From: David Marchand <david.marchand@6wind.com>
To: Stephen Hemminger <stephen@networkplumber.org>
Cc: "dev@dpdk.org" <dev@dpdk.org>, Stephen Hemminger <shemming@brocade.com>
Subject: Re: [dpdk-dev] [PATCH 1/4] pci: allow access to PCI config space
Date: Mon, 9 Feb 2015 10:45:01 +0100 [thread overview]
Message-ID: <CALwxeUu7Ur-yXMcKu7Zkq16zqR6f5vU8DEX4Z8CET9G2ttjKgA@mail.gmail.com> (raw)
In-Reply-To: <1423247795-22399-2-git-send-email-stephen@networkplumber.org>
Hello Stephen,
- It looks a bit odd to me, we end up with something asymetric between uio
/ vfio wrt pci config space.
Can we an api consistent between the two ?
Does this mean that your pmd cannot work / has not been used with vfio ?
- Anyway, I suppose we could reuse this api to remove the RTE_PCI_CONFIG
#ifdef / hardcoded stuff from linux eal / igb_uio.
Opinion ?
--
David Marchand
On Fri, Feb 6, 2015 at 7:36 PM, Stephen Hemminger <
stephen@networkplumber.org> wrote:
> From: Stephen Hemminger <shemming@brocade.com>
>
> Some drivers need ability to access PCI config (for example for power
> management). This adds an abstraction to do this; only implemented
> on Linux, but should be possible on BSD.
>
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> ---
> lib/librte_eal/common/include/rte_pci.h | 29
> +++++++++++++++++++++++++
> lib/librte_eal/linuxapp/eal/eal_pci.c | 15 +++++++++++++
> lib/librte_eal/linuxapp/eal/eal_pci_uio.c | 10 +++++++++
> lib/librte_eal/linuxapp/eal/rte_eal_version.map | 2 ++
> 4 files changed, 56 insertions(+)
>
> diff --git a/lib/librte_eal/common/include/rte_pci.h
> b/lib/librte_eal/common/include/rte_pci.h
> index 66ed793..a78081f 100644
> --- a/lib/librte_eal/common/include/rte_pci.h
> +++ b/lib/librte_eal/common/include/rte_pci.h
> @@ -152,6 +152,7 @@ struct rte_pci_device {
> uint16_t max_vfs; /**< sriov enable if not
> zero */
> int numa_node; /**< NUMA node connection
> */
> struct rte_devargs *devargs; /**< Device user arguments
> */
> + int config_fd; /**< PCI config access */
> };
>
> /** Any PCI device identifier (vendor, device, ...) */
> @@ -298,6 +299,34 @@ void rte_eal_pci_register(struct rte_pci_driver
> *driver);
> */
> void rte_eal_pci_unregister(struct rte_pci_driver *driver);
>
> +/**
> + * Read PCI config space.
> + *
> + * @param device
> + * A pointer to a rte_pci_device structure describing the device
> + * to use
> + * @param buf
> + * A data buffer where the bytes should be read into
> + * @param size
> + * The length of the data buffer.
> + */
> +int rte_eal_pci_read_config(const struct rte_pci_device *device,
> + void *buf, size_t len, off_t offset);
> +
> +/**
> + * Write PCI config space.
> + *
> + * @param device
> + * A pointer to a rte_pci_device structure describing the device
> + * to use
> + * @param buf
> + * A data buffer containing the bytes should be written
> + * @param size
> + * The length of the data buffer.
> + */
> +int rte_eal_pci_write_config(const struct rte_pci_device *device,
> + const void *buf, size_t len, off_t offset);
> +
> #ifdef __cplusplus
> }
> #endif
> diff --git a/lib/librte_eal/linuxapp/eal/eal_pci.c
> b/lib/librte_eal/linuxapp/eal/eal_pci.c
> index b5f5410..5bcfffa 100644
> --- a/lib/librte_eal/linuxapp/eal/eal_pci.c
> +++ b/lib/librte_eal/linuxapp/eal/eal_pci.c
> @@ -233,6 +233,7 @@ pci_scan_one(const char *dirname, uint16_t domain,
> uint8_t bus,
> dev->addr.bus = bus;
> dev->addr.devid = devid;
> dev->addr.function = function;
> + dev->config_fd = -1;
>
> /* get vendor id */
> snprintf(filename, sizeof(filename), "%s/vendor", dirname);
> @@ -592,6 +593,20 @@ rte_eal_pci_probe_one_driver(struct rte_pci_driver
> *dr, struct rte_pci_device *d
> return 1;
> }
>
> +/* Read PCI config space. */
> +int rte_eal_pci_read_config(const struct rte_pci_device *device,
> + void *buf, size_t len, off_t offset)
> +{
