by Justin P. Mattock

[permalink] [raw]

[permalink] [raw]

On Fri, 18 Jun 2010 13:26:32 -0700
"Justin P. Mattock" <[email protected]> wrote:

> On 06/18/2010 01:05 PM, Jesse Barnes wrote:
> > On Fri, 18 Jun 2010 12:59:32 -0700
> > "Justin P. Mattock"<[email protected]> wrote:
> >> just added this in(as a test), and the retval warning still shows up.
> >> with the last post I just added a printk was that legit, and if so what
> >> else might be added to it to make it complete and proper?
> >
> > What's the full warning? Seems like printing the value should have
> > been enough to shut up gcc...
> >
>
> this is the warning messg after applying yinghai's patch:
>
> CC drivers/pci/setup-bus.o
> drivers/pci/setup-bus.c: In function
> 'pci_assign_unassigned_bridge_resources':
> drivers/pci/setup-bus.c:868:6: warning: variable 'retval' set but not used

Right because Yinghai's patch just sets retval but doesn't actually use
it anywhere.

> if I add a printk then gcc is content.. patch below, but not the best at
> creating printk's(the whole % thing messes me up) but here goes:
>
> From 48e15b87072c6b4286d943c55bfe2ae26d358795 Mon Sep 17 00:00:00 2001
> From: Justin P. Mattock <[email protected]>
> Date: Fri, 18 Jun 2010 13:23:27 -0700
> Subject: [PATCH 4/4] bus.c_add_print
> Signed-off-by: Justin P. Mattock <[email protected]>
>
> ---
> drivers/pci/setup-bus.c | 1 +
> 1 files changed, 1 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
> index 66cb8f4..806b766 100644
> --- a/drivers/pci/setup-bus.c
> +++ b/drivers/pci/setup-bus.c
> @@ -919,6 +919,7 @@ again:
>
> enable_all:
> retval = pci_reenable_device(bridge);
> + printk(KERN_DEBUG "PCI%d: re-enabling device\n", retval);
> pci_set_master(bridge);
> pci_enable_bridges(parent);
> }

Again, this doesn't have the if (retval) condition around the printk; I
don't want to see this message everytime regardless. Also the message
is misleading, it should be something like:
dev_err(&bridge->dev, "failed to re-enable device: %d\n", retval)
instead. PCI%d makes it look like we're talking about a specific bus
or something and not an error code.

--
Jesse Barnes, Intel Open Source Technology Center

2010-06-18 21:12:46

by Justin P. Mattock

[permalink] [raw]

Subject: Re: [PATCH 4/5]pci:setup_bus.c Fix warning: variable 'retval' set but not used

On 06/18/2010 01:46 PM, Jesse Barnes wrote:
> On Fri, 18 Jun 2010 13:26:32 -0700
> "Justin P. Mattock"<[email protected]> wrote:
>
>> On 06/18/2010 01:05 PM, Jesse Barnes wrote:
>>> On Fri, 18 Jun 2010 12:59:32 -0700
>>> "Justin P. Mattock"<[email protected]> wrote:
>>>> just added this in(as a test), and the retval warning still shows up.
>>>> with the last post I just added a printk was that legit, and if so what
>>>> else might be added to it to make it complete and proper?
>>>
>>> What's the full warning? Seems like printing the value should have
>>> been enough to shut up gcc...
>>>
>>
>> this is the warning messg after applying yinghai's patch:
>>
>> CC drivers/pci/setup-bus.o
>> drivers/pci/setup-bus.c: In function
>> 'pci_assign_unassigned_bridge_resources':
>> drivers/pci/setup-bus.c:868:6: warning: variable 'retval' set but not used
>
> Right because Yinghai's patch just sets retval but doesn't actually use
> it anywhere.
>

that's what is confusing..(not being used, but is being used, but gcc
says it's not used..) :-)

>> if I add a printk then gcc is content.. patch below, but not the best at
>> creating printk's(the whole % thing messes me up) but here goes:
>>
>> From 48e15b87072c6b4286d943c55bfe2ae26d358795 Mon Sep 17 00:00:00 2001
>> From: Justin P. Mattock<[email protected]>
>> Date: Fri, 18 Jun 2010 13:23:27 -0700
>> Subject: [PATCH 4/4] bus.c_add_print
>> Signed-off-by: Justin P. Mattock<[email protected]>
>>
>> ---
>> drivers/pci/setup-bus.c | 1 +
>> 1 files changed, 1 insertions(+), 0 deletions(-)
>>
>> diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
>> index 66cb8f4..806b766 100644
>> --- a/drivers/pci/setup-bus.c
>> +++ b/drivers/pci/setup-bus.c
>> @@ -919,6 +919,7 @@ again:
>>
>> enable_all:
>> retval = pci_reenable_device(bridge);
>> + printk(KERN_DEBUG "PCI%d: re-enabling device\n", retval);
>> pci_set_master(bridge);
>> pci_enable_bridges(parent);
>> }
>
> Again, this doesn't have the if (retval) condition around the printk; I
> don't want to see this message everytime regardless. Also the message
> is misleading, it should be something like:
> dev_err(&bridge->dev, "failed to re-enable device: %d\n", retval)
> instead. PCI%d makes it look like we're talking about a specific bus
> or something and not an error code.
>

o.k. I admit I looked at other printk's in this file to get an idea of
what I might do.. saw PCI%d and figured it would print
"PCI: re-enabling device"
but didnt think it was an error... reason for putting KERN_DEBUG.

here is what the new patch looks like:

From f910375438be06497d0524bff146c26cafca272b Mon Sep 17 00:00:00 2001
From: Justin P. Mattock <[email protected]>
Date: Fri, 18 Jun 2010 14:08:37 -0700
Subject: [PATCH 4/4] setup-pci_test
Signed-off-by: Justin P. Mattock <[email protected]>

---
drivers/pci/setup-bus.c | 3 +++
1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
index 66cb8f4..2ab5f1e 100644
--- a/drivers/pci/setup-bus.c
+++ b/drivers/pci/setup-bus.c
@@ -919,6 +919,9 @@ again:

enable_all:
retval = pci_reenable_device(bridge);
+ if (retval) {
+ dev_err(&bridge->dev, "failed to re-enable device: %d\n", retval);
+ }
pci_set_master(bridge);
pci_enable_bridges(parent);
}
--
1.7.1.rc1.21.gf3bd6

should I have put if (!retval) instead
should I put "failed to re-enable bridge device"
is there an exit code needed?

if not and all is good then I can resend this out..

Justin P. Mattock