linux-ide.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Niklas Cassel <cassel@kernel.org>
To: Igor Pylypiv <ipylypiv@google.com>
Cc: Damien Le Moal <dlemoal@kernel.org>,
	John Garry <john.g.garry@oracle.com>,
	Jason Yan <yanaijie@huawei.com>,
	"James E.J. Bottomley" <jejb@linux.ibm.com>,
	"Martin K. Petersen" <martin.petersen@oracle.com>,
	Jack Wang <jinpu.wang@cloud.ionos.com>,
	Hannes Reinecke <hare@suse.de>,
	Xiang Chen <chenxiang66@hisilicon.com>,
	Artur Paszkiewicz <artur.paszkiewicz@intel.com>,
	Bart Van Assche <bvanassche@acm.org>,
	TJ Adams <tadamsjr@google.com>,
	linux-ide@vger.kernel.org, linux-scsi@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v8 1/7] ata: libata-sata: Factor out NCQ Priority configuration helpers
Date: Fri, 8 Mar 2024 11:03:21 +0100	[thread overview]
Message-ID: <ZeriaUWlhBqp4Q77@ryzen> (raw)
In-Reply-To: <20240307214418.3812290-2-ipylypiv@google.com>

On Thu, Mar 07, 2024 at 01:44:12PM -0800, Igor Pylypiv wrote:
> Export libata NCQ Priority configuration helpers to be reused
> for libsas managed SATA devices.
> 
> Switched locking from spin_lock_irq() to spin_lock_irqsave().
> In the future someone might call these helper functions when interrupts
> are disabled. spin_unlock_irq() could lead to a premature re-enabling
> of interrupts, whereas spin_unlock_irqrestore() restores the interrupt
> state to its condition prior to the spin_lock_irqsave() call.

Seems like a mistake in the existing code, why would
ata_ncq_prio_supported_show() and ata_ncq_prio_enable_show()
use spin_lock_irq() ?

Seems like spin_lock() would be better.

For ata_ncq_prio_enable_store(), you probably want to
spin_lock_irq() or spin_lock_irqsave().


Anyway, like you said, as you are now creating helper functions:
ata_ncq_prio_supported(), ata_ncq_prio_enabled(), ata_ncq_prio_enable()
these function might no longer only be called from sysfs code,
so it is probably a bad idea to let these functions use spin_lock_irq().

However, can't ata_ncq_prio_supported() and ata_ncq_prio_enabled()
still use a simple spin_lock(), why would they need to disable irqs?

Damien, you are the author of ata_ncq_prio_supported_show(), thoughts?


