[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <1340915108.2411.26.camel@zim.stowe>
Date: Thu, 28 Jun 2012 14:25:08 -0600
From: Myron Stowe <mstowe@...hat.com>
To: Bjorn Helgaas <bhelgaas@...gle.com>
Cc: Myron Stowe <myron.stowe@...hat.com>, linux-pci@...r.kernel.org,
linux@....linux.org.uk, ralf@...ux-mips.org, tglx@...utronix.de,
mingo@...hat.com, hpa@...or.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/9] PCI: Remove redundant debug output in pci_do_fixups
On Tue, 2012-06-26 at 16:26 -0600, Bjorn Helgaas wrote:
> On Thu, Jun 21, 2012 at 2:24 PM, Myron Stowe <myron.stowe@...hat.com> wrote:
> > When the boot argument 'initcall_debug' is specified, redundant debug
> > output occurs for each device as a quirk is applied:
> > ...
> > pci 0000:00:1a.0: calling quirk_usb_early_handoff+0x0/0x620
> > calling quirk_usb_early_handoff+0x0/0x620 @ 1 for 0000:00:1a.0
> > pci fixup quirk_usb_early_handoff+0x0/0x620 returned after 32 usecs for 0000:00: 1a.0
> > ...
> >
> > This patch removes the redundancy by eliminating the first debug output
> > occurence in the sequence shown above when 'initcall_debug' is specified.
>
> Here's what I don't like about this: adding "initcall_debug" *removes*
> some output. My expectation is that it would only *add* output.
Well, the point to me was that it only removed output that was
completely redundant when "initcall_debug" was true.
I *really* wanted to change the "initcall_debug" output to follow the
convention already in place via 'dev_dbg()' but did not as I'm vaguely
aware of some scripting that parses "initcall_debug" output and did not
want to break that.
>
> > Signed-off-by: Myron Stowe <myron.stowe@...hat.com>
> > ---
> >
> > drivers/pci/quirks.c | 5 +++--
> > 1 files changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
> > index a2d9d33..9c93558 100644
> > --- a/drivers/pci/quirks.c
> > +++ b/drivers/pci/quirks.c
> > @@ -2953,11 +2953,12 @@ static void pci_do_fixups(struct pci_dev *dev, struct pci_fixup *f,
> > f->vendor == (u16) PCI_ANY_ID) &&
> > (f->device == dev->device ||
> > f->device == (u16) PCI_ANY_ID)) {
> > - dev_dbg(&dev->dev, "calling %pF\n", f->hook);
> > if (initcall_debug)
> > do_one_fixup_debug(f->hook, dev);
> > - else
> > + else {
> > + dev_dbg(&dev->dev, "calling %pF\n", f->hook);
> > f->hook(dev);
>
> This part isn't something you changed, but I also think it's a bit
> ugly that we have two possible call sites for the quirk: either inside
> do_one_fixup_debug() or directly in pci_do_fixups(). I wonder if this
> could be restructured a bit in the style of initcall_debug_start() and
> initcall_debug_report(), so we could have this:
>
> ktime_t calltime;
>
> calltime = initcall_debug_start(dev);
> f->hook(dev);
> initcall_debug_report(dev, calltime);
>
> where initcall_debug_report() would only print something when
> initcall_debug is enabled.
>
Nice observation! Such a change would yield a better result. Perhaps,
transitioning to this approach would also eventually help with my
earlier peeve concerning the redundant output.
Thanks,
Myron
> > + }
> > }
> > }
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists