Skip to contentSkip to conversation

Remove stale crashkernel= example from documentation

PATCH In review Olaf Hering 3 participants 7 messages Sep 4, 2019

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Remove stale crashkernel= example from documentation

A plain crashkernel=size is apparently not supported by the code anymore. In case kdump ever worked like that, the code which removed support for this notation did not update the documentation.

---
 docs/misc/kexec_and_kdump.txt | 14 ++------------
 1 file changed, 2 insertions(+), 12 deletions(-)
docs/misc/kexec_and_kdump.txt +2 -12
index 0842b3d58f..fea62ffa5c 100644
@@ -116,17 +116,7 @@ to run without disrupting the memory used by the first kernel. This area is
 called the crash kernel region and is reserved using the crashkernel
 command line parameter to the Xen hypervisor. It has two forms:
 
-  i) crashkernel=size
-
-     This is the simplest and recommended way to reserve the crash kernel
-     region. Just specify how large the region should be and the hypervisor
-     will find a good location for it. A good size to start with is 128Mb
-
-     e.g.
-
-     crashkernel=128M
-
-  ii) crashkernel=size@base
+  i) crashkernel=size@base
 
       In this form the base address is provided in addition to
       the size. Use this if auto-placement doesn't work for some reason.
@@ -136,7 +126,7 @@ command line parameter to the Xen hypervisor. It has two forms:
 
       e.g. crashkernel=128M@256M
 
-  iii) crashkernel=size,below=offset
+  ii) crashkernel=size,below=offset
 
       This allows us to place the crash kernel within the usuable address
       space without having to worry about a specific phyiscal address.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#1 lore raw
Quoted 5 lines · Olaf Hering

A plain crashkernel=size is apparently not supported by the code anymore. In case kdump ever worked like that, the code which removed support for this notation did not update the documentation.

That sounds like an accidental regression in parsing of crashkernel=, rather than a deliberate action.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#2 lore raw

Am Wed, 4 Sep 2019 10:18:41 +0100
schrieb Andrew Cooper <[email protected]>:

Quoted 2 lines

That sounds like an accidental regression in parsing of crashkernel=, rather than a deliberate action.

Maybe just the lack of b49225dc9df336405292dc08862b4c7c9d887bd6 in vendor binaries... It is likely broken since 4.10. I have not tried staging.

Olaf

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#3 lore raw
Quoted 7 lines · Olaf Hering

Am Wed, 4 Sep 2019 10:18:41 +0100
schrieb Andrew Cooper <[email protected]>:

Quoted 2 lines

That sounds like an accidental regression in parsing of crashkernel=, rather than a deliberate action.

Maybe just the lack of b49225dc9df336405292dc08862b4c7c9d887bd6 in vendor binaries...

But this change was only to deal with the bogus log message. The handling was still correct (and the option was being honored). I also can't see how this would be different now. If you've really observed the option not working, please provide more detail.

Quoted 1 line

It is likely broken since 4.10. I have not tried staging.

Because of the issue just being cosmetic, I didn't consider in necessary to backport the change.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#4 lore raw

Am Wed, 4 Sep 2019 14:19:23 +0200
schrieb Jan Beulich <[email protected]>:

Quoted 5 lines · Olaf Hering
Quoted 1 line

Maybe just the lack of b49225dc9df336405292dc08862b4c7c9d887bd6 in vendor binaries...

But this change was only to deal with the bogus log message. The handling was still correct (and the option was being honored). I also can't see how this would be different now.

Is that true? My interpretation of the code path is that no colon and nothing after a size value will lead to EINVAL. With this change any unknown string will cause EINVAL.

Olaf

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#5 lore raw
Quoted 11 lines · Olaf Hering

Am Wed, 4 Sep 2019 14:19:23 +0200
schrieb Jan Beulich <[email protected]>:

Quoted 5 lines · Olaf Hering
Quoted 1 line

Maybe just the lack of b49225dc9df336405292dc08862b4c7c9d887bd6 in vendor binaries...

But this change was only to deal with the bogus log message. The handling was still correct (and the option was being honored). I also can't see how this would be different now.

Is that true? My interpretation of the code path is that no colon and nothing after a size value will lead to EINVAL.

First of all - does "the code" here mean master/staging, or any release branch? And then, yes, on release branches there will be EINVAL, but as said before kexec_crash_area.size will get/remain set nevertheless (as the error path doesn't zap any of the earlier parsing outcome).

Quoted 1 line

With this change any unknown string will cause EINVAL.

Which is what should happen for unknown (i.e. unsupported) strings, shouldn't it?

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#6 lore raw

Am Wed, 4 Sep 2019 16:22:37 +0200
schrieb Jan Beulich <[email protected]>:

Quoted 5 lines

First of all - does "the code" here mean master/staging, or any release branch? And then, yes, on release branches there will be EINVAL, but as said before kexec_crash_area.size will get/remain set nevertheless (as the error path doesn't zap any of the earlier parsing outcome).

The code is anything since staging-4.10, in my case 4.11 and 4.12. And yes, the side effect is setting the values.

I have tried a number of Xen/dom0/kexec-tools/crashkernel= combinations. SLE12SP4 works, SLE15SP1 fails. I was looking for "crash kernel" hints in xl dmesg, only much later I spotted the "KDump:" line. The mentioned entries are not present in /proc/iomem.
All of that led to the conclusion that crashkernel=size does not work.

It turned out the tooling is not dom0 aware, it fails to translate "console=com1 com1=57600"+"console=hvc0" into "console=ttyS0,57600" for the kdump kernel. This is a common pattern of ignorance.

And finally the kdump kernel fails to initialize hardware due to low memory and SWIOTBL? errors. Once drm.ko and usbcore.ko are disabled, the first kdump ever was generated on this system.

Quoted 3 lines
Quoted 1 line

With this change any unknown string will cause EINVAL.

Which is what should happen for unknown (i.e. unsupported) strings, shouldn't it?

Yes.

Olaf

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel