linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ram Pai <linuxram@us.ibm.com>
To: Yinghai Lu <yinghai@kernel.org>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>,
	Ram Pai <linuxram@us.ibm.com>,
	Dominik Brodowski <linux@dominikbrodowski.net>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 4/9] PCI: Fix cardbus bridge resources as optional size handling
Date: Wed, 8 Feb 2012 12:35:17 +0800	[thread overview]
Message-ID: <20120208043517.GD8930@ram-ThinkPad-T61> (raw)
In-Reply-To: <1328424908-6385-5-git-send-email-yinghai@kernel.org>

On Sat, Feb 04, 2012 at 10:55:03PM -0800, Yinghai Lu wrote:
> We should not set the requested size to -2.
> 
> that will confuse the resource list sorting with align when SIZEALIGN is used.
> 
> Change to STARTALIGN and pass align from start.
> 
> We are safe to do that just as we do that regular pci bridge.
> 
> In the long run, we should just treat cardbus like regular pci bridge.
> 
> Also fix when realloc is not passed, should keep the requested size.
> 
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>
> ---
>  drivers/pci/setup-bus.c |   63 ++++++++++++++++++++++++++--------------------
>  1 files changed, 36 insertions(+), 27 deletions(-)
> 
> diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
> index d5897c3..3b3932a 100644
> --- a/drivers/pci/setup-bus.c
> +++ b/drivers/pci/setup-bus.c
> @@ -898,21 +898,30 @@ static void pci_bus_size_cardbus(struct pci_bus *bus,
>  {
>  	struct pci_dev *bridge = bus->self;
>  	struct resource *b_res = &bridge->resource[PCI_BRIDGE_RESOURCES];
> +	resource_size_t b_res_3_size = pci_cardbus_mem_size * 2;
>  	u16 ctrl;
> 
>  	/*
>  	 * Reserve some resources for CardBus.  We reserve
>  	 * a fixed amount of bus space for CardBus bridges.
>  	 */
> -	b_res[0].start = 0;
> -	b_res[0].flags |= IORESOURCE_IO | IORESOURCE_SIZEALIGN;
> -	if (realloc_head)
> -		add_to_list(realloc_head, bridge, b_res, pci_cardbus_io_size, 0 /* dont care */);
> +	b_res[0].start = pci_cardbus_io_size;
> +	b_res[0].end = b_res[0].start + pci_cardbus_io_size - 1;
> +	b_res[0].flags |= IORESOURCE_IO | IORESOURCE_STARTALIGN;
> +	if (realloc_head) {
> +		b_res[0].end -= pci_cardbus_io_size;
> +		add_to_list(realloc_head, bridge, b_res, pci_cardbus_io_size,
> +				pci_cardbus_io_size);
> +	}
> 
> -	b_res[1].start = 0;
> -	b_res[1].flags |= IORESOURCE_IO | IORESOURCE_SIZEALIGN;
> -	if (realloc_head)
> -		add_to_list(realloc_head, bridge, b_res+1, pci_cardbus_io_size, 0 /* dont care */);
> +	b_res[1].start = pci_cardbus_io_size;
> +	b_res[1].end = b_res[1].start + pci_cardbus_io_size - 1;
> +	b_res[1].flags |= IORESOURCE_IO | IORESOURCE_STARTALIGN;
> +	if (realloc_head) {
> +		b_res[1].end -= pci_cardbus_io_size;
> +		add_to_list(realloc_head, bridge, b_res+1, pci_cardbus_io_size,
> +				 pci_cardbus_io_size);
> +	}
> 
>  	/* MEM1 must not be pref mmio */
>  	pci_read_config_word(bridge, PCI_CB_BRIDGE_CONTROL, &ctrl);
> @@ -939,28 +948,28 @@ static void pci_bus_size_cardbus(struct pci_bus *bus,
>  	 * twice the size.
>  	 */
>  	if (ctrl & PCI_CB_BRIDGE_CTL_PREFETCH_MEM0) {
> -		b_res[2].start = 0;
> -		b_res[2].flags |= IORESOURCE_MEM | IORESOURCE_PREFETCH | IORESOURCE_SIZEALIGN;
> -		if (realloc_head)
> -			add_to_list(realloc_head, bridge, b_res+2, pci_cardbus_mem_size, 0 /* dont care */);
> +		b_res[2].start = pci_cardbus_mem_size;
> +		b_res[2].end = b_res[2].start + pci_cardbus_mem_size - 1;
> +		b_res[2].flags |= IORESOURCE_MEM | IORESOURCE_PREFETCH |
> +				  IORESOURCE_STARTALIGN;
> +		if (realloc_head) {
> +			b_res[2].end -= pci_cardbus_mem_size;
> +			add_to_list(realloc_head, bridge, b_res+2,
> +				 pci_cardbus_mem_size, pci_cardbus_mem_size);
> +		}
> 
> -		b_res[3].start = 0;
> -		b_res[3].flags |= IORESOURCE_MEM | IORESOURCE_SIZEALIGN;
> -		if (realloc_head)
> -			add_to_list(realloc_head, bridge, b_res+3, pci_cardbus_mem_size, 0 /* dont care */);

