linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bjorn Helgaas <helgaas@kernel.org>
To: "Saheed O. Bolarinwa" <refactormyself@gmail.com>
Cc: linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org, kw@linux.com
Subject: Re: [RFC PATCH 1/4] PCI/ASPM: Remove struct pcie_link_state.parent
Date: Sun, 26 Sep 2021 17:05:37 -0500	[thread overview]
Message-ID: <20210926220537.GA591345@bhelgaas> (raw)
In-Reply-To: <20210916085206.2268-2-refactormyself@gmail.com>

On Thu, Sep 16, 2021 at 10:52:03AM +0200, Saheed O. Bolarinwa wrote:
> From: "Bolarinwa O. Saheed" <refactormyself@gmail.com>
> 
> Information cached in struct pcie_link_state.parent is accessible
> via struct pci_dev.
> 
> This patch:
>  - removes *parent* from the *struct pcie_link_state*
>  - adjusts all references to it to access the information directly
> 
> Signed-off-by: Bolarinwa O. Saheed <refactormyself@gmail.com>
> ---
> OPINION: the checkpatch.pl scring warns on this line:
> 	`BUG_ON(root->pdev->bus->parent->self);`
> however, I think if a root device reports a parent, that is serious!

Do you mean this warning?

  WARNING: Missing a blank line after declarations
  #967: FILE: drivers/pci/pcie/aspm.c:967:
  +	struct pcie_link_state *link;
  +	BUG_ON(root->pdev->bus->parent->self);

That's just complaining about a blank line, so no big deal.  You could
resolve that by adding the blank line in this patch.

The fact that we use BUG_ON() at all *is* a real problem.  See the
comments at the BUG() definition.  We should rework this so that
condition is either impossible and we can just remove the BUG_ON(), or
we can deal with it gracefully.  But this would be material for a
different patch.

>  drivers/pci/pcie/aspm.c | 18 +++++++++++-------
>  1 file changed, 11 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/pci/pcie/aspm.c b/drivers/pci/pcie/aspm.c
> index 013a47f587ce..48b83048aa30 100644
> --- a/drivers/pci/pcie/aspm.c
> +++ b/drivers/pci/pcie/aspm.c
> @@ -50,7 +50,6 @@ struct pcie_link_state {
>  	struct pci_dev *pdev;		/* Upstream component of the Link */
>  	struct pci_dev *downstream;	/* Downstream component, function 0 */
>  	struct pcie_link_state *root;	/* pointer to the root port link */
> -	struct pcie_link_state *parent;	/* pointer to the parent Link state */
>  	struct list_head sibling;	/* node in link_list */
>  
>  	/* ASPM state */
> @@ -379,6 +378,7 @@ static void encode_l12_threshold(u32 threshold_us, u32 *scale, u32 *value)
>  static void pcie_aspm_check_latency(struct pci_dev *endpoint)
>  {
>  	u32 latency, l1_switch_latency = 0;
> +	struct pci_dev *parent;
>  	struct aspm_latency *acceptable;
>  	struct pcie_link_state *link;
>  
> @@ -419,7 +419,8 @@ static void pcie_aspm_check_latency(struct pci_dev *endpoint)
>  			link->aspm_capable &= ~ASPM_STATE_L1;
>  		l1_switch_latency += 1000;
>  
> -		link = link->parent;
> +		parent = link->pdev->bus->parent->self;
> +		link = !parent ? NULL : parent->link_state;

I love the direction of this patch, but this chain of pointers
(link->pdev->bus->parent->self) is a little over the top and is
repeated several times here.

Can we simplify it a bit by making a helper function?  It's similar
but not quite the same as pci_upstream_bridge().

And maybe reverse the condition to avoid the negation?

  link = parent ? parent->link_state : NULL;

>  	}
>  }
>  
> @@ -793,9 +794,11 @@ static void pcie_config_aspm_link(struct pcie_link_state *link, u32 state)
>  
>  static void pcie_config_aspm_path(struct pcie_link_state *link)
>  {
> +	struct pci_dev *parent;
>  	while (link) {
>  		pcie_config_aspm_link(link, policy_to_aspm_state(link));
> -		link = link->parent;
> +		parent = link->pdev->bus->parent->self;
> +		link = !parent ? NULL : parent->link_state;
>  	}
>  }
>  
> @@ -872,8 +875,7 @@ static struct pcie_link_state *alloc_pcie_link_state(struct pci_dev *pdev)
>  			return NULL;
>  		}
>  
> -		link->parent = parent;
> -		link->root = link->parent->root;
> +		link->root = parent->root;
>  	}
>  
>  	list_add(&link->sibling, &link_list);
> @@ -962,7 +964,7 @@ void pcie_aspm_init_link_state(struct pci_dev *pdev)
>  static void pcie_update_aspm_capable(struct pcie_link_state *root)
>  {
>  	struct pcie_link_state *link;
> -	BUG_ON(root->parent);
> +	BUG_ON(root->pdev->bus->parent->self);
>  	list_for_each_entry(link, &link_list, sibling) {
>  		if (link->root != root)
>  			continue;
> @@ -985,6 +987,7 @@ static void pcie_update_aspm_capable(struct pcie_link_state *root)
>  /* @pdev: the endpoint device */
>  void pcie_aspm_exit_link_state(struct pci_dev *pdev)
>  {
> +	struct pci_dev *parent_dev;
>  	struct pci_dev *parent = pdev->bus->self;
>  	struct pcie_link_state *link, *root, *parent_link;
>  
> @@ -1002,7 +1005,8 @@ void pcie_aspm_exit_link_state(struct pci_dev *pdev)
>  
>  	link = parent->link_state;
>  	root = link->root;
> -	parent_link = link->parent;
> +	parent_dev = link->pdev->bus->parent->self;
> +	parent_link = !parent_dev ? NULL : parent_dev->link_state;
>  
>  	/* All functions are removed, so just disable ASPM for the link */
>  	pcie_config_aspm_link(link, 0);
> -- 
> 2.20.1
> 

  reply	other threads:[~2021-09-26 22:05 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-16  8:52 [RFC PATCH 0/3] PCI/ASPM: Remove unncessary linked list in aspm.c Saheed O. Bolarinwa
2021-09-16  8:52 ` [RFC PATCH 1/4] PCI/ASPM: Remove struct pcie_link_state.parent Saheed O. Bolarinwa
2021-09-26 22:05   ` Bjorn Helgaas [this message]
2021-09-16  8:52 ` [RFC PATCH 2/4] PCI/ASPM: Remove struct pcie_link_state.root Saheed O. Bolarinwa
2021-09-16  8:52 ` [RFC PATCH 3/4] PCI/ASPM: Remove struct pcie_link_state.downstream Saheed O. Bolarinwa
2021-09-16  8:52 ` [RFC PATCH 4/4] PCI/ASPM: Remove unncessary linked list defined within aspm.c Saheed O. Bolarinwa

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=20210926220537.GA591345@bhelgaas \
    --to=helgaas@kernel.org \
    --cc=kw@linux.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=refactormyself@gmail.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).