All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mika Westerberg <mika.westerberg@linux.intel.com>
To: Bjorn Helgaas <helgaas@kernel.org>
Cc: Bjorn Helgaas <bhelgaas@google.com>,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Len Brown <lenb@kernel.org>,
	Mario.Limonciello@dell.com,
	Michael Jamet <michael.jamet@intel.com>,
	Yehezkel Bernat <YehezkelShB@gmail.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	Lukas Wunner <lukas@wunner.de>,
	linux-pci@vger.kernel.org, linux-acpi@vger.kernel.org
Subject: Re: [PATCH v5 2/9] PCI: Take bridge window alignment into account when distributing resources
Date: Thu, 3 May 2018 15:39:41 +0300	[thread overview]
Message-ID: <20180503123941.GP2355@lahna.fi.intel.com> (raw)
In-Reply-To: <20180501201546.GC11698@bhelgaas-glaptop.roam.corp.google.com>

On Tue, May 01, 2018 at 03:32:46PM -0500, Bjorn Helgaas wrote:
> This is where it gets hard for me -- I'm not really comfortable if we
> have to convince ourselves that code is correct by testing every
> scenario.  It's a lot better if we can convince ourselves by reasoning
> about what the code does.  That's not very reliable either, but if we
> understand the code, we at least have a hope of being able to fix the
> bugs we missed in our reasoning.

I took another look at the code and we can calculate everything upfront
before resources get distributed to hotplug bridges. I also tried and it
still works on my test systems. Does the below patch look more
acceptable to you?

diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
index 072784f55ea5..cbc80dd5e8d8 100644
--- a/drivers/pci/setup-bus.c
+++ b/drivers/pci/setup-bus.c
@@ -1881,6 +1881,7 @@ static void pci_bus_distribute_available_resources(struct pci_bus *bus,
 	unsigned int normal_bridges = 0, hotplug_bridges = 0;
 	struct resource *io_res, *mmio_res, *mmio_pref_res;
 	struct pci_dev *dev, *bridge = bus->self;
+	resource_size_t align, io_align, mem_align;
 
 	io_res = &bridge->resource[PCI_BRIDGE_RESOURCES + 0];
 	mmio_res = &bridge->resource[PCI_BRIDGE_RESOURCES + 1];
@@ -1919,27 +1920,43 @@ static void pci_bus_distribute_available_resources(struct pci_bus *bus,
 			normal_bridges++;
 	}
 
+	io_align = window_alignment(bus, IORESOURCE_IO);
+	mem_align = window_alignment(bus, IORESOURCE_MEM);
+
 	for_each_pci_bridge(dev, bus) {
-		const struct resource *res;
+		struct resource *res;
 
 		if (dev->is_hotplug_bridge)
 			continue;
 
 		/*
 		 * Reduce the available resource space by what the
-		 * bridge and devices below it occupy.
+		 * bridge and devices below it occupy taking into
+		 * account alignment if it differs from the default.
 		 */
 		res = &dev->resource[PCI_BRIDGE_RESOURCES + 0];
-		if (!res->parent && available_io > resource_size(res))
+		if (!res->parent && available_io > resource_size(res)) {
 			remaining_io -= resource_size(res);
+			align = pci_resource_alignment(dev, res);
+			if (align > io_align)
+				remaining_io -= align - io_align;
+		}
 
 		res = &dev->resource[PCI_BRIDGE_RESOURCES + 1];
-		if (!res->parent && available_mmio > resource_size(res))
+		if (!res->parent && available_mmio > resource_size(res)) {
 			remaining_mmio -= resource_size(res);
+			align = pci_resource_alignment(dev, res);
+			if (align > mem_align)
+				remaining_mmio -= align - mem_align;
+		}
 
 		res = &dev->resource[PCI_BRIDGE_RESOURCES + 2];
-		if (!res->parent && available_mmio_pref > resource_size(res))
+		if (!res->parent && available_mmio_pref > resource_size(res)) {
 			remaining_mmio_pref -= resource_size(res);
+			align = pci_resource_alignment(dev, res);
+			if (align > mem_align)
+				remaining_mmio_pref -= align - mem_align;
+		}
 	}
 