Kind regards,
Niklas

> 
> Acked-by: Damien Le Moal <dlemoal@kernel.org>
> Reviewed-by: Jason Yan <yanaijie@huawei.com>
> Reviewed-by: Hannes Reinecke <hare@suse.de>
> Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
> ---
>  drivers/ata/libata-sata.c | 160 +++++++++++++++++++++++++++++---------
>  include/linux/libata.h    |   6 ++
>  2 files changed, 128 insertions(+), 38 deletions(-)
> 
> diff --git a/drivers/ata/libata-sata.c b/drivers/ata/libata-sata.c
> index 0fb1934875f2..a8d773003d74 100644
> --- a/drivers/ata/libata-sata.c
> +++ b/drivers/ata/libata-sata.c
> @@ -848,80 +848,143 @@ DEVICE_ATTR(link_power_management_policy, S_IRUGO | S_IWUSR,
>  	    ata_scsi_lpm_show, ata_scsi_lpm_store);
>  EXPORT_SYMBOL_GPL(dev_attr_link_power_management_policy);
>  
> -static ssize_t ata_ncq_prio_supported_show(struct device *device,
> -					   struct device_attribute *attr,
> -					   char *buf)
> +/**
> + *	ata_ncq_prio_supported - Check if device supports NCQ Priority
> + *	@ap: ATA port of the target device
> + *	@sdev: SCSI device
> + *	@supported: Address of a boolean to store the result
> + *
> + *	Helper to check if device supports NCQ Priority feature.
> + *
> + *	Context: Any context. Takes and releases @ap->lock.
> + *
> + *	Return:
> + *	* %0		- OK. Status is stored into @supported
> + *	* %-ENODEV	- Failed to find the ATA device
> + */
> +int ata_ncq_prio_supported(struct ata_port *ap, struct scsi_device *sdev,
> +			   bool *supported)
>  {
> -	struct scsi_device *sdev = to_scsi_device(device);
> -	struct ata_port *ap = ata_shost_to_port(sdev->host);
>  	struct ata_device *dev;
> -	bool ncq_prio_supported;
> +	unsigned long flags;
>  	int rc = 0;
>  
> -	spin_lock_irq(ap->lock);
> +	spin_lock_irqsave(ap->lock, flags);
>  	dev = ata_scsi_find_dev(ap, sdev);
>  	if (!dev)
>  		rc = -ENODEV;
>  	else
> -		ncq_prio_supported = dev->flags & ATA_DFLAG_NCQ_PRIO;
> -	spin_unlock_irq(ap->lock);
> +		*supported = dev->flags & ATA_DFLAG_NCQ_PRIO;
> +	spin_unlock_irqrestore(ap->lock, flags);
> +
> +	return rc;
> +}
> +EXPORT_SYMBOL_GPL(ata_ncq_prio_supported);
> +
> +static ssize_t ata_ncq_prio_supported_show(struct device *device,
> +					   struct device_attribute *attr,
> +					   char *buf)
> +{
> +	struct scsi_device *sdev = to_scsi_device(device);
> +	struct ata_port *ap = ata_shost_to_port(sdev->host);
> +	bool supported;
> +	int rc;
> +
> +	rc = ata_ncq_prio_supported(ap, sdev, &supported);
> +	if (rc)
> +		return rc;
>  
> -	return rc ? rc : sysfs_emit(buf, "%u\n", ncq_prio_supported);
> +	return sysfs_emit(buf, "%d\n", supported);
>  }
>  
>  DEVICE_ATTR(ncq_prio_supported, S_IRUGO, ata_ncq_prio_supported_show, NULL);
>  EXPORT_SYMBOL_GPL(dev_attr_ncq_prio_supported);
>  
> -static ssize_t ata_ncq_prio_enable_show(struct device *device,
> -					struct device_attribute *attr,
> -					char *buf)
> +/**
> + *	ata_ncq_prio_enabled - Check if NCQ Priority is enabled
> + *	@ap: ATA port of the target device
> + *	@sdev: SCSI device
> + *	@enabled: Address of a boolean to store the result
> + *
> + *	Helper to check if NCQ Priority feature is enabled.
> + *
> + *	Context: Any context. Takes and releases @ap->lock.
> + *
> + *	Return:
> + *	* %0		- OK. Status is stored into @enabled
> + *	* %-ENODEV	- Failed to find the ATA device
> + */
> +int ata_ncq_prio_enabled(struct ata_port *ap, struct scsi_device *sdev,
> +			 bool *enabled)
>  {
> -	struct scsi_device *sdev = to_scsi_device(device);
> -	struct ata_port *ap = ata_shost_to_port(sdev->host);
>  	struct ata_device *dev;
> -	bool ncq_prio_enable;
> +	unsigned long flags;
>  	int rc = 0;
>  
> -	spin_lock_irq(ap->lock);
> +	spin_lock_irqsave(ap->lock, flags);
>  	dev = ata_scsi_find_dev(ap, sdev);
>  	if (!dev)
>  		rc = -ENODEV;
>  	else
> -		ncq_prio_enable = dev->flags & ATA_DFLAG_NCQ_PRIO_ENABLED;
> -	spin_unlock_irq(ap->lock);
> +		*enabled = dev->flags & ATA_DFLAG_NCQ_PRIO_ENABLED;
> +	spin_unlock_irqrestore(ap->lock, flags);
