public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Bjorn Helgaas <bjorn.helgaas@hp.com>
To: Len Brown <lenb@kernel.org>
To: Andi Kleen <andi@firstfloor.org>
Cc: linux-acpi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: Adam Belay <ambx1@neo.rr.com>
Cc: Adam M Belay <abelay@mit.edu>
Cc: Li Shaohua <shaohua.li@intel.com>
Cc: Matthieu Castet <castet.matthieu@free.fr>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Rene Herman <rene.herman@keyaccess.nl>
Cc: Jaroslav Kysela <perex@perex.cz>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Takashi Iwai <tiwai@suse.de>
Cc: Jiri Slaby <jirislaby@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Subject: [patch 07/28] PNPACPI: keep disabled resources when parsing current config
Date: Fri, 27 Jun 2008 16:56:58 -0600	[thread overview]
Message-ID: <20080627225759.691183081@ldl.fc.hp.com> (raw)
In-Reply-To: 20080627225651.663174474@ldl.fc.hp.com

[-- Attachment #1: pnp-keep-disabled-resources-when-parsing-current-config --]
[-- Type: text/plain, Size: 3820 bytes --]

When we parse a device's _CRS data (the current resource settings),
we should keep track of everything we find, even if it's currently
disabled or invalid.

This is what we already do for ISAPNP and PNPBIOS, and it helps
keep things matched up when we subsequently re-encode resources.
For example, consider a device with (mem, irq0, irq1, io), where
irq0 is disabled.  If we drop irq0 when parsing the _CRS, we will
mistakenly put irq1 in the irq0 slot when we encode resources
for an _SRS call.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>

Index: work14/drivers/pnp/pnpacpi/rsparser.c
===================================================================
--- work14.orig/drivers/pnp/pnpacpi/rsparser.c	2008-06-17 14:47:29.000000000 -0600
+++ work14/drivers/pnp/pnpacpi/rsparser.c	2008-06-17 15:46:52.000000000 -0600
@@ -98,8 +98,10 @@ static void pnpacpi_parse_allocated_irqr
 	int irq, flags;
 	int p, t;
 
-	if (!valid_IRQ(gsi))
+	if (!valid_IRQ(gsi)) {
+		pnp_add_irq_resource(dev, gsi, IORESOURCE_DISABLED);
 		return;
+	}
 
 	/*
 	 * in IO-APIC mode, use overrided attribute. Two reasons:
@@ -248,24 +250,39 @@ static acpi_status pnpacpi_allocated_res
 		 * _CRS, but some firmware violates this, so parse them all.
 		 */
 		irq = &res->data.irq;
-		for (i = 0; i < irq->interrupt_count; i++) {
-			pnpacpi_parse_allocated_irqresource(dev,
-				irq->interrupts[i],
-				irq->triggering,
-				irq->polarity,
-				irq->sharable);
+		if (irq->interrupt_count == 0)
+			pnp_add_irq_resource(dev, 0, IORESOURCE_DISABLED);
+		else {
+			for (i = 0; i < irq->interrupt_count; i++) {
+				pnpacpi_parse_allocated_irqresource(dev,
+					irq->interrupts[i],
+					irq->triggering,
+					irq->polarity,
+				    irq->sharable);
+			}
+
+			/*
+			 * The IRQ encoder puts a single interrupt in each
+			 * descriptor, so if a _CRS descriptor has more than
+			 * one interrupt, we won't be able to re-encode it.
+			 */
+			if (pnp_can_write(dev) && irq->interrupt_count > 1) {
+				dev_warn(&dev->dev, "multiple interrupts in "
+					 "_CRS descriptor; configuration can't "
+					 "be changed\n");
+				dev->capabilities &= ~PNP_WRITE;
+			}
 		}
 		break;
 
 	case ACPI_RESOURCE_TYPE_DMA:
 		dma = &res->data.dma;
-		if (dma->channel_count > 0) {
+		if (dma->channel_count > 0 && dma->channels[0] != (u8) -1)
 			flags = dma_flags(dma->type, dma->bus_master,
 					  dma->transfer);
-			if (dma->channels[0] == (u8) -1)
-				flags |= IORESOURCE_DISABLED;
-			pnp_add_dma_resource(dev, dma->channels[0], flags);
-		}
+		else
+			flags = IORESOURCE_DISABLED;
+		pnp_add_dma_resource(dev, dma->channels[0], flags);
 		break;
 
 	case ACPI_RESOURCE_TYPE_IO:
@@ -331,12 +348,29 @@ static acpi_status pnpacpi_allocated_res
 		if (extended_irq->producer_consumer == ACPI_PRODUCER)
 			return AE_OK;
 
