25 10:00 ` Rusty Russell
2013-03-26 19:39 ` Michael S. Tsirkin
2013-03-27 0:07 ` Rusty Russell
2013-03-27 0:22 ` H. Peter Anvin
2013-03-27 2:31 ` H. Peter Anvin
2013-03-27 11:26 ` Michael S. Tsirkin
2013-03-27 14:21 ` H. Peter Anvin
2013-03-27 11:25 ` Michael S. Tsirkin
2013-03-28 4:50 ` H. Peter Anvin
2013-03-30 3:19 ` Rusty Russell
2013-04-02 22:51 ` H. Peter Anvin
2013-04-03 6:10 ` Rusty Russell
2013-04-03 11:22 ` Michael S. Tsirkin
2013-04-03 14:10 ` H. Peter Anvin
2013-04-03 14:35 ` Michael S. Tsirkin
2013-04-03 14:35 ` H. Peter Anvin
2013-04-03 17:02 ` Michael S. Tsirkin
2013-04-04 5:48 ` Rusty Russell
2013-04-04 8:25 ` Michael S. Tsirkin
2013-04-05 1:25 ` Rusty Russell [this message]
2013-03-21 8:29 ` [PATCH 17/22] virtio_pci_legacy: cleanup struct virtio_pci_vq_info Rusty Russell
2013-03-21 8:29 ` [PATCH 18/22] virtio_pci: share structure between legacy and modern Rusty Russell
2013-03-21 8:29 ` [PATCH 19/22] virtio_pci: share interrupt/notify handlers " Rusty Russell
2013-03-21 8:29 ` [PATCH 20/22] virtio_pci: share virtqueue setup/teardown between modern and legacy driver Rusty Russell
2013-03-21 8:29 ` [PATCH 21/22] virtio_pci: simplify common helpers Rusty Russell
2013-03-21 8:29 ` [PATCH 22/22] virtio_pci: fix finalize_features in modern driver Rusty Russell
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=874nfln4yv.fsf@rustcorp.com.au \
--to=rusty@rustcorp.com.au \
--cc=hpa@zytor.com \
--cc=mst@redhat.com \
--cc=virtualization@lists.linux-foundation.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.