[PATCH v3] vmcore: Add a kernel parameter vmcore_device_dump

Dave Young dyoung at redhat.com
Mon May 27 22:18:02 PDT 2019


On 05/27/19 at 01:05pm, Kairui Song wrote:
> On Mon, May 27, 2019 at 2:45 AM Bhupesh Sharma <bhsharma at redhat.com> wrote:
> >
> > On Fri, May 24, 2019 at 6:25 PM Dave Young <dyoung at redhat.com> wrote:
> > >
> > > On 05/24/19 at 02:29pm, Kairui Song wrote:
> > > > Since commit 2724273e8fd0 ("vmcore: add API to collect hardware dump in
> > > > second kernel"), drivers is allowed to add device related dump data to
> > > > vmcore as they want by using the device dump API. This have a potential
> > > > issue, the data is stored in memory, drivers may append too much data
> > > > and use too much memory. The vmcore is typically used in a kdump kernel
> > > > which runs in a pre-reserved small chunk of memory. So as a result it
> > > > will make kdump unusable at all due to OOM issues.
> > > >
> > > > So introduce new vmcore_device_dump= kernel parameter, and disable
> > > > device dump by default. User can enable it only if device dump data is
> > > > required for debugging, and have the chance to increase the kdump
> > > > reserved memory accordingly before device dump fails kdump.
> > > >
> > > > Signed-off-by: Kairui Song <kasong at redhat.com>
> > > >
> > > > ---
> > > >
> > > >  Update from V2:
> > > >   - Improve related docs
> > > >
> > > >  Update from V1:
> > > >   - Use bool parameter to turn it on/off instead of letting user give
> > > >     the size limit. Size of device dump is hard to determine.
> > > >
> > > >  Documentation/admin-guide/kernel-parameters.txt | 14 ++++++++++++++
> > > >  fs/proc/Kconfig                                 |  6 ++++--
> > > >  fs/proc/vmcore.c                                | 13 +++++++++++++
> > > >  3 files changed, 31 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
> > > > index 138f6664b2e2..3706ad9e1d97 100644
> > > > --- a/Documentation/admin-guide/kernel-parameters.txt
> > > > +++ b/Documentation/admin-guide/kernel-parameters.txt
> > > > @@ -5078,6 +5078,20 @@
> > > >                       decrease the size and leave more room for directly
> > > >                       mapped kernel RAM.
> > > >
> > > > +     vmcore_device_dump=     [KNL,KDUMP]
> > > > +                     Format: {"off" | "on"}
> > > > +                     Depends on CONFIG_PROC_VMCORE_DEVICE_DUMP.
> > > > +                     This parameter allows enable or disable device dump
> > > > +                     for vmcore on kernel start-up.
> > > > +                     Device dump allows drivers to append dump data to
> > > > +                     vmcore so you can collect driver specified debug info.
> > > > +                     Note that the drivers could append the data without
> > > > +                     any limit, and the data is stored in memory, this may
> > > > +                     bring a significant memory stress. If you want to turn
> > > > +                     on this option, make sure you have reserved enough memory
> > > > +                     with crashkernel= parameter.
> > > > +                     default: off
> > > > +
> > > >       vmcp_cma=nn[MG] [KNL,S390]
> > > >                       Sets the memory size reserved for contiguous memory
> > > >                       allocations for the vmcp device driver.
> > > > diff --git a/fs/proc/Kconfig b/fs/proc/Kconfig
> > > > index 817c02b13b1d..1a7a38976bb0 100644
> > > > --- a/fs/proc/Kconfig
> > > > +++ b/fs/proc/Kconfig
> > > > @@ -56,8 +56,10 @@ config PROC_VMCORE_DEVICE_DUMP
> > > >         recovery kernel's initramfs to collect its underlying device
> > > >         snapshot.
> > > >
> > > > -       If you say Y here, the collected device dumps will be added
> > > > -       as ELF notes to /proc/vmcore.
> > > > +       If you say Y here, a new kernel parameter 'vmcore_device_dump'
> > > > +       will be available. You can then enable device dump by passing
> > >
> > > "a new kernel parameter 'vmcore_device_dump' will be available" is not
> > > necessary, "new" is a not a clear word.  I suggest to remove this
> > > sentence.
> > >
> > > s/You can then/You can
> >
> > I agree with Dave. We are just trying to say here that even if
> > CONFIG_PROC_VMCORE_DEVICE_DUMP is set to Y, one can still disable the
> > device dump feature by passing parameter 'vmcore_device_dump=off' to
> > the kernel.
> >
> > May be you can use the wording I mentioned in the v2 patch review,
> > which tried to convey a similar meaning.
> >
> > With the change addressed:
> > Reviewed-by: Bhupesh Sharma <bhsharma at redhat.com>
> >
> > Thanks,
> > Bhupesh
> >
> OK, How about:
> 
>   If you say Y here, device dump is still disabled by default.
>   You can enable device dump by passing 'vmcore_device_dump=on'
>   to kernel, the collected device dumps will be added as ELF
>   notes to /proc/vmcore.
> 
> If you think this is good I'll send V4 including the changes.

Kairui, looks good.

rethink about it, to be even simple, just replace the __setup with core_param like
crash_kexec_post_notifiers, no need on/off. 

> 
> -- 
> Best Regards,
> Kairui Song

Thanks
Dave



More information about the kexec mailing list