DPDK patches and discussions
 help / color / mirror / Atom feed
From: Thomas Monjalon <thomas.monjalon@6wind.com>
To: zhida zang <zhida.zang@intel.com>
Cc: dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] scripts: enable extended tag of PCIe
Date: Mon, 23 Mar 2015 12:52:31 +0100	[thread overview]
Message-ID: <3200414.B2YSnFKgSE@xps13> (raw)
In-Reply-To: <1422593822-15531-1-git-send-email-zhida.zang@intel.com>

Hi,

This patch needs review and documentation.
It's going to be dropped if nobody cares.

There were some previous discussions about it:
	http://dpdk.org/ml/archives/dev/2015-February/012708.html


2015-01-30 12:57, zhida zang:
> As 'extended tag' of PCIe needs to be enabled for i40e high performance,
> Linux command of 'setpci' can be used to check and set the corresponding 
> bit of 'extended tag' of PCIe configuration space. The script is to check
> and set the right bit in PCIe configuration space to enable 'extended tag'.
> 
> Signed-off-by: Zhida Zang <zhida.zang@intel.com>
> ---
>  tools/set_pci.py | 124 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 124 insertions(+)
>  create mode 100755 tools/set_pci.py
> 
> diff --git a/tools/set_pci.py b/tools/set_pci.py
> new file mode 100755
> index 0000000..e242efb
> --- /dev/null
> +++ b/tools/set_pci.py
> @@ -0,0 +1,124 @@
> +#! /usr/bin/python
> +import sys
> +import os
> +import subprocess
> +import getopt
> +from os.path import basename
> +
> +# The register to check if extended tag is supported or not.
> +PCI_DEV_CAP_REG = 0xA4
> +# The control register which contains the bit to enable/disable 'extended tag'.
> +PCI_DEV_CTRL_REG = 0xA8
> +# The mask of 'extended tag' in capability register.
> +PCI_DEV_CAP_EXT_TAG_MASK = 0x20
> +# The mask of 'extended tag' in control register.
> +PCI_DEV_CTRL_EXT_TAG_MASK = 0x100
> +
> +dev_ids = {}
> +flag = "Set"
> +
> +
> +def usage():
> +    '''Print usage information for the program'''
> +    argv0 = basename(sys.argv[0])
> +    print """
> +Usage:
> +------
> +
> +    %(argv0)s [options] DEVICE1 DEVICE2 ....
> +
> +where DEVICE1, DEVICE2 etc, are specified via PCI "domain:bus:slot.func" syntax
> +or "bus:slot.func" syntax. For devices bound to Linux kernel drivers, they may
> +also be referred to by Linux interface name e.g. eth0, eth1, em0, em1, etc.
> +
> +Options:
> +    --help, --usage:
> +        Display usage information and quit
> +
> +    -s --set:
> +        Set the following pci device
> +
> +    -u --Unset:
> +        Unset the following pci device
> +
> +Examples:
> +---------
> +To set pci 0a:00.0
> +        %(argv0)s -s 0a:00.0
> +        %(argv0)s --set 0a:00.0
> +
> +To unset 0000:01:00.0
> +        %(argv0)s -u 0000:01:00.0
> +        %(argv0)s --unset 0000:01:00.0
> +
> +To set 0000:02:00.0 and 0000:02:00.1
> +        %(argv0)s -s 02:00.0 02:00.1
> +
> +    """ % locals()  # replace items from local variables
> +
> +
> +def parse_args():
> +    global flag
> +    global dev_ids
> +    if len(sys.argv) <= 1:
> +        usage()
> +        sys.exit(0)
> +    try:
> +        opts, dev_ids = getopt.getopt(
> +            sys.argv[1:],
> +            "su",
> +            ["help", "usage", "set", "unset"]
> +            )
> +    except getopt.GetoptError, error:
> +        print str(error)
> +        print "Run '%s --usage' for further information" % sys.argv[0]
> +        sys.exit(1)
> +
> +    for opt, arg in opts:
> +        if opt == "--help" or opt == "--usage":
> +            usage()
> +            sys.exit(0)
> +        if opt == "-s" or opt == "--set":
> +            flag = "Set"
> +        if opt == "-u" or opt == "--unset":
> +            flag = "Unset"
> +
> +
> +def check_output(args, stderr=None):
> +    '''Run a command and capture its output'''
> +    return subprocess.Popen(
> +        args,
> +        stdout=subprocess.PIPE,
> +        stderr=stderr
> +        ).communicate()[0]
> +
> +
> +def set_pci():
> +    if len(dev_ids) == 0:
> +        print "Error: No devices specified."
> +        print "Run '%s --usage' for further information" % sys.argv[0]
> +        sys.exit(1)
> +    param_cap = "%x.W" % PCI_DEV_CAP_REG
> +    for k in range(len(dev_ids)):
> +        val = check_output(["setpci", "-s", dev_ids[k], param_cap])
> +        if (not (int(val, 16) & PCI_DEV_CAP_EXT_TAG_MASK)):
> +            print dev_ids[k], "Not supported"
> +            continue
> +        if (int(val, 16) & PCI_DEV_CTRL_EXT_TAG_MASK):
> +            continue
> +        param_ctrl = "%x.W" % PCI_DEV_CTRL_REG
> +        val = check_output(["setpci", "-s", dev_ids[k], param_ctrl])
> +        if flag == "Set":
> +            val = int(val, 16) | PCI_DEV_CTRL_EXT_TAG_MASK
> +        else:
> +            val = int(val, 16) & ~PCI_DEV_CTRL_EXT_TAG_MASK
> +        param_ctrl = "%x.W=%x" % (PCI_DEV_CTRL_REG, val)
> +        check_output(["setpci", "-s", dev_ids[k], param_ctrl])
> +
> +
> +def main():
> +    parse_args()
> +    set_pci()
> +
> +if __name__ == "__main__":
> +    main()
> 

  reply	other threads:[~2015-03-23 11:53 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <setpciv1>
2015-01-30  4:57 ` zhida zang
2015-03-23 11:52   ` Thomas Monjalon [this message]
2015-03-24  1:07     ` Zhang, Helin
2015-03-25  3:32       ` Liu, Yong
2015-03-30  5:30         ` Lu, Patrick
2015-03-30  6:11           ` Zhang, Helin

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=3200414.B2YSnFKgSE@xps13 \
    --to=thomas.monjalon@6wind.com \
    --cc=dev@dpdk.org \
    --cc=zhida.zang@intel.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).