Linux PCI subsystem development
 help / color / mirror / Atom feed
From: Dan Williams <dan.j.williams@intel.com>
To: "Daisuke Kobayashi (Fujitsu)" <kobayashi.da-06@fujitsu.com>,
	"'Dan Williams'" <dan.j.williams@intel.com>,
	"linux-cxl@vger.kernel.org" <linux-cxl@vger.kernel.org>
Cc: "Yasunori Gotou (Fujitsu)" <y-goto@fujitsu.com>,
	"linux-pci@vger.kernel.org" <linux-pci@vger.kernel.org>,
	"mj@ucw.cz" <mj@ucw.cz>
Subject: RE: [PATCH v3 1/3] Add sysfs attribute for CXL 1.1 device link status
Date: Mon, 8 Apr 2024 14:43:27 -0700	[thread overview]
Message-ID: <661464ff8492a_2583ad29419@dwillia2-xfh.jf.intel.com.notmuch> (raw)
In-Reply-To: <OSAPR01MB7182F74119B93BB6D944C1B1BA032@OSAPR01MB7182.jpnprd01.prod.outlook.com>

Daisuke Kobayashi (Fujitsu) wrote:
[..]
> I would like to report on some additional findings.
> The process of registering cxl_rcd_groups to struct pci_driver.driver.dev_groups seems
> to not generate a file in sysfs when looking at the contents of the module_pci_driver() macro.

Oh, apologies, and thanks for taking a deeper look. The failure is
because my suggested example led you astray. I suggested this:

diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
index 2ff361e756d6..eec04f103aa8 100644
--- a/drivers/cxl/pci.c
+++ b/drivers/cxl/pci.c
@@ -971,6 +971,7 @@ static struct pci_driver cxl_pci_driver = {
        .err_handler            = &cxl_error_handlers,
        .driver = {
                .probe_type     = PROBE_PREFER_ASYNCHRONOUS,
+               .dev_groups     = cxl_rcd_groups,
        },
 }; 


...the correct place to put it is here:

@@ -969,6 +969,7 @@ static struct pci_driver cxl_pci_driver = {
        .id_table               = cxl_mem_pci_tbl,
        .probe                  = cxl_pci_probe,
        .err_handler            = &cxl_error_handlers,
+       .dev_groups             = cxl_rcd_groups,
        .driver = {
                .probe_type     = PROBE_PREFER_ASYNCHRONOUS,
        },


...otherwise __pci_register_driver() will overwrite it. This is a subtle
bug given probe_type is directly initialized in .driver.

> For this feature, I think it would be best to output the values to a directory
> of /sys/bus/pci/devices/<pci-addr>/. To output to this directory, the attribute would
> need to be registered to pci_dev.dev.
> My current understanding is that the best way to do this would be to register the attribute
> with device_add_groups(&pdev->dev, cxl_rcd_groups) on probe and remove the files with

No, the dynamic sysfs registration APIs should be avoided when possible.
The above fix is what you need.

  reply	other threads:[~2024-04-08 21:43 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-12  8:05 [PATCH v3 0/3] Display cxl1.1 device link status Kobayashi,Daisuke
2024-03-12  8:05 ` [PATCH v3 1/3] Add sysfs attribute for CXL 1.1 " Kobayashi,Daisuke
2024-03-26 19:51   ` Dan Williams
2024-03-28  1:47     ` Daisuke Kobayashi (Fujitsu)
2024-04-03  9:40       ` Daisuke Kobayashi (Fujitsu)
2024-04-05  8:31         ` Daisuke Kobayashi (Fujitsu)
2024-04-08 21:43           ` Dan Williams [this message]
2024-04-09  4:55             ` Daisuke Kobayashi (Fujitsu)
2024-04-05 17:25         ` Jonathan Cameron
2024-04-08 21:32         ` Dan Williams
2024-04-09 14:59   ` Bjorn Helgaas
2024-04-09 15:00     ` Bjorn Helgaas
2024-03-12  8:05 ` [PATCH v3 2/3] Remove conditional branch that is not suitable for cxl1.1 devices Kobayashi,Daisuke
2024-03-26 20:00   ` Dan Williams
2024-03-27  8:26     ` Daisuke Kobayashi (Fujitsu)
2024-03-12  8:05 ` [PATCH v3 3/3] Add function to display cxl1.1 device link status Kobayashi,Daisuke
2024-03-26 20:05   ` Dan Williams
2024-03-27  8:27     ` Daisuke Kobayashi (Fujitsu)
2024-03-29 22:23   ` Martin Mareš
2024-03-30  1:15     ` Dan Williams
2024-03-31  1:03       ` Martin Mareš
2024-04-01 17:47         ` Dan Williams
2024-04-02  7:09           ` Daisuke Kobayashi (Fujitsu)
2024-03-25  4:49 ` [PATCH v3 0/3] Display " Daisuke Kobayashi (Fujitsu)
2024-03-26 19:15 ` Dan Williams
2024-03-27  8:24   ` Daisuke Kobayashi (Fujitsu)

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=661464ff8492a_2583ad29419@dwillia2-xfh.jf.intel.com.notmuch \
    --to=dan.j.williams@intel.com \
    --cc=kobayashi.da-06@fujitsu.com \
    --cc=linux-cxl@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=mj@ucw.cz \
    --cc=y-goto@fujitsu.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