Patchwork [1/2] acpiphp: create companion ACPI devices before creating PCI devices

login
register
mail settings
Submitter Jiang Liu
Date Jan. 30, 2013, 4:10 p.m.
Message ID <1359562210-26536-1-git-send-email-jiang.liu@huawei.com>
Download mbox | patch
Permalink /patch/216940/
State Accepted
Headers show

Comments

Jiang Liu - Jan. 30, 2013, 4:10 p.m.
From: Jiang Liu <jiang.liu@huawei.com>

With commit 4f535093cf8f6da8c "PCI: Put pci_dev in device tree as
early as possible", companion ACPI devices should be created before
creating correspoding PCI devices, otherwise it will break the ACPI
PCI binding logic.

Signed-off-by: Jiang Liu <jiang.liu@huawei.com>
---
Hi Bjorn,
	This patch set applies to your pci/yinghai-root-bus branch.
There are still other potential bugs in the acpiphp driver under
investigation. So I will send out these two first to catch up with
the 3.9 merging window.
	Thanks!
	Gerry
---
 drivers/pci/hotplug/acpiphp_glue.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
Bjorn Helgaas - Feb. 1, 2013, 11:06 p.m.
On Wed, Jan 30, 2013 at 9:10 AM, Jiang Liu <liuj97@gmail.com> wrote:
> From: Jiang Liu <jiang.liu@huawei.com>
>
> With commit 4f535093cf8f6da8c "PCI: Put pci_dev in device tree as
> early as possible", companion ACPI devices should be created before
> creating correspoding PCI devices, otherwise it will break the ACPI
> PCI binding logic.
>
> Signed-off-by: Jiang Liu <jiang.liu@huawei.com>
> ---
> Hi Bjorn,
>         This patch set applies to your pci/yinghai-root-bus branch.
> There are still other potential bugs in the acpiphp driver under
> investigation. So I will send out these two first to catch up with
> the 3.9 merging window.
>         Thanks!
>         Gerry

I assume this fixes some sort of user-visible issue with acpiphp.
What failure does the user see?

> ---
>  drivers/pci/hotplug/acpiphp_glue.c |    6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c
> index bd784ff..acb7af2 100644
> --- a/drivers/pci/hotplug/acpiphp_glue.c
> +++ b/drivers/pci/hotplug/acpiphp_glue.c
> @@ -839,6 +839,9 @@ static int __ref enable_device(struct acpiphp_slot *slot)
>         if (slot->flags & SLOT_ENABLED)
>                 goto err_exit;
>
> +       list_for_each_entry(func, &slot->funcs, sibling)
> +               acpiphp_bus_add(func);
> +
>         num = pci_scan_slot(bus, PCI_DEVFN(slot->device, 0));
>         if (num == 0) {
>                 /* Maybe only part of funcs are added. */
> @@ -862,9 +865,6 @@ static int __ref enable_device(struct acpiphp_slot *slot)
>                 }
>         }
>
> -       list_for_each_entry(func, &slot->funcs, sibling)
> -               acpiphp_bus_add(func);
> -
>         pci_bus_assign_resources(bus);
>         acpiphp_sanitize_bus(bus);
>         acpiphp_set_hpp_values(bus);
> --
> 1.7.9.5
>
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Bjorn Helgaas - Feb. 1, 2013, 11:17 p.m.
On Fri, Feb 1, 2013 at 4:06 PM, Bjorn Helgaas <bhelgaas@google.com> wrote:
> On Wed, Jan 30, 2013 at 9:10 AM, Jiang Liu <liuj97@gmail.com> wrote:
>> From: Jiang Liu <jiang.liu@huawei.com>
>>
>> With commit 4f535093cf8f6da8c "PCI: Put pci_dev in device tree as
>> early as possible", companion ACPI devices should be created before
>> creating correspoding PCI devices, otherwise it will break the ACPI
>> PCI binding logic.
>>
>> Signed-off-by: Jiang Liu <jiang.liu@huawei.com>

Applied to pci/yinghai-root-bus for v3.9.  Thanks!

