From: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
To: Rusty Russell <rusty@rustcorp.com.au>
Cc: netdev@vger.kernel.org, Max Krasnyansky <maxk@qualcomm.com>,
virtualization@lists.linux-foundation.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/5] /dev/vring: simple userspace-kernel ringbuffer interface.
Date: Sat, 19 Apr 2008 14:22:15 +0400 [thread overview]
Message-ID: <20080419102214.GA21952@2ka.mipt.ru> (raw)
In-Reply-To: <200804181439.49051.rusty@rustcorp.com.au>
Hi.
On Fri, Apr 18, 2008 at 02:39:48PM +1000, Rusty Russell (rusty@rustcorp.com.au) wrote:
> +int vring_get_buffer(struct vring_info *vr,
> + struct iovec *in_iov,
> + unsigned int *num_in, unsigned long *in_len,
> + struct iovec *out_iov,
> + unsigned int *num_out, unsigned long *out_len)
> +{
> + unsigned int i, in = 0, out = 0;
> + unsigned long dummy;
> + u16 avail, last_avail, head;
> + struct vring_desc d;
Should this whole function and vring_used_buffer() be protected with
vr->lock mutex?
> + if (unlikely(get_user(avail, &vr->ring.avail->idx)))
> + return -EFAULT;
> + if (unlikely(get_user(last_avail, &vring_last_avail(&vr->ring))))
> + return -EFAULT;
> +
> + if (last_avail == avail)
> + return 0;
> +
> + if (!in_len)
> + in_len = &dummy;
> + if (!out_len)
> + out_len = &dummy;
> +
> + *in_len = *out_len = 0;
> +
> + if (unlikely(get_user(head, &vr->ring.avail->ring[last_avail
> + & vr->mask])))
> + return -EFAULT;
> +
> + i = head;
> + do {
> + if (unlikely(i >= vr->ring.num)) {
> + pr_debug("vring: bad index: %u\n", i);
> + return -EINVAL;
> + }
> +
> + if (copy_from_user(&d, &vr->ring.desc[i], sizeof(d)) != 0)
> + return -EFAULT;
> +
> + if (d.flags & VRING_DESC_F_WRITE) {
> + /* Check for length and iovec overflows */
> + if (!num_in) {
> + pr_debug("vring: writable desc %u in ring %p\n",
> + i, vr->ring.desc);
> + return -EINVAL;
> + }
> + if (in == *num_in || *in_len + d.len < *in_len)
> + return -E2BIG;
> + in_iov[in].iov_len = d.len;
> + *in_len += d.len;
> + in_iov[in].iov_base = (void __user *)(long)d.addr;
> + in++;
> + } else {
> + if (!num_out) {
> + pr_debug("vring: readable desc %u in ring %p\n",
> + i, vr->ring.desc);
> + return -EINVAL;
> + }
> + if (out == *num_out || *out_len + d.len < *out_len)
> + return -E2BIG;
> + out_iov[out].iov_len = d.len;
> + *out_len += d.len;
> + out_iov[out].iov_base = (void __user *)(long)d.addr;
> + out++;
> + }
> +
> + i = d.next;
> + } while (d.flags & VRING_DESC_F_NEXT);
> +
> + if (num_in)
> + *num_in = in;
> + if (num_out)
> + *num_out = out;
> +
> + last_avail++;
> + put_user(last_avail, &vring_last_avail(&vr->ring));
> +
> + /* 0 is a valid head, so add one. */
> + return head + 1;
> +}
> +EXPORT_SYMBOL_GPL(vring_get_buffer);
> +
> +/**
> + * vring_used_buffer - return a used buffer to the vring
> + * @vr: the vring
> + * @id: the id returned from vring_get_buffer
> + * @len: the total bytes *written* to the buffer
> + */
> +void vring_used_buffer(struct vring_info *vr, int id, u32 len)
> +{
> + struct vring_used_elem used;
> + u16 used_idx;
> +
> + BUG_ON(id <= 0 || id > vr->ring.num);
> +
> + used.id = id - 1;
> + used.len = len;
> + if (get_user(used_idx, &vr->ring.used->idx) != 0)
> + return;
> +
> + if (copy_to_user(&vr->ring.used->ring[used_idx & vr->mask], &used,
> + sizeof(used)))
> + return;
> +
> + wmb();
> + used_idx++;
> + put_user(used_idx, &vr->ring.used->idx);
> +}
> +EXPORT_SYMBOL_GPL(vring_used_buffer);
--
Evgeniy Polyakov
next prev parent reply other threads:[~2008-04-19 10:23 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-04-18 4:33 [PATCH 0/5] High-speed tun receive and xmit Rusty Russell
2008-04-18 4:35 ` [PATCH 1/5] virtio: put last_used and last_avail index into ring itself Rusty Russell
2008-04-18 4:39 ` [PATCH 2/5] /dev/vring: simple userspace-kernel ringbuffer interface Rusty Russell
2008-04-18 4:41 ` [PATCH 3/5] /dev/vring limit and base ioctls Rusty Russell
2008-04-18 4:42 ` [PATCH 4/5] tun: vringfd receive support Rusty Russell
2008-04-18 4:43 ` [PATCH 5/5] tun: vringfd xmit support Rusty Russell
2008-04-18 11:31 ` Andrew Morton
2008-04-18 15:15 ` Rusty Russell
2008-04-18 16:24 ` Ray Lee
2008-04-18 19:06 ` Andrew Morton
2008-04-19 14:41 ` Rusty Russell
2008-04-19 17:51 ` Andrew Morton
2008-04-19 1:54 ` Andrew Morton
2008-04-18 11:46 ` pradeep singh rautela
2008-04-18 14:25 ` Ray Lee
2008-04-18 18:01 ` pradeep singh rautela
2008-04-18 11:18 ` [PATCH 2/5] /dev/vring: simple userspace-kernel ringbuffer interface Andrew Morton
2008-04-18 14:32 ` Rusty Russell
2008-04-18 18:59 ` Andrew Morton
2008-04-18 19:38 ` Michael Kerrisk
2008-04-19 16:41 ` Rusty Russell
2008-04-20 0:16 ` David Miller
2008-04-19 15:02 ` Jonathan Corbet
2008-04-19 10:22 ` Evgeniy Polyakov [this message]
2008-04-19 16:05 ` Rusty Russell
2008-04-19 16:33 ` Evgeniy Polyakov
2008-04-19 16:45 ` 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=20080419102214.GA21952@2ka.mipt.ru \
--to=johnpol@2ka.mipt.ru \
--cc=linux-kernel@vger.kernel.org \
--cc=maxk@qualcomm.com \
--cc=netdev@vger.kernel.org \
--cc=rusty@rustcorp.com.au \
--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 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).