Re: [PATCH V2] Change ACPI IPMI support to "default y"

[Russ Anderson <rja@sgi.com>]

On Fri, Feb 21, 2014 at 02:17:12AM +0000, Zheng, Lv wrote:
> Hi,
> 
> Sorry for interrupting you.
> I have some information that may be helpful for your discussion.
> Please find them in the inlined replies.
> Well, I don't want to join the fight, just for your informations. :-)

I don't want to join the fight, either.

I have not looked at your code changes but the description
looks like the right direction.


> > From: linux-acpi-owner@vger.kernel.org [mailto:linux-acpi-owner@vger.kernel.org] On Behalf Of Russ Anderson
> > Sent: Friday, February 21, 2014 7:59 AM
> > 
> > On Thu, Feb 20, 2014 at 11:09:42PM +0000, Matthew Garrett wrote:
> > > On Thu, 2014-02-20 at 16:45 -0600, Russ Anderson wrote:
> > > > On Thu, Feb 20, 2014 at 10:26:45PM +0000, Matthew Garrett wrote:
> > >
> > > > > Because I'm trying to ensure that the default behaviour of the kernel is
> > > > > to *work*. Defaulting to having IPMI be modular means that the default
> > > > > behaviour of the kernel, as far as the ACPI spec goes, is to be broken.
> > > >
> > > > The ACPI spec requires IPMI functionality before a module loads at
> > > > boot time?  And the kernel is *broken* if it does not support ACIP IPMI
> > > > functionality before module load time?  Really?
> > >
> > > There's no mechanism to ensure that IPMI support will be loaded before
> > > ACPI calls attempt to access IPMI operation regions. Really.
> > 
> > And no mechanism can be added to ensure that ACPI call are
> > not attempted before IPMI is initialized?  A flag or lock
> > or exported symbol indicating IPMI support is ready.
> 
> In fact there is a workaround solution I've posted here:
> https://patchwork.kernel.org/patch/2831851/
> The updated version of this patch can be found at:
> https://bugzilla.kernel.org/attachment.cgi?id=112611
> It is the acpi-ipmi13.patch file.
> 
> This solution may change the meaning of ACPI spec defined _REG.
> So we may need a better solution.
> 
> But after merging this patch, it is safe to unload acpi_ipmi at users' wishes.
> Without solutions to solve region handler uninstallation races, it is not safe to unload acpi_ipmi module.
> You can see crashes in the description of this patch.
> 
> The ipmi_si module is using a different way to unload itself which has not been tested by me.
> You can find it in Documentation/IPMI.txt by searching "hot and remove of interfaces" in this file.
> 
> > 
> > > > > ACPI 4.0 includes support for IPMI operation regions. Modular IPMI means
> > > > > that the kernel will spend a significant amount of time (potentially
> > > > > until a user manually loads a driver) failing to implement part of the
> > > > > IPMI specification. That's a problem, and the correct fix is to ensure
> > > > > that the kernel always implements IPMI support.
> > > >
> > > > The ACPI spec says ipmi_si cannot be a driver?  Really?
> > > > What is the real problem you are trying to solve?
> > >
> > > The most straightforward case is that of an ACPI power meter.
> > 
> > So it is just a matter of making sure ipmi_si modules loads before
> > the ACPI power meter module loads, right?  module dependency issue.
> 
> I agree.
> I think there should be relationship between ACPI region and Linux kernel modules.
> In fact on the well-known operating system, _REG is always invoked at the end of the IRP_PNP_START_DEVICE completions.
> But we may still be able to return -EPROBE_DEFER in the power meter driver when it detects failures caused by the readiness state of the region handlers.
> 
> I didn't work further on this issue when solving the reported bug:
> https://bugzilla.kernel.org/show_bug.cgi?id=46741
> 
> > 
> > >                                                                Several
> > > vendors implement this with an IPMI operation region. Calling any of the
> > > power meter functions will trigger access to that IPMI operation region,
> > > which will fail. This may result in driver initialisation failing. There
> > > is no express dependency between the power meter driver and ipmi_si,
> > > because the spec envisages IPMI support as basic kernel functionality.
> > > It's meant to be there before you start loading any other drivers.
> > 
> > The spec "envisages"?  I get there is a dependency, that IPMI driver
> > needs to be loaded before ACIP power meter.  This isn't the first
> > case of a driver being dependent on another driver.  That doesn't
> > mean IPMI driver must be built into the kernel.
> > 
> > > > > Now, you've described some other problems. I don't disagree that those
> > > > > are problems. The correct thing for us to do with those problems is to
> > > > > fix them, not to simply change the kernel defaults such that it's
> > > > > possible for users to choose between two differently broken states. I'm
> > > > > absolutely willing to help, as long as you're willing to put some
> > > > > reasonable amount of effort into describing them.
> > > >
> > > > How about ACPI IPMI functionality starts when the ipmi_si
> > > > module loads at boot time.
> 
> Actually, I have a solution implemented this.
> You can find it at:
> https://bugzilla.kernel.org/attachment.cgi?id=112611
> It is the acpi-ipmi14.patch file.
> 
> The patch will hand the maintenance-ship of acpi_ipmi to IPMI community.
> I'm not sure it is proper to merge it by Linux upstreams.
> 
> Thanks and best regards
> -Lv
> 
> > >
> > > I've repeatedly asked for you to provide detailed descriptions of the
> > > problems you've seen because I have a genuine interest in fixing them.
> > > If you're just going to childishly refuse then this discussion is
> > > pointless.
> > 
> > The distro cases I would point you at are marked private.
> > And you do not have access to our internal support system.
> > A simple google search for "kipmi0" shows a lot of reports of
> > high cpu utilization.
> > 
> > And I'm old enough to appreciate being called childishly.  :-)
> > 
> > --
> > Russ Anderson,  Kernel and Performance Software Team Manager
> > SGI - Silicon Graphics Inc          rja@sgi.com
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
Russ Anderson,  Kernel and Performance Software Team Manager
SGI - Silicon Graphics Inc          rja@sgi.com