diff mbox series

failing to set value to 0 in Grub2ConfigFile

Message ID alpine.LFD.2.21.1908132143580.3016@austen3.home (mailing list archive)
State New, archived
Headers show
Series failing to set value to 0 in Grub2ConfigFile | expand

Commit Message

Michael Young Aug. 13, 2019, 9:02 p.m. UTC
I have been looking at the pygrub code to see if it is possible to cope 
with grub files with BLSCFG and spotted this minor issue in GrubConf.py 
where the code intends to replace ${saved_entry} and ${next_entry} with 0 
but doesn't succeed.

Signed-off-by: Michael Young <m.a.young@durham.ac.uk>
From a08eff9b1b881dc61f9427153706e2d5b3bd0e01 Mon Sep 17 00:00:00 2001
From: Michael Young <m.a.young@durham.ac.uk>
Date: Tue, 13 Aug 2019 21:15:02 +0100
Subject: [PATCH] failing to set value to 0 in Grub2ConfigFile

In Grub2ConfigFile the code to handle ${saved_entry} and ${next_entry}
sets arg = "0" but this now does nothing following
"tools/pygrub: Make pygrub understand default entry in string format"
d1b93ea2615bd789ee28901f1f1c05ffb319cb61
which replaced arg.strip() with arg_strip in the following line.
This patch restores the previous behaviour.
---
 tools/pygrub/src/GrubConf.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Andrew Cooper Aug. 13, 2019, 9:09 p.m. UTC | #1
On 13/08/2019 22:02, YOUNG, MICHAEL A. wrote:
> I have been looking at the pygrub code to see if it is possible to cope 
> with grub files with BLSCFG and spotted this minor issue in GrubConf.py 
> where the code intends to replace ${saved_entry} and ${next_entry} with 0 
> but doesn't succeed.
>
> Signed-off-by: Michael Young <m.a.young@durham.ac.uk>

Ah - this looks suspiciously like it might be the bugfix for an issue
reported by Steven.

Steven - do you mind giving this patch a try for your "Fedora 30 DomU -
pygrub always boots the second menu option" problem?

~Andrew
From a08eff9b1b881dc61f9427153706e2d5b3bd0e01 Mon Sep 17 00:00:00 2001
From: Michael Young <m.a.young@durham.ac.uk>
Date: Tue, 13 Aug 2019 21:15:02 +0100
Subject: [PATCH] failing to set value to 0 in Grub2ConfigFile

In Grub2ConfigFile the code to handle ${saved_entry} and ${next_entry}
sets arg = "0" but this now does nothing following
"tools/pygrub: Make pygrub understand default entry in string format"
d1b93ea2615bd789ee28901f1f1c05ffb319cb61
which replaced arg.strip() with arg_strip in the following line.
This patch restores the previous behaviour.
---
 tools/pygrub/src/GrubConf.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/pygrub/src/GrubConf.py b/tools/pygrub/src/GrubConf.py
index 594139bac7..73f1bbed2f 100644
--- a/tools/pygrub/src/GrubConf.py
+++ b/tools/pygrub/src/GrubConf.py
@@ -440,7 +440,7 @@ class Grub2ConfigFile(_GrubConfigFile):
                     arg_strip = arg.strip()
                     if arg_strip == "${saved_entry}" or arg_strip == "${next_entry}":
                         logging.warning("grub2's saved_entry/next_entry not supported")
-                        arg = "0"
+                        arg_strip = "0"
                     setattr(self, self.commands[com], arg_strip)
                 else:
                     logging.info("Ignored directive %s" %(com,))
Michael Young Aug. 13, 2019, 9:51 p.m. UTC | #2
On Tue, 13 Aug 2019, Andrew Cooper wrote:

> On 13/08/2019 22:02, YOUNG, MICHAEL A. wrote:
>> I have been looking at the pygrub code to see if it is possible to cope
>> with grub files with BLSCFG and spotted this minor issue in GrubConf.py
>> where the code intends to replace ${saved_entry} and ${next_entry} with 0
>> but doesn't succeed.
>>
>> Signed-off-by: Michael Young <m.a.young@durham.ac.uk>
>
> Ah - this looks suspiciously like it might be the bugfix for an issue
> reported by Steven.
>
> Steven - do you mind giving this patch a try for your "Fedora 30 DomU -
> pygrub always boots the second menu option" problem?