>  
> -	return rc ? rc : sysfs_emit(buf, "%u\n", ncq_prio_enable);
> +	return rc;
>  }
> +EXPORT_SYMBOL_GPL(ata_ncq_prio_enabled);
>  
> -static ssize_t ata_ncq_prio_enable_store(struct device *device,
> -					 struct device_attribute *attr,
> -					 const char *buf, size_t len)
> +static ssize_t ata_ncq_prio_enable_show(struct device *device,
> +					struct device_attribute *attr,
> +					char *buf)
>  {
>  	struct scsi_device *sdev = to_scsi_device(device);
> -	struct ata_port *ap;
> -	struct ata_device *dev;
> -	long int input;
> -	int rc = 0;
> +	struct ata_port *ap = ata_shost_to_port(sdev->host);
> +	bool enabled;
> +	int rc;
>  
> -	rc = kstrtol(buf, 10, &input);
> +	rc = ata_ncq_prio_enabled(ap, sdev, &enabled);
>  	if (rc)
>  		return rc;
> -	if ((input < 0) || (input > 1))
> -		return -EINVAL;
>  
> -	ap = ata_shost_to_port(sdev->host);
> -	dev = ata_scsi_find_dev(ap, sdev);
> -	if (unlikely(!dev))
> -		return  -ENODEV;
> +	return sysfs_emit(buf, "%d\n", enabled);
> +}
> +
> +/**
> + *	ata_ncq_prio_enable - Enable/disable NCQ Priority
> + *	@ap: ATA port of the target device
> + *	@sdev: SCSI device
> + *	@enable: true - enable NCQ Priority, false - disable NCQ Priority
> + *
> + *	Helper to enable/disable NCQ Priority feature.
> + *
> + *	Context: Any context. Takes and releases @ap->lock.
> + *
> + *	Return:
> + *	* %0		- OK. Status is stored into @enabled
> + *	* %-ENODEV	- Failed to find the ATA device
> + *	* %-EINVAL	- NCQ Priority is not supported or CDL is enabled
> + */
> +int ata_ncq_prio_enable(struct ata_port *ap, struct scsi_device *sdev,
> +			bool enable)
> +{
> +	struct ata_device *dev;
> +	unsigned long flags;
> +	int rc = 0;
> +
> +	spin_lock_irqsave(ap->lock, flags);
>  
> -	spin_lock_irq(ap->lock);
> +	dev = ata_scsi_find_dev(ap, sdev);
> +	if (!dev) {
> +		rc = -ENODEV;
> +		goto unlock;
> +	}
>  
>  	if (!(dev->flags & ATA_DFLAG_NCQ_PRIO)) {
>  		rc = -EINVAL;
>  		goto unlock;
>  	}
>  
> -	if (input) {
> +	if (enable) {
>  		if (dev->flags & ATA_DFLAG_CDL_ENABLED) {
>  			ata_dev_err(dev,
>  				"CDL must be disabled to enable NCQ priority\n");
> @@ -934,9 +997,30 @@ static ssize_t ata_ncq_prio_enable_store(struct device *device,
>  	}
>  
>  unlock:
> -	spin_unlock_irq(ap->lock);
> +	spin_unlock_irqrestore(ap->lock, flags);
> +
> +	return rc;
> +}
> +EXPORT_SYMBOL_GPL(ata_ncq_prio_enable);
> +
> +static ssize_t ata_ncq_prio_enable_store(struct device *device,
> +					 struct device_attribute *attr,
> +					 const char *buf, size_t len)
> +{
> +	struct scsi_device *sdev = to_scsi_device(device);
> +	struct ata_port *ap = ata_shost_to_port(sdev->host);
> +	bool enable;
> +	int rc;
> +
> +	rc = kstrtobool(buf, &enable);
> +	if (rc)
> +		return rc;
> +
> +	rc = ata_ncq_prio_enable(ap, sdev, enable);
> +	if (rc)
> +		return rc;
>  
> -	return rc ? rc : len;
> +	return len;
>  }
>  
>  DEVICE_ATTR(ncq_prio_enable, S_IRUGO | S_IWUSR,
> diff --git a/include/linux/libata.h b/include/linux/libata.h
> index 26d68115afb8..6dd9a4f9ca7c 100644
> --- a/include/linux/libata.h
> +++ b/include/linux/libata.h
> @@ -1157,6 +1157,12 @@ extern int ata_scsi_change_queue_depth(struct scsi_device *sdev,
>  				       int queue_depth);
>  extern int ata_change_queue_depth(struct ata_port *ap, struct scsi_device *sdev,
>  				  int queue_depth);
> +extern int ata_ncq_prio_supported(struct ata_port *ap, struct scsi_device *sdev,
> +				  bool *supported);
> +extern int ata_ncq_prio_enabled(struct ata_port *ap, struct scsi_device *sdev,
> +				bool *enabled);
> +extern int ata_ncq_prio_enable(struct ata_port *ap, struct scsi_device *sdev,
> +			       bool enable);
>  extern struct ata_device *ata_dev_pair(struct ata_device *adev);
>  extern int ata_do_set_mode(struct ata_link *link, struct ata_device **r_failed_dev);
>  extern void ata_scsi_port_error_handler(struct Scsi_Host *host, struct ata_port *ap);
> -- 
> 2.44.0.278.ge034bb2e1d-goog
> 