>> ---
>> Hi Bjorn,
>>         This patch set applies to your pci/yinghai-root-bus branch.
>> There are still other potential bugs in the acpiphp driver under
>> investigation. So I will send out these two first to catch up with
>> the 3.9 merging window.
>>         Thanks!
>>         Gerry
>
> I assume this fixes some sort of user-visible issue with acpiphp.
> What failure does the user see?
>
>> ---
>>  drivers/pci/hotplug/acpiphp_glue.c |    6 +++---
>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c
>> index bd784ff..acb7af2 100644
>> --- a/drivers/pci/hotplug/acpiphp_glue.c
>> +++ b/drivers/pci/hotplug/acpiphp_glue.c
>> @@ -839,6 +839,9 @@ static int __ref enable_device(struct acpiphp_slot *slot)
>>         if (slot->flags & SLOT_ENABLED)
>>                 goto err_exit;
>>
>> +       list_for_each_entry(func, &slot->funcs, sibling)
>> +               acpiphp_bus_add(func);
>> +
>>         num = pci_scan_slot(bus, PCI_DEVFN(slot->device, 0));
>>         if (num == 0) {
>>                 /* Maybe only part of funcs are added. */
>> @@ -862,9 +865,6 @@ static int __ref enable_device(struct acpiphp_slot *slot)
>>                 }
>>         }
>>
>> -       list_for_each_entry(func, &slot->funcs, sibling)
>> -               acpiphp_bus_add(func);
>> -
>>         pci_bus_assign_resources(bus);
>>         acpiphp_sanitize_bus(bus);
>>         acpiphp_set_hpp_values(bus);
>> --
>> 1.7.9.5
>>
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Yijing Wang - Feb. 2, 2013, 2:24 a.m.
On 2013/2/2 7:06, Bjorn Helgaas wrote:
> On Wed, Jan 30, 2013 at 9:10 AM, Jiang Liu <liuj97@gmail.com> wrote:
>> From: Jiang Liu <jiang.liu@huawei.com>
>>
>> With commit 4f535093cf8f6da8c "PCI: Put pci_dev in device tree as
>> early as possible", companion ACPI devices should be created before
>> creating correspoding PCI devices, otherwise it will break the ACPI
>> PCI binding logic.
>>
>> Signed-off-by: Jiang Liu <jiang.liu@huawei.com>
>> ---
>> Hi Bjorn,
>>         This patch set applies to your pci/yinghai-root-bus branch.
>> There are still other potential bugs in the acpiphp driver under
>> investigation. So I will send out these two first to catch up with
>> the 3.9 merging window.
>>         Thanks!
>>         Gerry
> 
> I assume this fixes some sort of user-visible issue with acpiphp.
> What failure does the user see?

Hi Bjorn,
   Before applied this patch, after hot remove and hot add pci device by acpiphp,
we can find firmware_node under the pci device was missed.

