From: Jan Viktorin <viktorin@rehivetech.com>
To: dev@dpdk.org
Cc: Jan Viktorin <viktorin@rehivetech.com>,
Anatoly Burakov <anatoly.burakov@intel.com>,
David Marchand <david.marchand@6wind.com>,
Keith Wiles <keith.wiles@intel.com>,
Santosh Shukla <sshukla@mvista.com>,
Stephen Hemminger <stephen@networkplumber.org>,
Shreyansh Jain <shreyansh.jain@nxp.com>
Subject: [dpdk-dev] [PATCH v2 10/16] vfio: extract setup logic out of pci_vfio_map_resource
Date: Mon, 13 Jun 2016 15:02:00 +0200 [thread overview]
Message-ID: <1465822926-23742-11-git-send-email-viktorin@rehivetech.com> (raw)
In-Reply-To: <1465822926-23742-1-git-send-email-viktorin@rehivetech.com>
In-Reply-To: <1461937456-22943-1-git-send-email-viktorin@rehivetech.com>
The setup logic access the global vfio_cfg variable that will be moved in the
following commits. We need to separate all accesses to this variable to a
general code.
Signed-off-by: Jan Viktorin <viktorin@rehivetech.com>
---
v2
* fixed lost viable code (A. Burakov)
---
lib/librte_eal/linuxapp/eal/eal_pci_vfio.c | 85 +++++++++++++++++-------------
1 file changed, 49 insertions(+), 36 deletions(-)
diff --git a/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c b/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c
index 2b3dd2e..27ce6a8 100644
--- a/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c
+++ b/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c
@@ -411,37 +411,22 @@ clear_current_group(void)
vfio_cfg.vfio_groups[vfio_cfg.vfio_group_idx].fd = -1;
}
-
-/*
- * map the PCI resources of a PCI device in virtual memory (VFIO version).
- * primary and secondary processes follow almost exactly the same path
+/**
+ * Setup vfio_cfg for the device indentified by its address. It discovers
+ * the configured I/O MMU groups or sets a new one for the device. If a new
+ * groups is assigned, the DMA mapping is performed.
+ * Returns 0 on success, a negative value on failure and a positive value in
+ * case the given device cannot be managed this way.
*/
-int
-pci_vfio_map_resource(struct rte_pci_device *dev)
+static int pci_vfio_setup_device(const char *pci_addr, int *vfio_dev_fd,
+ struct vfio_device_info *device_info)
{
struct vfio_group_status group_status = {
.argsz = sizeof(group_status)
};
- struct vfio_device_info device_info = { .argsz = sizeof(device_info) };
- int vfio_group_fd, vfio_dev_fd;
+ int vfio_group_fd;
int iommu_group_no;
- char pci_addr[PATH_MAX] = {0};
- struct rte_pci_addr *loc = &dev->addr;
- int i, ret, msix_bar;
- struct mapped_pci_resource *vfio_res = NULL;
- struct mapped_pci_res_list *vfio_res_list = RTE_TAILQ_CAST(rte_vfio_tailq.head, mapped_pci_res_list);
-
- struct pci_map *maps;
- uint32_t msix_table_offset = 0;
- uint32_t msix_table_size = 0;
- uint32_t ioport_bar;
-
- dev->intr_handle.fd = -1;
- dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
-
- /* store PCI address string */
- snprintf(pci_addr, sizeof(pci_addr), PCI_PRI_FMT,
- loc->domain, loc->bus, loc->devid, loc->function);
+ int ret;
/* get group number */
ret = pci_vfio_get_group_no(pci_addr, &iommu_group_no);
@@ -476,8 +461,8 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
}
/*
- * at this point, we know at least one port on this device is bound to VFIO,
- * so we can proceed to try and set this particular port up
+ * at this point, we know that this group is viable (meaning, all devices
+ * are either bound to VFIO or not bound to anything)
*/
/* check if the group is viable */
@@ -495,11 +480,6 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
return -1;
}
- /*
- * at this point, we know that this group is viable (meaning, all devices
- * are either bound to VFIO or not bound to anything)
- */
-
/* check if group does not have a container yet */
if (!(group_status.flags & VFIO_GROUP_FLAGS_CONTAINER_SET)) {
@@ -546,8 +526,8 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
}
/* get a file descriptor for the device */
- vfio_dev_fd = ioctl(vfio_group_fd, VFIO_GROUP_GET_DEVICE_FD, pci_addr);
- if (vfio_dev_fd < 0) {
+ *vfio_dev_fd = ioctl(vfio_group_fd, VFIO_GROUP_GET_DEVICE_FD, pci_addr);
+ if (*vfio_dev_fd < 0) {
/* if we cannot get a device fd, this simply means that this
* particular port is not bound to VFIO
*/
@@ -557,14 +537,47 @@ pci_vfio_map_resource(struct rte_pci_device *dev)
}
/* test and setup the device */
- ret = ioctl(vfio_dev_fd, VFIO_DEVICE_GET_INFO, &device_info);
+ ret = ioctl(*vfio_dev_fd, VFIO_DEVICE_GET_INFO, device_info);
if (ret) {
RTE_LOG(ERR, EAL, " %s cannot get device info, "
"error %i (%s)\n", pci_addr, errno, strerror(errno));
- close(vfio_dev_fd);
+ close(*vfio_dev_fd);
return -1;
}
+ return 0;
+}
+
+/*
+ * map the PCI resources of a PCI device in virtual memory (VFIO version).