  reply	other threads:[~2024-03-08 10:03 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-07 21:44 [PATCH v8 0/7] NCQ Priority sysfs sttributes for libsas Igor Pylypiv
2024-03-07 21:44 ` [PATCH v8 1/7] ata: libata-sata: Factor out NCQ Priority configuration helpers Igor Pylypiv
2024-03-08 10:03   ` Niklas Cassel [this message]
2024-03-12  6:17     ` Damien Le Moal
2024-03-12 15:37       ` Niklas Cassel
2024-03-12 15:37   ` Niklas Cassel
2024-03-07 21:44 ` [PATCH v8 2/7] scsi: libsas: Define NCQ Priority sysfs attributes for SATA devices Igor Pylypiv
2024-03-08 10:10   ` Niklas Cassel
2024-03-26  9:53   ` Geert Uytterhoeven
2024-03-26 10:05     ` Damien Le Moal
2024-03-26 10:06     ` Damien Le Moal
2024-03-26 10:16       ` Geert Uytterhoeven
2024-03-26 11:52         ` Damien Le Moal
2024-03-07 21:44 ` [PATCH v8 3/7] scsi: pm80xx: Add libsas SATA sysfs attributes group Igor Pylypiv
2024-03-07 21:44 ` [PATCH v8 4/7] scsi: mvsas: " Igor Pylypiv
2024-03-08 10:09   ` Niklas Cassel
2024-03-07 21:44 ` [PATCH v8 5/7] scsi: hisi_sas: " Igor Pylypiv
2024-03-08 10:09   ` Niklas Cassel
2024-03-07 21:44 ` [PATCH v8 6/7] scsi: aic94xx: " Igor Pylypiv
2024-03-08 10:10   ` Niklas Cassel
2024-03-07 21:44 ` [PATCH v8 7/7] scsi: isci: " Igor Pylypiv
2024-03-25 20:02 ` [PATCH v8 0/7] NCQ Priority sysfs sttributes for libsas Martin K. Petersen

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=ZeriaUWlhBqp4Q77@ryzen \
    --to=cassel@kernel.org \
    --cc=artur.paszkiewicz@intel.com \
    --cc=bvanassche@acm.org \
    --cc=chenxiang66@hisilicon.com \
    --cc=dlemoal@kernel.org \
    --cc=hare@suse.de \
    --cc=ipylypiv@google.com \
    --cc=jejb@linux.ibm.com \
    --cc=jinpu.wang@cloud.ionos.com \
    --cc=john.g.garry@oracle.com \
    --cc=linux-ide@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=martin.petersen@oracle.com \
    --cc=tadamsjr@google.com \
    --cc=yanaijie@huawei.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).