Error - nsxfeval - And uncool silence from kernel hackers.
Booted 2.6.16-rc2 on my AMD x86_64 notebook and saw something new in the log (different from 2.6.15):
ACPI: Subsystem revision 20060127
ACPI: Interpreter enabled
ACPI: Using IOAPIC for interrupt routing
ACPI Error (nsxfeval-0242): Handle is NULL and Pathname is relative [20060127]
[Same error repeated 3 more times]
Searching the archives gives:
linux-2.6.16-rc1-mm2: new ACPI errors [2006-01-21]
http://marc.theaimsgroup.com/?l=linux-acpi&m=113786840630844&w=2
Wherein the reporter points to a previous display of a similar looking message (but not exactly like it) in:
Re: 2.6.12.3 clock drifting twice too fast (amd64) [2005-08-17]
http://marc.theaimsgroup.com/?l=linux-kernel&m=112428492427357&w=2
He did not receive any reply/response to his report and that is a MAJOR error as far as kernel programmer <--> kernel tester relationship goes.
If the error is harmless, then for the love of Embla just jot down a quick reply, and hide the freakin message from user eyes. Now I have to waste additional time asking the same thing: "Is this a real error? Is it OK for my dsdt to lack these _PRx objects?".
My machine doesn't seem to behave differently from 2.6.15 so a gut feeling says 'Disregard the message'. But will the message now linger on just like that PCI error message that never got any love (yeah Greg KH, I'm blaming you here):
Re: PCI error on x86_64 2.6.13 kernel [2005-11-19]
http://marc.theaimsgroup.com/?l=linux-kernel&m=113236418300012&w=2
In that message you ask me to do a git bisect, Greg, since I had seen when the message first turned up. So I did, and identified the offender as git-299de0343c7d18448a69c635378342e9214b14af After that... complete silence, not even an ACK.
And the PCI error _still_ stares a user in the face, 3 months later in 2.6.16-rc2:
PCI: Failed to allocate mem resource #6:20000@f0000000 for 0000:01:00.0
You've written a slew of documents about how users, testers and patch submitters should behave towards the linux kernel hackers, Greg. Perhaps it's time to point the pen in the other direction. Let me start the job for you:
"1) Silence is _not_ golden."
Please, we the users breathe and bleed just like you do.
As for the PCI error, I made a patch against 2.6.15-rc6 backing out the offending parts. The message went away and nothing bad happened to my machine:
arch/i386/pci/i386.c +3 -10
@@ -106,16 +106,11 @@if ((dev = bus->self)) {for (idx =PCI_BRIDGE_RESOURCES ; idx <PCI_NUM_RESOURCES ; idx++) { r = &dev->resource[idx]; -if (!r->flags) +if (!r->start)continue ; pr =pci_find_parent_resource (dev, r); -if (!r->start || !pr ||request_resource (pr, r) <0 ) { +if (!pr ||request_resource (pr, r) <0 )printk (KERN_ERR "PCI: Cannot allocate resource region %d of bridge %s\n" , idx,pci_name (dev)); -/* Something is wrong with the region. -Invalidate the resource to prevent child -resource allocations in this range. */ - r->flags =0 ; - } } }pcibios_allocate_bus_resources (&bus->children);
@@ -188,8 +183,6 @@ } } -pci_assign_unassigned_resources (); -return 0 ; }
@@ -215,7 +208,7 @@pci_read_config_word (dev,PCI_COMMAND , &cmd); old_cmd = cmd; -for (idx =0 ; idx <PCI_NUM_RESOURCES ; idx++) { +for (idx=0 ; idx<6 ; idx++) {/* Only set up the requested stuff */ if (!(mask & (1 <<idx)))continue ;
drivers/pci/setup-bus.c +0 -2
@@ -268,8 +268,6 @@for (i =0 ; i <PCI_BUS_NUM_RESOURCES ; i++) { r = bus->resource[i]; -if (r == &ioport_resource || r == &iomem_resource) -continue ;if (r && (r->flags & type_mask) == type && !r->parent)return r; }
Mvh
Mats Johannesson
--
Quoted 3 lines · Voluspa
Booted 2.6.16-rc2 on my AMD
x86_64notebook and saw something new in the log (different from 2.6.15):
So, 2.6.16-rc2 works just fine, with out your reversal of that one patch?
Quoted 3 lines
"1) Silence is _not_ golden."
Please, we the users breathe and bleed just like you do.
Sorry, and like you, we are human and drop things on the floor all the time. Sorry for not following up on this before.
Persistance is a good trait to have when dealing with overworked kernel developers :)
thanks,
greg k-h
Quoted 7 lines · On Thu, 9 Feb 2006 15:32:33 -0800 Greg KH wrote:
Quoted 3 lines · Voluspa
Booted 2.6.16-rc2 on my AMD
x86_64notebook and saw something new in the log (different from 2.6.15):So, 2.6.16-rc2 works just fine, with out your reversal of that one patch?
Eh, I bundled two issues in my rant. 1) nsxfeval, which the ACPI guys have confirmed is pure noise and can be forgotten. 2) "PCI: Failed to allocate mem resource" which has been there since the commit I pointed to, and never has effected the machine negatively (as far as I can tell). So yes, 2.6.16-rc2 seems fine, and I don't patch it for that PCI thing. It's just that the message is present and will worry people 'forever' if nothing is done.
There is another issue, but it's not a regression. I'll file a bugzilla for that in a couple of days. Or a week...