Detailed info as bellow:
linux-ha2:/sys/bus/pci/slots # ls /sys/bus/pci/devices/0000\:05\:00.0/ -al
total 0
drwxr-xr-x 3 root root         0 Jan 31 10:19 .
drwxr-xr-x 9 root root         0 Jan 31 10:19 ..
-r--r--r-- 1 root root     65536 Jan 31 10:31 boot_vga
-rw-r--r-- 1 root root     65536 Jan 31 10:31 broken_parity_status
-r--r--r-- 1 root root     65536 Jan 31 10:19 class
-rw-r--r-- 1 root root      4096 Jan 31 10:31 config
-r--r--r-- 1 root root     65536 Jan 31 10:31 consistent_dma_mask_bits
-rw-r--r-- 1 root root     65536 Jan 31 10:31 d3cold_allowed
-r--r--r-- 1 root root     65536 Jan 31 10:19 device
-r--r--r-- 1 root root     65536 Jan 31 10:31 dma_mask_bits
-rw------- 1 root root     65536 Jan 31 10:31 enable
lrwxrwxrwx 1 root root         0 Jan 31 10:31 firmware_node -> ../../../LNXSYSTM:00/device:00/ACPI0004:00/PNP0A08:00/device:07/device:08
-r--r--r-- 1 root root     65536 Jan 31 10:19 irq
-r--r--r-- 1 root root     65536 Jan 31 10:31 local_cpulist
-r--r--r-- 1 root root     65536 Jan 31 10:20 local_cpus
-r--r--r-- 1 root root     65536 Jan 31 10:31 modalias
-rw-r--r-- 1 root root     65536 Jan 31 10:31 msi_bus
-r--r--r-- 1 root root     65536 Jan 31 10:31 numa_node
drwxr-xr-x 2 root root         0 Jan 31 10:31 power
--w--w---- 1 root root     65536 Jan 31 10:31 remove
--w--w---- 1 root root     65536 Jan 31 10:31 rescan
-r--r--r-- 1 root root     65536 Jan 31 10:19 resource
-rw------- 1 root root  16777216 Jan 31 10:31 resource0
-rw------- 1 root root 268435456 Jan 31 10:31 resource1
-rw------- 1 root root 268435456 Jan 31 10:31 resource1_wc
-rw------- 1 root root  33554432 Jan 31 10:31 resource3
-rw------- 1 root root  33554432 Jan 31 10:31 resource3_wc
-rw------- 1 root root       128 Jan 31 10:31 resource5
-rw------- 1 root root    524288 Jan 31 10:31 rom
lrwxrwxrwx 1 root root         0 Jan 31 10:19 subsystem -> ../../../../bus/pci
-r--r--r-- 1 root root     65536 Jan 31 10:19 subsystem_device
-r--r--r-- 1 root root     65536 Jan 31 10:19 subsystem_vendor
-rw-r--r-- 1 root root     65536 Jan 31 10:19 uevent
-r--r--r-- 1 root root     65536 Jan 31 10:19 vendor
linux-ha2:/sys/bus/pci/slots # cd 0
linux-ha2:/sys/bus/pci/slots/0 # cat address
0000:05:00
linux-ha2:/sys/bus/pci/slots/0 # cat power
1
linux-ha2:/sys/bus/pci/slots/0 # echo 0 > power
linux-ha2:/sys/bus/pci/slots/0 # echo 1 > power
linux-ha2:/sys/bus/pci/slots/0 # ls /sys/bus/pci/devices/0000\:05\:00.0/ -al
total 0
drwxr-xr-x 3 root root         0 Jan 31 10:31 .
drwxr-xr-x 9 root root         0 Jan 31 10:19 ..
-r--r--r-- 1 root root     65536 Jan 31 10:31 boot_vga
-rw-r--r-- 1 root root     65536 Jan 31 10:31 broken_parity_status
-r--r--r-- 1 root root     65536 Jan 31 10:31 class
-rw-r--r-- 1 root root      4096 Jan 31 10:31 config
-r--r--r-- 1 root root     65536 Jan 31 10:31 consistent_dma_mask_bits
-rw-r--r-- 1 root root     65536 Jan 31 10:31 d3cold_allowed
-r--r--r-- 1 root root     65536 Jan 31 10:31 device
-r--r--r-- 1 root root     65536 Jan 31 10:31 dma_mask_bits
-rw------- 1 root root     65536 Jan 31 10:31 enable
-r--r--r-- 1 root root     65536 Jan 31 10:31 irq
-r--r--r-- 1 root root     65536 Jan 31 10:31 local_cpulist
-r--r--r-- 1 root root     65536 Jan 31 10:31 local_cpus
-r--r--r-- 1 root root     65536 Jan 31 10:31 modalias
-rw-r--r-- 1 root root     65536 Jan 31 10:31 msi_bus
-r--r--r-- 1 root root     65536 Jan 31 10:31 numa_node
drwxr-xr-x 2 root root         0 Jan 31 10:31 power
--w--w---- 1 root root     65536 Jan 31 10:31 remove
--w--w---- 1 root root     65536 Jan 31 10:31 rescan
-r--r--r-- 1 root root     65536 Jan 31 10:31 resource
-rw------- 1 root root  16777216 Jan 31 10:31 resource0
-rw------- 1 root root 268435456 Jan 31 10:31 resource1
-rw------- 1 root root 268435456 Jan 31 10:31 resource1_wc
-rw------- 1 root root  33554432 Jan 31 10:31 resource3
-rw------- 1 root root  33554432 Jan 31 10:31 resource3_wc
-rw------- 1 root root       128 Jan 31 10:31 resource5
-rw------- 1 root root    524288 Jan 31 10:31 rom
lrwxrwxrwx 1 root root         0 Jan 31 10:31 subsystem -> ../../../../bus/pci
-r--r--r-- 1 root root     65536 Jan 31 10:31 subsystem_device
-r--r--r-- 1 root root     65536 Jan 31 10:31 subsystem_vendor
-rw-r--r-- 1 root root     65536 Jan 31 10:31 uevent
-r--r--r-- 1 root root     65536 Jan 31 10:31 vendor
linux-ha2:/sys/bus/pci/slots/0 #

