linux-scsi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Damien Le Moal <damien.lemoal@opensource.wdc.com>
To: Steffen Maier <maier@linux.ibm.com>,
	linux-scsi@vger.kernel.org,
	"Martin K . Petersen" <martin.petersen@oracle.com>,
	linux-ide@vger.kernel.org
Cc: Bart Van Assche <bvanassche@acm.org>
Subject: Re: [PATCH] scsi: fix scsi device attributes registration
Date: Fri, 12 Nov 2021 08:03:39 +0900	[thread overview]
Message-ID: <9eb04b1a-52a6-a8d3-b5b7-04df8cb4c2e2@opensource.wdc.com> (raw)
In-Reply-To: <0cdf8e28-2fcc-9aa6-e8cb-ef1c0c77bd69@linux.ibm.com>

On 2021/11/11 19:35, Steffen Maier wrote:
> On 11/11/21 09:45, Damien Le Moal wrote:
>> Since the sdev_gendev device of a scsi device defines its attributes
>> using scsi_sdev_attr_groups as the groups field value of its device
>> type, the execution of device_add() in scsi_sysfs_add_sdev() register
>> with sysfs only the attributes defined using scsi_sdev_attr_groups. As
>> a results, the attributes defined by an LLD using the scsi host
>> sdev_groups attribute groups are never registered with sysfs and not
>> visible to the users.
>>
>> Fix this problem by removing scsi_sdev_attr_groups and manually setting
>> the groups field of a scsi device sdev_gendev to point to the scsi
>> device gendev_attr_groups. As the first entry of this array of
>> attribute groups is scsi_sdev_attr_group, using gendev_attr_groups as
>> the gendev groups result in all defined attributes to be created in
>> sysfs when device_add() is called.
>>
>> Fixes: 92c4b58b15c5 ("scsi: core: Register sysfs attributes earlier")
>> cc: Bart Van Assche <bvanassche@acm.org>
>> Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
>> ---
>>   drivers/scsi/scsi_sysfs.c | 7 +------
>>   1 file changed, 1 insertion(+), 6 deletions(-)
>>
>> diff --git a/drivers/scsi/scsi_sysfs.c b/drivers/scsi/scsi_sysfs.c
>> index d3d362289ecc..92c92853f516 100644
>> --- a/drivers/scsi/scsi_sysfs.c
>> +++ b/drivers/scsi/scsi_sysfs.c
>> @@ -1301,11 +1301,6 @@ static struct attribute_group scsi_sdev_attr_group = {
>>   	.is_bin_visible = scsi_sdev_bin_attr_is_visible,
>>   };
>>   
>> -static const struct attribute_group *scsi_sdev_attr_groups[] = {
>> -	&scsi_sdev_attr_group,
>> -	NULL
>> -};
>> -
>>   static int scsi_target_add(struct scsi_target *starget)
>>   {
>>   	int error;
>> @@ -1575,7 +1570,6 @@ int scsi_sysfs_add_host(struct Scsi_Host *shost)
>>   static struct device_type scsi_dev_type = {
>>   	.name =		"scsi_device",
>>   	.release =	scsi_device_dev_release,
>> -	.groups =	scsi_sdev_attr_groups,
>>   };
>>   
>>   void scsi_sysfs_device_initialize(struct scsi_device *sdev)
>> @@ -1601,6 +1595,7 @@ void scsi_sysfs_device_initialize(struct scsi_device *sdev)
>>   		}
>>   	}
>>   	WARN_ON_ONCE(j >= ARRAY_SIZE(sdev->gendev_attr_groups));
>> +	sdev->sdev_gendev.groups = sdev->gendev_attr_groups;
>>   
>>   	device_initialize(&sdev->sdev_dev);
>>   	sdev->sdev_dev.parent = get_device(&sdev->sdev_gendev);
>>
> 
> Damien, which kernel were you using?
> 
> Isn't this fixed by?:
> 
> next:
> 
> https://lore.kernel.org/linux-scsi/163478764102.7011.9375895285870786953.b4-ty@oracle.com/t/#mab0eeb4a8d8db95c3ace0013bfef775736e124cb
> ("scsi: core: Fix early registration of sysfs attributes for scsi_device")
> https://git.kernel.org/pub/scm/linux/kernel/git/mkp/scsi.git/commit/?h=5.16/scsi-staging&id=3a71f0f7a51259b3cb95d79cac1e19dcc5e89ce9
> https://git.kernel.org/pub/scm/linux/kernel/git/mkp/scsi.git/commit/?h=5.16/scsi-queue&id=3a71f0f7a51259b3cb95d79cac1e19dcc5e89ce9
> https://git.kernel.org/pub/scm/linux/kernel/git/mkp/scsi.git/commit/?h=for-next&id=3a71f0f7a51259b3cb95d79cac1e19dcc5e89ce9
> =>
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next-history.git/commit/?h=next-20211028&id=503f375baa99edff894eb1a534d2ac0b4f799573
> 
> soon in vanilla:
> 
> https://lore.kernel.org/linux-scsi/163612851260.17201.4106345384610850520.pr-tracker-bot@kernel.org/t/#md79869a3966ccceb30eef658b85743e49cf31dc1

I discovered the problem while testing libata tree which is currently based on
Linus master. So I completely missed this fix, which is even simpler than what I
sent :) I should have checked Martin's tree first !
Thanks for pointing this out. I tested and everything is fine.

Martin, Bart,

My apologies for the noise :)

-- 
Damien Le Moal
Western Digital Research

      reply	other threads:[~2021-11-11 23:03 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-11  8:45 Damien Le Moal
2021-11-11  9:56 ` Christoph Hellwig
2021-11-11 10:35 ` Steffen Maier
2021-11-11 23:03   ` Damien Le Moal [this message]

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=9eb04b1a-52a6-a8d3-b5b7-04df8cb4c2e2@opensource.wdc.com \
    --to=damien.lemoal@opensource.wdc.com \
    --cc=bvanassche@acm.org \
    --cc=linux-ide@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=maier@linux.ibm.com \
    --cc=martin.petersen@oracle.com \
    --subject='Re: [PATCH] scsi: fix scsi device attributes registration' \
    /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

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).