Sadly I don't think it is that simple and to it properly would require 
parsing if clauses in the grub file and also reading variables from the 
grubenv file.

I do however have an idea which might work which is to ignore anything in 
if clauses, read the grubenv file (which I now have a hacky way of doing) 
and treating the value of next_entry or saved_entry as the setting for
the default kernel to pick. If I finish a patch that does this I will post
it on the list, but I very much doubt it will be of commitable quality.

 	Michael Young
Steven Haigh Aug. 14, 2019, 5:14 a.m. UTC | #3
I've had a tinker with the patch - I don't have a Fedora build system 
atm - so I just edited the file on the Dom0 and removed the pyc/pyo 
files. Same issue:

    pyGRUB  version 0.6
 
┌────────────────────────────────────────────────────────────────────────┐
 │ Fedora (5.2.6-200.fc30.x86_64) 30 (Thirty)                         
    │
 │ Fedora (0-rescue-ee4b18b1898e4bf2b36ff71077b23b5e) 30 (Thirty)     
    │
 │                                                                    
    │
 │                                                                    
    │
 │                                                                    
    │
 │                                                                    
    │
 │                                                                    
    │
 │                                                                    
    │
 
└────────────────────────────────────────────────────────────────────────┘
     Use the ^ and v keys to select which entry is highlighted.
     Press enter to boot the selected OS, 'e' to edit the
     commands before booting, 'a' to modify the kernel arguments
     before booting, or 'c' for a command line.

The rescue entry is selected in the above example.

My crappy hack has been to edit /usr/libexec/xen/bin/pygrub and add 
sel=0 as follows:

    def image_index(self):
        if isinstance(self.cf.default, int):
            sel = self.cf.default
        elif self.cf.default.isdigit():
            sel = int(self.cf.default)
            sel = 0
        else:

I know this is horrible!

I'm still disabling BLSCFG in /etc/default/grub - otherwise the pygrub 
menu is completely empty.

I don't know what the solution is right now - but I do somewhat agree 
with ignoring anything inside an if statement in grub.cfg - as the 
logic is ignored anyway. Do you still need to read the grubenv in doing 
this?

I assume the read for grubenv is to get the 'saved_entry' value?

Steven Haigh


Andrew Cooper Aug. 14, 2019, 11:17 a.m. UTC | #4
On 13/08/2019 22:02, YOUNG, MICHAEL A. wrote:
> I have been looking at the pygrub code to see if it is possible to cope 
> with grub files with BLSCFG and spotted this minor issue in GrubConf.py 
> where the code intends to replace ${saved_entry} and ${next_entry} with 0 
> but doesn't succeed.
>
> Signed-off-by: Michael Young <m.a.young@durham.ac.uk>

Irrespective of other issues, this is clearly a good bugfix.  Committed.

~Andrew
diff mbox series

Patch

diff --git a/tools/pygrub/src/GrubConf.py b/tools/pygrub/src/GrubConf.py
index 594139bac7..73f1bbed2f 100644
--- a/tools/pygrub/src/GrubConf.py
+++ b/tools/pygrub/src/GrubConf.py
@@ -440,7 +440,7 @@  class Grub2ConfigFile(_GrubConfigFile):
                     arg_strip = arg.strip()
                     if arg_strip == "${saved_entry}" or arg_strip == "${next_entry}":
                         logging.warning("grub2's saved_entry/next_entry not supported")
-                        arg = "0"
+                        arg_strip = "0"
                     setattr(self, self.commands[com], arg_strip)
                 else:
                     logging.info("Ignored directive %s" %(com,))