 	/*
@@ -1964,7 +1981,7 @@ static void pci_bus_distribute_available_resources(struct pci_bus *bus,
 				available_io, available_mmio,
 				available_mmio_pref);
 		} else if (dev->is_hotplug_bridge) {
-			resource_size_t align, io, mmio, mmio_pref;
+			resource_size_t io, mmio, mmio_pref;
 
 			/*
 			 * Distribute available extra resources equally
-- 
2.17.0

  reply	other threads:[~2018-05-03 12:39 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-04-16 10:34 [PATCH v5 0/9] PCI: Fixes and cleanups for native PCIe and ACPI hotplug Mika Westerberg
2018-04-16 10:34 ` [PATCH v5 1/9] PCI: Take all bridges into account when calculating bus numbers for extension Mika Westerberg
2018-04-16 10:34 ` [PATCH v5 2/9] PCI: Take bridge window alignment into account when distributing resources Mika Westerberg
2018-04-25 22:38   ` Bjorn Helgaas
2018-04-26 12:23     ` Mika Westerberg
2018-05-01 20:32       ` Bjorn Helgaas
2018-05-03 12:39         ` Mika Westerberg [this message]
2018-04-16 10:34 ` [PATCH v5 3/9] PCI: pciehp: Clear Presence Detect and Data Link Layer Status Changed on resume Mika Westerberg
2018-05-01 21:52   ` Bjorn Helgaas
2018-05-02 11:55     ` Mika Westerberg
2018-05-02 13:41       ` Bjorn Helgaas
2018-05-03 10:42         ` Mika Westerberg
2018-05-03 23:01           ` Bjorn Helgaas
2018-05-04  7:20             ` Mika Westerberg
2018-05-30 10:40             ` Lukas Wunner
2018-05-30 13:27               ` Mika Westerberg
2018-05-04  7:18     ` Lukas Wunner
2018-05-04  8:02       ` Mika Westerberg
2018-04-16 10:34 ` [PATCH v5 4/9] ACPI / hotplug / PCI: Do not scan all bridges when native PCIe hotplug is used Mika Westerberg
     [not found]   ` <20180502204932.GG11698@bhelgaas-glaptop.roam.corp.google.com>
2018-05-03 10:22     ` Mika Westerberg
2018-05-05  0:04       ` Bjorn Helgaas
2018-05-07 11:34         ` Mika Westerberg
2018-05-07 20:37           ` Bjorn Helgaas
2018-04-16 10:34 ` [PATCH v5 5/9] ACPI / hotplug / PCI: Mark stale PCI devices disconnected Mika Westerberg
2018-04-16 10:34 ` [PATCH v5 6/9] PCI: Move resource distribution for a single bridge outside of the loop Mika Westerberg
2018-04-24 23:05   ` Bjorn Helgaas
2018-04-25  7:29     ` Mika Westerberg
2018-04-16 10:34 ` [PATCH v5 7/9] PCI: Document return value of pci_scan_bridge() and pci_scan_bridge_extend() Mika Westerberg
2018-04-16 10:34 ` [PATCH v5 8/9] PCI: Improve "partially hidden behind bridge" log message Mika Westerberg
2018-04-16 10:34 ` [PATCH v5 9/9] ACPI / hotplug / PCI: Drop unnecessary parentheses Mika Westerberg

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=20180503123941.GP2355@lahna.fi.intel.com \
    --to=mika.westerberg@linux.intel.com \
    --cc=Mario.Limonciello@dell.com \
    --cc=YehezkelShB@gmail.com \
    --cc=andriy.shevchenko@linux.intel.com \
    --cc=bhelgaas@google.com \
    --cc=helgaas@kernel.org \
    --cc=lenb@kernel.org \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=lukas@wunner.de \
    --cc=michael.jamet@intel.com \
    --cc=rjw@rjwysocki.net \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.