-		for (i = 0; i < extended_irq->interrupt_count; i++) {
-			pnpacpi_parse_allocated_irqresource(dev,
-				extended_irq->interrupts[i],
-				extended_irq->triggering,
-				extended_irq->polarity,
-				extended_irq->sharable);
+		if (extended_irq->interrupt_count == 0)
+			pnp_add_irq_resource(dev, 0, IORESOURCE_DISABLED);
+		else {
+			for (i = 0; i < extended_irq->interrupt_count; i++) {
+				pnpacpi_parse_allocated_irqresource(dev,
+					extended_irq->interrupts[i],
+					extended_irq->triggering,
+					extended_irq->polarity,
+					extended_irq->sharable);
+			}
+
+			/*
+			 * The IRQ encoder puts a single interrupt in each
+			 * descriptor, so if a _CRS descriptor has more than
+			 * one interrupt, we won't be able to re-encode it.
+			 */
+			if (pnp_can_write(dev) &&
+			    extended_irq->interrupt_count > 1) {
+				dev_warn(&dev->dev, "multiple interrupts in "
+					 "_CRS descriptor; configuration can't "
+					 "be changed\n");
+				dev->capabilities &= ~PNP_WRITE;
+			}
 		}
 		break;
 

-- 

  parent reply	other threads:[~2008-06-27 23:02 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-06-27 22:56 [patch 00/28] PNP: convert fixed tables to lists, v4 Bjorn Helgaas
2008-06-27 22:56 ` [patch 01/28] PNP: add detail to debug resource dump Bjorn Helgaas
2008-06-27 22:56 ` [patch 02/28] PNP: remove pnp_resource.index Bjorn Helgaas
2008-06-27 22:56 ` [patch 03/28] PNP: add pnp_resource_type() internal interface Bjorn Helgaas
2008-06-27 22:56 ` [patch 04/28] PNP: add pnp_resource_type_name() helper function Bjorn Helgaas
2008-06-27 22:56 ` [patch 05/28] PNP: make pnp_{port,mem,etc}_start(), et al work for invalid resources Bjorn Helgaas
2008-06-27 22:56 ` [patch 06/28] PNP: replace pnp_resource_table with dynamically allocated resources Bjorn Helgaas
2008-06-27 22:56 ` Bjorn Helgaas [this message]
2008-06-27 22:56 ` [patch 08/28] PNP: remove ratelimit on add resource failures Bjorn Helgaas
2008-06-27 22:57 ` [patch 09/28] PNP: dont sort by type in /sys/.../resources Bjorn Helgaas
2008-06-27 22:57 ` [patch 10/28] PNP: add pnp_possible_config() -- can a device could be configured this way? Bjorn Helgaas
2008-06-27 22:57 ` [patch 11/28] PNP: whitespace/coding style fixes Bjorn Helgaas
2008-06-28 19:56   ` Joe Perches
2008-06-28 20:02     ` Rene Herman
2008-06-27 22:57 ` [patch 12/28] PNP: define PNP-specific IORESOURCE_IO_* flags alongside IRQ, DMA, MEM Bjorn Helgaas
2008-06-27 22:57 ` [patch 13/28] PNP: make resource option structures private to PNP subsystem Bjorn Helgaas
2008-06-27 22:57 ` [patch 14/28] PNP: introduce pnp_irq_mask_t typedef Bjorn Helgaas
2008-06-27 22:57 ` [patch 15/28] PNP: increase I/O port & memory option address sizes Bjorn Helgaas
2008-06-27 22:57 ` [patch 16/28] PNP: improve resource assignment debug Bjorn Helgaas
2008-06-27 22:57 ` [patch 17/28] PNP: in debug resource dump, make empty list obvious Bjorn Helgaas
2008-06-27 22:57 ` [patch 18/28] PNP: make resource assignment functions return 0 (success) or -EBUSY (failure) Bjorn Helgaas
2008-06-27 22:57 ` [patch 19/28] PNP: remove redundant pnp_can_configure() check Bjorn Helgaas
2008-06-27 22:57 ` [patch 20/28] PNP: centralize resource option allocations Bjorn Helgaas
2008-06-27 22:57 ` [patch 21/28] PNPACPI: ignore _PRS interrupt numbers larger than PNP_IRQ_NR Bjorn Helgaas
2008-06-27 22:57 ` [patch 22/28] PNP: rename pnp_register_*_resource() local variables Bjorn Helgaas
2008-06-27 22:57 ` [patch 23/28] PNP: support optional IRQ resources Bjorn Helgaas
2008-06-27 22:57 ` [patch 24/28] PNP: remove extra 0x100 bit from option priority Bjorn Helgaas
2008-06-27 22:57 ` [patch 25/28] ISAPNP: handle independent options following dependent ones Bjorn Helgaas
2008-06-27 22:57 ` [patch 26/28] PNP: convert resource options to single linked list Bjorn Helgaas
2008-06-27 22:57 ` [patch 27/28] PNP: avoid legacy IDE IRQs Bjorn Helgaas
2008-06-27 22:57 ` [patch 28/28] PNPACPI: add support for HP vendor-specific CCSR descriptors Bjorn Helgaas
2008-06-28  4:36 ` [patch 00/28] PNP: convert fixed tables to lists, v4 Len Brown
  -- strict thread matches above, loose matches on Subject: below --
2008-06-17 22:58 [patch 00/28] PNP: convert fixed tables to lists, v3 Bjorn Helgaas
2008-06-17 22:58 ` [patch 07/28] PNPACPI: keep disabled resources when parsing current config Bjorn Helgaas

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=20080627225759.691183081@ldl.fc.hp.com \
    --to=bjorn.helgaas@hp.com \
    --cc=lenb@kernel.org \
    --cc=linux-acpi@vger.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