Re: [PATCH] Assign memory resource with specified alignment.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, 24 Feb 2009 09:24:13 -0800
Jesse Barnes <jbarnes@xxxxxxxxxxxxxxxx> wrote:

> On Tuesday, February 17, 2009 2:19:39 am Yuji Shimada wrote:
> > > If resource_alignment is a per-device attribute, doesn't it make
> > > sense to put the file in /sys/bus/pci/devices/resource_alignment ?
> >
> > I don't think it is easy to add "/sys/bus/pci/devices/resource_alignment".
> >
> > In case of "/sys/bus/pci/resource_alignment", we can use BUS_ATTR.
> > In case of "/sys/bus/pci/devices/SSSS:BB:DD.F/resource_alignment", we
> > can use DEVICE_ATTR.
> >
> > But in case of "/sys/bus/pci/devices/resource_alignment", what can we
> > use?
> 
> Yeah, might be a little trickier.  Really it's a property of the bus scan & 
> resource allocation, so another option would be (with Alex's new stuff) to 
> make it part of the actual rescan command somehow.

"rescan" file will be added to "/sys/bus/pci".
If I add "resource_alignment" file to "/sys/bus/pci", both of
them will be in the same directory.
I think this is reasonable.

I'd like to add "/sys/bus/pci/resource_alignment" file.

> > > > +config RESOURCE_ALIGNMENT_PARAM
> > > > +	bool
> > > > +	depends on PCI_QUIRKS
> > > > +	default y
> > >
> > > No Kconfig help?
> >
> > I will write the help.
> >
> > > And you really want to default y?
> >
> > Yes. This does not cause the issue, I think.
> 
> How much bigger does it make the code?  I'm not even sure it's worth having a 
> config option for it; it shouldn't affect default behavior at all and doesn't 
> add a whole bunch of new stuff, so just adding it unconditionally should be 
> ok.

I will add it unconditionally.

> > > > +			if (sscanf(p, "%x:%x.%x%n",
> > > > +					&bus, &slot, &func, &count) != 3) {
> > > > +				/* Invalid format */
> > > > +				printk(KERN_ERR "PCI: Can't parse "
> > > > +					"resource_alignment parameter: %s\n",
> > >
> > > I personally prefer all printk text to be on one line to improve
> > > greppability. I think it's one of the valid reasons to break
> > > 80-column rule.
> >
> > CordingStyle in the kernel tree says "Long strings are as well broken
> > into shorter strings". I would not like to break this.
> 
> I won't argue too much about CodingStyle, but printk strings are an exception 
> to the 80 column rule, since as Alex says splitting them makes grepping for 
> error messages an exercise in frustration.

I will change the splitted text to one line text.

Thanks,
--
Yuji Shimada

--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [DMA Engine]     [Linux Coverity]     [Linux USB]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Greybus]

  Powered by Linux