DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Burakov, Anatoly" <anatoly.burakov@intel.com>
To: Thomas Monjalon <thomas@monjalon.net>
Cc: Takeshi Yoshimura <tyos@jp.ibm.com>,
	dev@dpdk.org, David Christensen <drc@linux.vnet.ibm.com>
Subject: Re: [dpdk-dev] [PATCH] vfio: retry creating sPAPR DMA window
Date: Wed, 10 Jul 2019 17:01:31 +0100	[thread overview]
Message-ID: <80114dab-b80a-9670-01af-735031f01c68@intel.com> (raw)
In-Reply-To: <2134455.0e53NjsZy4@xps>

On 10-Jul-19 1:39 PM, Thomas Monjalon wrote:
> 10/07/2019 14:35, Burakov, Anatoly:
>> On 10-Jul-19 1:17 PM, Thomas Monjalon wrote:
>>> 10/07/2019 12:32, Burakov, Anatoly:
>>>> On 07-Jun-19 3:28 AM, Takeshi Yoshimura wrote:
>>>>> sPAPR allows only page_shift from VFIO_IOMMU_SPAPR_TCE_GET_INFO ioctl.
>>>>> However, Linux 4.17 or before returns incorrect page_shift for Power9.
>>>>> I added the code for retrying creation of sPAPR DMA window.
>>>>>
>>>>> Signed-off-by: Takeshi Yoshimura <tyos@jp.ibm.com>
>>>>> ---
>>>>>     lib/librte_eal/linux/eal/eal_vfio.c | 26 +++++++++++++++++++++++---
>>>>>     1 file changed, 23 insertions(+), 3 deletions(-)
>>>>>
>>>>> diff --git a/lib/librte_eal/linux/eal/eal_vfio.c b/lib/librte_eal/linux/eal/eal_vfio.c
>>>>> index 6892a2c14..f16c5c3c0 100644
>>>>> --- a/lib/librte_eal/linux/eal/eal_vfio.c
>>>>> +++ b/lib/librte_eal/linux/eal/eal_vfio.c
>>>>> @@ -1448,9 +1448,29 @@ vfio_spapr_create_new_dma_window(int vfio_container_fd,
>>>>>     	/* create new DMA window */
>>>>>     	ret = ioctl(vfio_container_fd, VFIO_IOMMU_SPAPR_TCE_CREATE, create);
>>>>>     	if (ret) {
>>>>> -		RTE_LOG(ERR, EAL, "  cannot create new DMA window, "
>>>>> -				"error %i (%s)\n", errno, strerror(errno));
>>>>> -		return -1;
>>>>> +		/* try possible page_shift and levels for workaround */
>>>>> +		uint32_t levels;
>>>>> +
>>>>> +		for (levels = 1; levels <= info.ddw.levels; levels++) {
>>>>> +			uint32_t pgsizes = info.ddw.pgsizes;
>>>>
>>>> +CC POWER maintainer from MAINTAINERS file.
>>>>
>>>> This is failing compilation on some older distros (Fedora 20, Ubuntu 14.04):
>>>>
>>>>
>>>> /root/dpdk/lib/librte_eal/linux/eal/eal_vfio.c: In function
>>>> ‘vfio_spapr_create_new_dma_window’:
>>>> /root/dpdk/lib/librte_eal/linux/eal/eal_vfio.c:1451:34: error: ‘struct
>>>> vfio_iommu_spapr_tce_info’ has no member named ‘ddw’
>>>> for (levels = 1; levels <= info.ddw.levels; levels++) {
>>>> ^
>>>> /root/dpdk/lib/librte_eal/linux/eal/eal_vfio.c:1452:27: error: ‘struct
>>>> vfio_iommu_spapr_tce_info’ has no member named ‘ddw’
>>>> uint32_t pgsizes = info.ddw.pgsizes;
>>>> ^
>>>> /root/dpdk/lib/librte_eal/linux/eal/eal_vfio.c: At top level:
>>>> cc1: error: unrecognized command line option
>>>> "-Wno-address-of-packed-member" [-Werror]
>>>> cc1: all warnings being treated as errors
>>>> make[7]: *** [eal_vfio.o] Error 1
>>>> make[7]: *** Waiting for unfinished jobs....
>>>> make[6]: *** [eal] Error 2
>>>> make[5]: *** [linux] Error 2
>>>> make[4]: *** [librte_eal] Error 2
>>>> make[3]: *** [lib] Error 2
>>>> make[2]: *** [all] Error 2
>>>> make[1]: *** [pre_install] Error 2
>>>> make: *** [install] Error 2
>>>>
>>>>
>>>> Looking at the eal_vfio.h we are handling case of missing sPAPR info/ddw
>>>> structures, but this isn't simply a case of a missing struct; rather it
>>>> is an issue of differing definitions between what recent kernels have,
>>>> and what older kernels had.
>>>>
>>>> @Thomas Do we still support these older distros?
>>>
>>> If the kernel is not maintained, we don't support it.
>>> For most distros, Linux < 3.16 is not maintained.
>>
>> This is on kernel 3.13 (Ubuntu 14.04 LTS which according to Canonical's
>> published schedule [1] has stopped receiving even maintenance updates,
>> and is only updated for select customers). It was never an LTS kernel to
>> begin with, i think.
> 
> So we can ignore this issue.
> 

It seems that it's also reproducible on Ubuntu 14.04 with kernel 3.19. 
Which is not an upstream LTS kernel, but it's >3.16.

It also looks like it may be an issue for us specifically, so i would 
suggest reverting the patch for rc2 if the submitter/maintainers don't 
come back and fix it.

-- 
Thanks,
Anatoly

  reply	other threads:[~2019-07-10 16:01 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-06-07  2:28 Takeshi Yoshimura
2019-07-04 16:01 ` Thomas Monjalon
2019-07-08 16:47   ` David Christensen
2019-07-08 17:45     ` Thomas Monjalon
2019-07-09 10:22       ` Burakov, Anatoly
2019-07-05  8:15 ` Burakov, Anatoly
2019-07-07 21:21   ` Thomas Monjalon
2019-07-10 10:32 ` Burakov, Anatoly
2019-07-10 12:17   ` Thomas Monjalon
2019-07-10 12:35     ` Burakov, Anatoly
2019-07-10 12:39       ` Thomas Monjalon
2019-07-10 16:01         ` Burakov, Anatoly [this message]
2019-07-11  0:50           ` David Christensen
2019-07-11  2:18         ` Takeshi T Yoshimura

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=80114dab-b80a-9670-01af-735031f01c68@intel.com \
    --to=anatoly.burakov@intel.com \
    --cc=dev@dpdk.org \
    --cc=drc@linux.vnet.ibm.com \
    --cc=thomas@monjalon.net \
    --cc=tyos@jp.ibm.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).