Quoted 9 lines
Quoted 3 lines
"1) Silence is _not_ golden."
Please, we the users breathe and bleed just like you do.
Sorry, and like you, we are human and drop things on the floor all the time. Sorry for not following up on this before.
Persistance is a good trait to have when dealing with overworked kernel developers :)
Yeah, I know, sorry if I jabbed too hard. I was caught up in a sadness for anyone who's mail is ignored. There is a form of humiliation when that happens on a high profile list like lkml or other kernel related lists. A mirror of the defeat can be found on a small server in Korpilombolo for the next two decades at least.
But I must say that during the years that I've read lkml, it _feels_ as if the reply-rate has improved significantly. The list is not as terrifying as it used to be.
Mvh
Mats Johannesson
--
Quoted 9 lines · Voluspa
Quoted 7 lines · On Thu, 9 Feb 2006 15:32:33 -0800 Greg KH wrote:
Quoted 3 lines · Voluspa
Booted 2.6.16-rc2 on my AMD
x86_64notebook and saw something new in the log (different from 2.6.15):So, 2.6.16-rc2 works just fine, with out your reversal of that one patch?
Quoted 7 lines
Eh, I bundled two issues in my rant. 1) nsxfeval, which the ACPI guys have confirmed is pure noise and can be forgotten. 2) "PCI: Failed to allocate mem resource" which has been there since the commit I pointed to, and never has effected the machine negatively (as far as I can tell). So yes, 2.6.16-rc2 seems fine, and I don't patch it for that PCI thing. It's just that the message is present and will worry people 'forever' if nothing is done.
As long as everything is working properly, you can happily ignore the issue and blame your bios vendor for making such a stupid mistake :)
Quoted 5 lines
Yeah, I know, sorry if I jabbed too hard. I was caught up in a sadness for anyone who's mail is ignored. There is a form of humiliation when that happens on a high profile list like lkml or other kernel related lists. A mirror of the defeat can be found on a small server in Korpilombolo for the next two decades at least.
I know it's tough when things like this are ignored, and no, you didn't jab to hard at all. We all need to be reminded when we forget things like this, myself especially at times :)
Quoted 3 lines
But I must say that during the years that I've read lkml, it _feels_ as if the reply-rate has improved significantly. The list is not as terrifying as it used to be.
Glad to hear it, we are trying here...
thanks,
greg k-h
Quoted 18 lines · On Thu, 9 Feb 2006 20:24:19 -0800 Greg KH wrote:
Quoted 9 lines · Voluspa
Quoted 7 lines · On Thu, 9 Feb 2006 15:32:33 -0800 Greg KH wrote:
Quoted 3 lines · Voluspa
Booted 2.6.16-rc2 on my AMD
x86_64notebook and saw something new in the log (different from 2.6.15):So, 2.6.16-rc2 works just fine, with out your reversal of that one patch?
Quoted 7 lines
Eh, I bundled two issues in my rant. 1) nsxfeval, which the ACPI guys have confirmed is pure noise and can be forgotten. 2) "PCI: Failed to allocate mem resource" which has been there since the commit I pointed to, and never has effected the machine negatively (as far as I can tell). So yes, 2.6.16-rc2 seems fine, and I don't patch it for that PCI thing. It's just that the message is present and will worry people 'forever' if nothing is done.
Quoted 2 lines
As long as everything is working properly, you can happily ignore the issue and blame your bios vendor for making such a stupid mistake :)
The issue came bouncing in from fedora-list to lkml. I've never cared one way or the other. And while I have no love for a BIOS maker (Phoenix) that only give you options like "Are you happy with the settings as is: Yes/Yes", the occasional kernel hacker has also been seen wearing brown paper bags.
Something like a quirk/white/black-list could be used to make it invisible. But as I said, me not really care (and neither do you ;-) So let's just bury this thread.
Mvh
Mats Johannesson
--
Quoted 11 lines · Greg KH
Quoted 7 lines · Voluspa
Yeah, I know, sorry if I jabbed too hard. I was caught up in a sadness for anyone who's mail is ignored. There is a form of humiliation when that happens on a high profile list like lkml or other kernel related lists. A mirror of the defeat can be found on a small server in Korpilombolo for the next two decades at least.
Quoted 5 lines
I know it's tough when things like this are ignored, and no, you didn't jab to hard at all. We all need to be reminded when we forget things like this, myself especially at times :)
Well, I still got an unanswered message here: http://marc.theaimsgroup.com/?l=linux-acpi&m=113924047806727&w=2
:-)
Might have been such a stupid question as not to warrant any attention, but even a "oh please leave us alone" would be better than this dratted silence hehe
Cheers,
-Tako
Quoted 3 lines · Tako Schotanus
Might have been such a stupid question as not to warrant any attention, but even a "oh please leave us alone" would be better than this dratted silence hehe
Don't be sad, it's probably just that nobody knows how to tackle the problem.
Quoted 9 lines · Erik Slagter
Quoted 2 lines · Tako Schotanus
Might have been such a stupid question as not to warrant any attention, but even a "oh please leave us alone" would be better than this dratted silence hehe
Don't be sad, it's probably just that nobody knows how to tackle the problem.
Ok, I'll try not to be :-)
But nobody who has even a hint or a tip where I could begin to look for an answer?
For example: IS there actually a difference between a system that has been shut down and woken up again and a system that has been reset? I just don't understand how I can get these errors when waking up the system but that they go away if I quickly press the reset button after waking the system.
Clutching at straws here! :-)
Cheers,
-Tako