From: wangyunjian <wangyunjian@huawei.com>
To: <dev@dpdk.org>
Cc: <anatoly.burakov@intel.com>, <hemant.agrawal@nxp.com>,
<sachin.saxena@nxp.com>, <jerry.lilijun@huawei.com>,
<xudingke@huawei.com>, Yunjian Wang <wangyunjian@huawei.com>,
<stable@dpdk.org>
Subject: [dpdk-stable] [dpdk-dev] [PATCH 1/2] vfio: fix check for vfio_group_fd
Date: Tue, 19 May 2020 11:42:00 +0800 [thread overview]
Message-ID: <1589859720-16224-1-git-send-email-wangyunjian@huawei.com> (raw)
From: Yunjian Wang <wangyunjian@huawei.com>
The issue is that a file descriptor at 0 is a valid one. Currently
the file not found, the return value will be set to 0. As a result,
it is impossible to distinguish between a correct descriptor and a
failed return value. Fix it to return -ENOENT instead of 0.
Fixes: b758423bc4fe ("vfio: fix race condition with sysfs")
Fixes: ff0b67d1c868 ("vfio: DMA mappinge")
Cc: stable@dpdk.org
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
---
lib/librte_eal/linux/eal_vfio.c | 23 +++++++++++++----------
lib/librte_eal/linux/eal_vfio_mp_sync.c | 4 ++--
2 files changed, 15 insertions(+), 12 deletions(-)
diff --git a/lib/librte_eal/linux/eal_vfio.c b/lib/librte_eal/linux/eal_vfio.c
index d26e1649a..8c5a13be6 100644
--- a/lib/librte_eal/linux/eal_vfio.c
+++ b/lib/librte_eal/linux/eal_vfio.c
@@ -293,7 +293,7 @@ vfio_open_group_fd(int iommu_group_num)
strerror(errno));
return -1;
}
- return 0;
+ return -ENOENT;
}
/* noiommu group found */
}
@@ -318,12 +318,12 @@ vfio_open_group_fd(int iommu_group_num)
vfio_group_fd = mp_rep->fds[0];
} else if (p->result == SOCKET_NO_FD) {
RTE_LOG(ERR, EAL, " bad VFIO group fd\n");
- vfio_group_fd = 0;
+ vfio_group_fd = -ENOENT;
}
}
free(mp_reply.msgs);
- if (vfio_group_fd < 0)
+ if (vfio_group_fd < 0 && vfio_group_fd != -ENOENT)
RTE_LOG(ERR, EAL, " cannot request group fd\n");
return vfio_group_fd;
}
@@ -379,9 +379,9 @@ vfio_get_group_fd(struct vfio_config *vfio_cfg,
}
vfio_group_fd = vfio_open_group_fd(iommu_group_num);
- if (vfio_group_fd <= 0) {
+ if (vfio_group_fd < 0) {
RTE_LOG(ERR, EAL, "Failed to open group %d\n", iommu_group_num);
- return -1;
+ return vfio_group_fd;
}
cur_grp->group_num = iommu_group_num;
@@ -728,11 +728,14 @@ rte_vfio_setup_device(const char *sysfs_base, const char *dev_addr,
/* get the actual group fd */
vfio_group_fd = rte_vfio_get_group_fd(iommu_group_num);
- if (vfio_group_fd < 0)
+ if (vfio_group_fd < 0 && vfio_group_fd != -ENOENT)
return -1;
- /* if group_fd == 0, that means the device isn't managed by VFIO */
- if (vfio_group_fd == 0) {
+ /*
+ * if vfio_group_fd == -ENOENT, that means the device
+ * isn't managed by VFIO
+ */
+ if (vfio_group_fd == -ENOENT) {
RTE_LOG(WARNING, EAL, " %s not managed by VFIO driver, skipping\n",
dev_addr);
return 1;
@@ -955,10 +958,10 @@ rte_vfio_release_device(const char *sysfs_base, const char *dev_addr,
/* get the actual group fd */
vfio_group_fd = rte_vfio_get_group_fd(iommu_group_num);
- if (vfio_group_fd <= 0) {
+ if (vfio_group_fd < 0) {
RTE_LOG(INFO, EAL, "rte_vfio_get_group_fd failed for %s\n",
dev_addr);
- ret = -1;
+ ret = vfio_group_fd;
goto out;
}
diff --git a/lib/librte_eal/linux/eal_vfio_mp_sync.c b/lib/librte_eal/linux/eal_vfio_mp_sync.c
index 5f2a5fc1d..6254696ae 100644
--- a/lib/librte_eal/linux/eal_vfio_mp_sync.c
+++ b/lib/librte_eal/linux/eal_vfio_mp_sync.c
@@ -44,9 +44,9 @@ vfio_mp_primary(const struct rte_mp_msg *msg, const void *peer)
r->req = SOCKET_REQ_GROUP;
r->group_num = m->group_num;
fd = rte_vfio_get_group_fd(m->group_num);
- if (fd < 0)
+ if (fd < 0 && fd != -ENOENT)
r->result = SOCKET_ERR;
- else if (fd == 0)
+ else if (fd == -ENOENT)
/* if VFIO group exists but isn't bound to VFIO driver */
r->result = SOCKET_NO_FD;
else {
--
2.23.0
next reply other threads:[~2020-05-19 3:42 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-19 3:42 wangyunjian [this message]
2020-05-19 7:43 ` David Marchand
2020-05-21 12:53 ` Burakov, Anatoly
2020-05-22 7:36 ` David Marchand
2020-05-27 10:45 ` Kevin Traynor
2020-07-29 11:43 ` wangyunjian
2020-09-17 11:44 ` Burakov, Anatoly
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=1589859720-16224-1-git-send-email-wangyunjian@huawei.com \
--to=wangyunjian@huawei.com \
--cc=anatoly.burakov@intel.com \
--cc=dev@dpdk.org \
--cc=hemant.agrawal@nxp.com \
--cc=jerry.lilijun@huawei.com \
--cc=sachin.saxena@nxp.com \
--cc=stable@dpdk.org \
--cc=xudingke@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).