The b_res_3_size has to be reduced to half here.
Otherwise it will try allocate 2*pci_cardbus_mem_size to the BAR3..

		b_res_3_size = pci_cardbus_mem_size;


> -	} else {
> -		b_res[3].start = 0;
> -		b_res[3].flags |= IORESOURCE_MEM | IORESOURCE_SIZEALIGN;
> -		if (realloc_head)
> -			add_to_list(realloc_head, bridge, b_res+3, pci_cardbus_mem_size * 2, 0 /* dont care */);
> +		/* reduce that to half */
> +		b_res_3_size = pci_cardbus_mem_size;

b_res3_3_size should not be updated here, since BAR3 has to be allocated 2*pci_cardbus_mem_size
resource.

RP


>  	}
> 
> -	/* set the size of the resource to zero, so that the resource does not
> -	 * get assigned during required-resource allocation cycle but gets assigned
> -	 * during the optional-resource allocation cycle.
> - 	 */
> -	b_res[0].start = b_res[1].start = b_res[2].start = b_res[3].start = 1;
> -	b_res[0].end = b_res[1].end = b_res[2].end = b_res[3].end = 0;
> +	b_res[3].start = pci_cardbus_mem_size;
> +	b_res[3].end = b_res[3].start + b_res_3_size - 1;
> +	b_res[3].flags |= IORESOURCE_MEM | IORESOURCE_STARTALIGN;
> +	if (realloc_head) {
> +		b_res[3].end -= b_res_3_size;
> +		add_to_list(realloc_head, bridge, b_res+3, b_res_3_size,
> +				 pci_cardbus_mem_size);
> +	}
>  }
> 
>  void __ref __pci_bus_size_bridges(struct pci_bus *bus,
> -- 
> 1.7.7


  reply	other threads:[~2012-02-08  4:35 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-02-05  6:54 [PATCH -v2 0/9] PCI : bridge resource reallocation patchset -- followup Yinghai Lu
2012-02-05  6:55 ` [PATCH 1/9] pci: Fix pci cardbus removal Yinghai Lu
2012-02-10 20:20   ` Jesse Barnes
2012-02-05  6:55 ` [PATCH 2/9] PCI: Fix /sys warning when sriov enabled card is hot removed Yinghai Lu
2012-02-10 20:39   ` Jesse Barnes
2012-02-05  6:55 ` [PATCH 3/9] PCI: Disable cardbus bridge MEM1 pref CTL Yinghai Lu
2012-02-10 20:46   ` Jesse Barnes
2012-02-10 20:54     ` Yinghai Lu
2012-02-10 22:11       ` Bjorn Helgaas
2012-02-05  6:55 ` [PATCH 4/9] PCI: Fix cardbus bridge resources as optional size handling Yinghai Lu
2012-02-08  4:35   ` Ram Pai [this message]
2012-02-08  4:48     ` Yinghai Lu
2012-02-08  5:01       ` Ram Pai
2012-02-08  6:11         ` Yinghai Lu
2012-02-10 20:52           ` Jesse Barnes
2012-02-10 20:56             ` Yinghai Lu
2012-02-05  6:55 ` [PATCH 5/9] PCI: Skip reset cardbus assigned resource during pci bus rescan Yinghai Lu
2012-02-05  6:55 ` [PATCH 6/9] PCI: Retry on type IORESOURCE_IO allocation Yinghai Lu
2012-02-05  6:55 ` [PATCH 7/9] PCI: Make pci bridge reallocating enabled/disabled Yinghai Lu
2012-02-10 20:56   ` Jesse Barnes
2012-02-05  6:55 ` [PATCH 8/9] PCI: print out suggestion about using pci=realloc Yinghai Lu
2012-02-10 20:59   ` Jesse Barnes
2012-02-05  6:55 ` [PATCH 9/9] PCI: only enable pci realloc when SRIOV bar is not assigned Yinghai Lu
2012-02-10 21:00   ` Jesse Barnes
2012-02-10 21:23     ` Yinghai Lu

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=20120208043517.GD8930@ram-ThinkPad-T61 \
    --to=linuxram@us.ibm.com \
    --cc=jbarnes@virtuousgeek.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=linux@dominikbrodowski.net \
    --cc=torvalds@linux-foundation.org \
    --cc=yinghai@kernel.org \
    /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).