[Crash-utility] Re: [PATCH] arm64: fix regression for the determination of VA_BITS and section_size_bits

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Tao,
> Could you explain why removing it can fix the issue? I cannot get a
> clue by reading the message. Also please put the root cause into the
> patch log as well.
> 
The first change is a mistake fix which can be ignored, since the dump file contains a wrong vmcoreinfo "NUMBER(TCR_EL1_T1SZ)=25\n"
which lead a wrong vabits_actual(27) got by arm64_set_va_bits_by_tcr(). The same issue as Guanyou comments: https://lists.crash-utility.osci.io/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxxxxxx/thread/LUCYZ5JBGEH2R422JFPJC5IUPMDRGCRS/

> 
> Looks like you have fixed 2 regressions, and I don't see the 2
> regressions are related to each other(correct me if I'm wrong), If so
> I would prefer you to send 2 individual patches, each dealing with
> only one regression. So we can track the patch history easier.
> 
So there is only one regression left. I would follow your suggestion.

> 
> Do you determine the android-12-GKI by checking
> CONFIG_ANDROID_KABI_RESERVE? If so could you also write it into the
> code comments, for people who don't have an android kernel background.
> 
Checking CONFIG_ANDROID_KABI_RESERVE is a easiest way which introduced on android12-5.10 with the google change:
https://android.googlesource.com/kernel/common/+/3e9d5f02f0e124633edb3a638fe614466c8d0a51%5E%21/

There is another way to  determine the android-12-GKI by checking linux_banner:
crash> linux_banner
linux_banner = $1 = "Linux version 5.10.209-android12-9-g5cf27268b585-dirty ..."

Which way do you think is better?

> 
> I think machdep->section_size_bits should be initialized as 0 at the
> start of arm64_get_section_size_bits(), the exit block will assume its
> value to be 0 if error happens.
The initial value of machdep->section_size_bits is supposed to be 0 since the global variable machdep is initialized as 0.
struct machdep_table machdep_table = { 0 };
struct machdep_table *machdep = &machdep_table;

Thanks
--
Crash-utility mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxxxxxx
https://${domain_name}/admin/lists/devel.lists.crash-utility.osci.io/
Contribution Guidelines: https://github.com/crash-utility/crash/wiki




[Index of Archives]     [Fedora Development]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [KDE Users]     [Fedora Tools]

 

Powered by Linux