> + return pread(device->config_fd, buf, len, offset);
> +}
> +
> +/* Write PCI config space. */
> +int rte_eal_pci_write_config(const struct rte_pci_device *device,
> + const void *buf, size_t len, off_t offset)
> +{
> + return pwrite(device->config_fd, buf, len, offset);
> +}
> +
> /* Init the PCI EAL subsystem */
> int
> rte_eal_pci_init(void)
> diff --git a/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> b/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> index e53f06b..0396eaa 100644
> --- a/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> +++ b/lib/librte_eal/linuxapp/eal/eal_pci_uio.c
> @@ -213,6 +213,7 @@ pci_get_uio_dev(struct rte_pci_device *dev, char
> *dstbuf,
> struct dirent *e;
> DIR *dir;
> char dirname[PATH_MAX];
> + char filename[PATH_MAX];
>
> /* depending on kernel version, uio can be located in uio/uioX
> * or uio:uioX */
> @@ -268,6 +269,15 @@ pci_get_uio_dev(struct rte_pci_device *dev, char
> *dstbuf,
> if (e == NULL)
> return -1;
>
> + /* Open fd for access to PCI config */
> + snprintf(filename, sizeof(filename), "%s/device/config", dirname);
> + dev->config_fd = open(filename, O_RDWR);
> + if (dev->config_fd < 0) {
> + RTE_LOG(ERR, EAL, "%s(): cannot open %s: %s\n",
> + __func__, filename, strerror(errno));
> + return -1;
> + }
> +
> /* create uio device if we've been asked to */
> if (internal_config.create_uio_dev &&
> pci_mknod_uio_dev(dstbuf, uio_num) < 0)
> diff --git a/lib/librte_eal/linuxapp/eal/rte_eal_version.map
> b/lib/librte_eal/linuxapp/eal/rte_eal_version.map
> index d36286e..626689c 100644
> --- a/lib/librte_eal/linuxapp/eal/rte_eal_version.map
> +++ b/lib/librte_eal/linuxapp/eal/rte_eal_version.map
> @@ -40,6 +40,8 @@ DPDK_2.0 {
> rte_eal_pci_probe;
> rte_eal_pci_register;
> rte_eal_pci_unregister;
> + rte_eal_pci_read_config;
> + rte_eal_pci_write_config;
> rte_eal_process_type;
> rte_eal_remote_launch;
> rte_eal_tailq_lookup;
> --
> 2.1.4
>
>
next prev parent reply other threads:[~2015-02-09 9:45 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-02-06 18:36 [dpdk-dev] [PATCH 0/4] Broadcom 10G NIC Poll Mode Driver Stephen Hemminger
2015-02-06 18:36 ` [dpdk-dev] [PATCH 1/4] pci: allow access to PCI config space Stephen Hemminger
2015-02-09 9:45 ` David Marchand [this message]
2015-02-10 23:49 ` Stephen Hemminger
[not found] ` <c058fdefd5564f3f96bd21593c9ea19a@BRMWP-EXMB11.corp.brocade.com>
2015-02-11 1:23 ` Stephen Hemminger
2015-02-06 18:36 ` [dpdk-dev] [PATCH 2/4] bcm: add BCM pci device ids Stephen Hemminger
2015-02-06 18:36 ` [dpdk-dev] [PATCH 3/4] bcm: new poll mode driver Stephen Hemminger
2015-02-06 18:36 ` [dpdk-dev] [PATCH 4/4] bcm: enable BCM poll mode driver in config Stephen Hemminger
2015-02-07 9:15 ` [dpdk-dev] [PATCH 0/4] Broadcom 10G NIC Poll Mode Driver Jun Xiao
2015-02-07 15:15 ` Stephen Hemminger
2015-02-07 16:05 ` Jun Xiao
2015-02-07 22:23 ` Stephen Hemminger
2015-02-12 14:00 ` Thomas Monjalon
2015-02-12 14:33 ` Stephen Hemminger
2015-02-12 14:56 ` Thomas Monjalon
[not found] ` <0a8c8ae93e354bf0a74dd8ade9d1cc70@BRMWP-EXMB11.corp.brocade.com>
2015-02-12 21:45 ` Stephen Hemminger
2015-02-13 8:31 ` Thomas Monjalon
2015-05-07 23:25 [dpdk-dev] [PATCH 0/4 v4] bnx2x: new poll mode driver Stephen Hemminger
2015-05-07 23:25 ` [dpdk-dev] [PATCH 1/4] pci: allow access to PCI config space Stephen Hemminger
2015-05-11 12:54 ` Neil Horman
[not found] ` <38426478085b4e779e18967cd1b6ae4f@BRMWP-EXMB11.corp.brocade.com>
2015-05-11 15:23 ` Stephen Hemminger
2015-05-11 15:37 ` Neil Horman
2015-05-11 17:30 ` Stephen Hemminger
2015-05-11 17:31 ` Stephen Hemminger
2015-05-12 9:56 ` Bruce Richardson
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=CALwxeUu7Ur-yXMcKu7Zkq16zqR6f5vU8DEX4Z8CET9G2ttjKgA@mail.gmail.com \
--to=david.marchand@6wind.com \
--cc=dev@dpdk.org \
--cc=shemming@brocade.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).