+ * primary and secondary processes follow almost exactly the same path
+ */
+int
+pci_vfio_map_resource(struct rte_pci_device *dev)
+{
+ struct vfio_device_info device_info = { .argsz = sizeof(device_info) };
+ char pci_addr[PATH_MAX] = {0};
+ int vfio_dev_fd;
+ struct rte_pci_addr *loc = &dev->addr;
+ int i, ret, msix_bar;
+ struct mapped_pci_resource *vfio_res = NULL;
+ struct mapped_pci_res_list *vfio_res_list = RTE_TAILQ_CAST(rte_vfio_tailq.head, mapped_pci_res_list);
+
+ struct pci_map *maps;
+ uint32_t msix_table_offset = 0;
+ uint32_t msix_table_size = 0;
+ uint32_t ioport_bar;
+
+ dev->intr_handle.fd = -1;
+ dev->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
+
+ /* store PCI address string */
+ snprintf(pci_addr, sizeof(pci_addr), PCI_PRI_FMT,
+ loc->domain, loc->bus, loc->devid, loc->function);
+
+ if ((ret = pci_vfio_setup_device(pci_addr, &vfio_dev_fd, &device_info)))
+ return ret;
+
/* get MSI-X BAR, if any (we have to know where it is because we can't
* easily mmap it when using VFIO) */
msix_bar = -1;
--
2.8.0
next prev parent reply other threads:[~2016-06-13 13:07 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-29 13:44 [dpdk-dev] [PATCH 00/15] Make VFIO support independent on PCI Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 01/15] vfio: fix include of eal_private.h to be local Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 02/15] vfio: move VFIO-specific stuff to eal_vfio.h Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 03/15] vfio: move common vfio constants " Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 04/15] vfio: move vfio_iommu_type and dma_map functions to eal_vfio Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 05/15] vfio: generalize pci_vfio_set_iommu_type Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 06/15] vfio: generalize pci_vfio_has_supported_extensions Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 07/15] vfio: move vfio-specific SOCKET_* constants Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 08/15] vfio: generalize pci_vfio_get_container_fd Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 09/15] vfio: generalize pci_vfio_get_group_no Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 10/15] vfio: extract setup logic out of pci_vfio_map_resource Jan Viktorin
2016-05-10 11:53 ` Burakov, Anatoly
2016-05-10 12:58 ` Jan Viktorin
2016-05-10 13:17 ` Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 11/15] vfio: move global vfio_cfg to eal_vfio.c Jan Viktorin
2016-05-10 11:56 ` Burakov, Anatoly
2016-05-10 12:55 ` Jan Viktorin
2016-05-13 15:34 ` Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 12/15] vfio: make vfio_*_dma_map and iommu_types private Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 13/15] vfio: rename and generalize eal_pci_vfio_mp_sync Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 14/15] vfio: initialize vfio out of the PCI subsystem Jan Viktorin
2016-04-29 13:44 ` [dpdk-dev] [PATCH 15/15] vfio: change VFIO init to be extendable Jan Viktorin
2016-05-10 11:50 ` Burakov, Anatoly
2016-05-10 12:54 ` Jan Viktorin
2016-05-10 13:15 ` Burakov, Anatoly
2016-05-10 12:18 ` [dpdk-dev] [PATCH 00/15] Make VFIO support independent on PCI Santosh Shukla
2016-05-10 12:45 ` Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 00/16] Make VFIO support less dependent " Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 01/16] vfio: fix include of eal_private.h to be local Jan Viktorin
2016-07-04 10:22 ` Burakov, Anatoly
2016-07-04 10:55 ` Jan Viktorin
2016-07-04 15:00 ` Jan Viktorin
2016-07-04 15:09 ` Burakov, Anatoly
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 02/16] vfio: move VFIO-specific stuff to eal_vfio.h Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 03/16] vfio: move common vfio constants " Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 04/16] vfio: move vfio_iommu_type and dma_map functions to eal_vfio Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 05/16] vfio: generalize pci_vfio_set_iommu_type Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 06/16] vfio: generalize pci_vfio_has_supported_extensions Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 07/16] vfio: move vfio-specific SOCKET_* constants Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 08/16] vfio: generalize pci_vfio_get_container_fd Jan Viktorin
2016-06-13 13:01 ` [dpdk-dev] [PATCH v2 09/16] vfio: generalize pci_vfio_get_group_no Jan Viktorin
2016-06-13 13:02 ` Jan Viktorin [this message]
2016-06-13 13:02 ` [dpdk-dev] [PATCH v2 11/16] vfio: move global vfio_cfg to eal_vfio.c Jan Viktorin
2016-06-13 13:02 ` [dpdk-dev] [PATCH v2 12/16] vfio: fix typo in doc for vfio_setup_device Jan Viktorin
2016-06-24 15:55 ` Mcnamara, John
2016-06-13 13:02 ` [dpdk-dev] [PATCH v2 13/16] vfio: make vfio_*_dma_map and iommu_types private Jan Viktorin
2016-06-13 13:02 ` [dpdk-dev] [PATCH v2 14/16] vfio: rename and generalize eal_pci_vfio_mp_sync Jan Viktorin
2016-06-13 13:02 ` [dpdk-dev] [PATCH v2 15/16] vfio: initialize vfio out of the PCI subsystem Jan Viktorin
2016-06-13 13:02 ` [dpdk-dev] [PATCH v2 16/16] vfio: change VFIO init to be extendable Jan Viktorin
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=1465822926-23742-11-git-send-email-viktorin@rehivetech.com \
--to=viktorin@rehivetech.com \
--cc=anatoly.burakov@intel.com \
--cc=david.marchand@6wind.com \
--cc=dev@dpdk.org \
--cc=keith.wiles@intel.com \
--cc=shreyansh.jain@nxp.com \
--cc=sshukla@mvista.com \
--cc=stephen@networkplumber.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).