From: "Gaëtan Rivet" <gaetan.rivet@6wind.com>
To: "Tan, Jianfeng" <jianfeng.tan@intel.com>
Cc: dev@dpdk.org, santosh.shukla@caviumnetworks.com,
jerin.jacob@caviumnetworks.com, anatoly.burakov@intel.com
Subject: Re: [dpdk-dev] [PATCH] pci: fix check uio bind
Date: Fri, 20 Oct 2017 22:08:22 +0200 [thread overview]
Message-ID: <20171020200822.GJ3596@bidouze.vm.6wind.com> (raw)
In-Reply-To: <38a126ec-7a6a-b43f-88b1-67e3b6477369@intel.com>
On Sat, Oct 21, 2017 at 12:47:14AM +0800, Tan, Jianfeng wrote:
> Hi Gaëtan,
>
>
> On 10/19/2017 7:42 PM, Gaëtan Rivet wrote:
> >Hi Jianfeng,
> >
> >On Thu, Oct 19, 2017 at 11:18:29AM +0000, Jianfeng Tan wrote:
> >>When checking if any devices bound to uio, we did not exclud
> >>those which are blacklisted (or in the case that a whitelist
> >>is specified).
> >>
> >>This patch fixes it by only checking whitelisted devices.
> >>
> >>Fixes: 815c7deaed2d ("pci: get IOMMU class on Linux")
> >>
> >>Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
> >>---
> >> lib/librte_eal/linuxapp/eal/eal_pci.c | 18 ++++++++++++++++++
> >> 1 file changed, 18 insertions(+)
> >>
> >>diff --git a/lib/librte_eal/linuxapp/eal/eal_pci.c b/lib/librte_eal/linuxapp/eal/eal_pci.c
> >>index b4dbf95..2b23d67 100644
> >>--- a/lib/librte_eal/linuxapp/eal/eal_pci.c
> >>+++ b/lib/librte_eal/linuxapp/eal/eal_pci.c
> >>@@ -516,8 +516,26 @@ static inline int
> >> pci_one_device_bound_uio(void)
> >> {
> >> struct rte_pci_device *dev = NULL;
> >>+ struct rte_devargs *devargs;
> >>+ int check_all = 1;
> >>+ int need_check;
> >>+
> >>+ if (rte_pci_bus.bus.conf.scan_mode == RTE_BUS_SCAN_WHITELIST)
> >>+ check_all = 0;
> >> FOREACH_DEVICE_ON_PCIBUS(dev) {
> >>+ devargs = dev->device.devargs;
> >>+
> >>+ need_check = 0;
> >>+ if (check_all)
> >Unless I'm mistaken, you will check blacklisted devices as well here.
>
> Thank you for pointing out this.
>
> I was referring to rte_pci_probe(), which also only check "probe_all" and
> (devargs && RTE_DEV_WHITELISTED); but turns out it double checks the
> blacklisted devices in rte_pci_probe_one_driver().
>
> I'll fix it.
>
> >The condition should be something like:
> >
> >if (check_all && devargs == NULL)
>
> >Which means that both ifs can be refactored as
> >
> >if ((check_all ^ (devargs != NULL)) == 0)
> > continue;
> >
> >Removing need_check. But it can be hard to read.
>
> Yes, I prefer to make it easy to understand. Please let me know if you are
> OK with below code (remove check_all):
>
> FOREACH_DEVICE_ON_PCIBUS(dev) {
> devargs = dev->device.devargs;
>
> need_check = 0;
> switch (rte_pci_bus.bus.conf.scan_mode) {
> case RTE_BUS_SCAN_UNDEFINED:
> need_check = 1;
> break;
> case RTE_BUS_SCAN_WHITELIST:
> if (devargs && devargs->policy ==
> RTE_DEV_WHITELISTED)
> need_check = 1;
> break;
> case RTE_BUS_SCAN_BLACKLIST:
> if (!devargs || devargs->policy !=
> RTE_DEV_BLACKLISTED)
> need_check = 1;
> break;
> }
>
> if (!need_check)
> continue;
> ...
I like the switch, two remarks however:
1. The SCAN_UNDEFINED basically means blacklist mode for the PCI bus.
This is the reason probe_all was set by testing for WHITELIST
mode: either of the other too would thus trigger the blacklist
behavior.
Thus, I think you could write a fallthrough case for UNDEFINED, that
would go into the BLACKLIST mode.
2. For pointers in general I would test against NULL instead of using
the unary '!'.
I think it is the general policy in DPDK to always explicitly check
against the constant value, but I personally think that for booleans
like need_check the "not" operator is ok.
So I will only highlight the !devargs :)
>
> Thanks,
> Jianfeng
--
Gaëtan Rivet
6WIND
next prev parent reply other threads:[~2017-10-20 20:08 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-19 11:18 Jianfeng Tan
2017-10-19 11:42 ` Gaëtan Rivet
2017-10-20 16:47 ` Tan, Jianfeng
2017-10-20 20:08 ` Gaëtan Rivet [this message]
2017-10-23 3:20 ` Tan, Jianfeng
2017-10-24 7:44 ` [dpdk-dev] [PATCH v2] " Jianfeng Tan
2017-10-24 8:25 ` Gaëtan Rivet
2017-10-26 21:48 ` Thomas Monjalon
2017-10-24 8:31 ` santosh
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=20171020200822.GJ3596@bidouze.vm.6wind.com \
--to=gaetan.rivet@6wind.com \
--cc=anatoly.burakov@intel.com \
--cc=dev@dpdk.org \
--cc=jerin.jacob@caviumnetworks.com \
--cc=jianfeng.tan@intel.com \
--cc=santosh.shukla@caviumnetworks.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).