So after hot add pci device again, we lost the firmware_node link under device.
firmware_node -> ../../../LNXSYSTM:00/device:00/ACPI0004:00/PNP0A08:00/device:07/device:08


> 
>> ---
>>  drivers/pci/hotplug/acpiphp_glue.c |    6 +++---
>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c
>> index bd784ff..acb7af2 100644
>> --- a/drivers/pci/hotplug/acpiphp_glue.c
>> +++ b/drivers/pci/hotplug/acpiphp_glue.c
>> @@ -839,6 +839,9 @@ static int __ref enable_device(struct acpiphp_slot *slot)
>>         if (slot->flags & SLOT_ENABLED)
>>                 goto err_exit;
>>
>> +       list_for_each_entry(func, &slot->funcs, sibling)
>> +               acpiphp_bus_add(func);
>> +
>>         num = pci_scan_slot(bus, PCI_DEVFN(slot->device, 0));
>>         if (num == 0) {
>>                 /* Maybe only part of funcs are added. */
>> @@ -862,9 +865,6 @@ static int __ref enable_device(struct acpiphp_slot *slot)
>>                 }
>>         }
>>
>> -       list_for_each_entry(func, &slot->funcs, sibling)
>> -               acpiphp_bus_add(func);
>> -
>>         pci_bus_assign_resources(bus);
>>         acpiphp_sanitize_bus(bus);
>>         acpiphp_set_hpp_values(bus);
>> --
>> 1.7.9.5
>>
> 
>

Patch

diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c
index bd784ff..acb7af2 100644
--- a/drivers/pci/hotplug/acpiphp_glue.c
+++ b/drivers/pci/hotplug/acpiphp_glue.c
@@ -839,6 +839,9 @@  static int __ref enable_device(struct acpiphp_slot *slot)
 	if (slot->flags & SLOT_ENABLED)
 		goto err_exit;
 
+	list_for_each_entry(func, &slot->funcs, sibling)
+		acpiphp_bus_add(func);
+
 	num = pci_scan_slot(bus, PCI_DEVFN(slot->device, 0));
 	if (num == 0) {
 		/* Maybe only part of funcs are added. */
@@ -862,9 +865,6 @@  static int __ref enable_device(struct acpiphp_slot *slot)
 		}
 	}
 
-	list_for_each_entry(func, &slot->funcs, sibling)
-		acpiphp_bus_add(func);
-
 	pci_bus_assign_resources(bus);
 	acpiphp_sanitize_bus(bus);
 	acpiphp_set_hpp_values(bus);