New memory tester application potentially to replace memtest86+: PCMemTest

[Neal Gompa]

Hey all,

I discovered today that there's a new replacement for memtest86+ that
appears to even have UEFI support called PCMemTest[0].

The main reason I call out to this is because we don't have a memory
tester offering in our UEFI boot variant for the Fedora live media,
and this is actively maintained (unlike memtest86+, which we currently
use...).

Mageia is shipping this starting with Mageia 8[1], and we should
consider shipping this with Fedora 34.

[0]: https://github.com/martinwhitaker/pcmemtest
[1]: https://wiki.mageia.org/en/Mageia_8_Release_Notes#PCMemTest

--
真実はいつも一つ！/ Always, there's only one truth!

[Roberto Ragusa]

On 2/7/21 10:07 AM, Neal Gompa wrote:
>  Hey all,
>  
>  I discovered today that there's a new replacement for memtest86+ that
>  appears to even have UEFI support called PCMemTest[0]. 
It can only be an alternative, not a replacement, since it is dropping features:

«In particular, no attempt is made to measure the cache and main memory speed,
or to identify and report the DRAM type.»

Regards.
-- 
    Roberto Ragusa    mail at robertoragusa.it

[drago01]

On Sunday, February 7, 2021, Roberto Ragusa <mail(a)robertoragusa.it&gt; wrote:

>  On 2/7/21 10:07 AM, Neal Gompa wrote:
>
> > Hey all,
> >
> > I discovered today that there's a new replacement for memtest86+ that
> > appears to even have UEFI support called PCMemTest[0].
> >
>
>  It can only be an alternative, not a replacement, since it is dropping
>  features:
>
>  «In particular, no attempt is made to measure the cache and main memory
>  speed,
>  or to identify and report the DRAM type.» 

 Which is nice to have but not really the point of a memory tester ...


>  Regards.
>  --
>     Roberto Ragusa    mail at robertoragusa.it
>  _______________________________________________
>  devel mailing list -- devel(a)lists.fedoraproject.org
>  To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct: https://docs.fedoraproject.org
>  /en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives: https://lists.fedoraproject.or
>  g/archives/list/devel(a)lists.fedoraproject.org
>

[Roberto Ragusa]

On 2/7/21 12:16 PM, drago01 wrote:
>  
>  On Sunday, February 7, 2021, Roberto Ragusa <mail(a)robertoragusa.it
> <mailto:[e-mail]>> wrote:
>  
>      It can only be an alternative, not a replacement, since it is dropping features:
>  
>      «In particular, no attempt is made to measure the cache and main memory speed,
>      or to identify and report the DRAM type.»
>  
>  
>    Which is nice to have but not really the point of a memory tester ... 
I'm pointing out that memtest86+ is not just a memory tester.
Is there any other tool covering that functionality?

-- 
    Roberto Ragusa    mail at robertoragusa.it

[Stephen John Smoogen]

On Sun, 7 Feb 2021 at 09:41, Roberto Ragusa <mail(a)robertoragusa.it&gt; wrote:

>  On 2/7/21 12:16 PM, drago01 wrote:
>  >
>  > On Sunday, February 7, 2021, Roberto Ragusa <mail(a)robertoragusa.it
>  <mailto:[e-mail]>> wrote:
>  >
>  >     It can only be an alternative, not a replacement, since it is
>  dropping features:
>  >
>  >     «In particular, no attempt is made to measure the cache and main
>  memory speed,
>  >     or to identify and report the DRAM type.»
>  >
>  >
>  >   Which is nice to have but not really the point of a memory tester ...
>
>  I'm pointing out that memtest86+ is not just a memory tester.
>  Is there any other tool covering that functionality?
>
>  
Probably not because for more and more modern hardware it gets harder and
harder to get that information. Heck even 'testing' memory is hard enough
these days because a lot of hardware is built around covering up any
problem with the hardware because it is expected to fail and the hardware
itself will see that, mark it bad and set something else up as good.

-- 
Stephen J Smoogen.

[Zbigniew Jędrzejewski-Szmek]

On Sun, Feb 07, 2021 at 11:03:38AM -0500, Stephen John Smoogen wrote:
>  On Sun, 7 Feb 2021 at 09:41, Roberto Ragusa
> <mail(a)robertoragusa.it&gt; wrote:
>  
>  > On 2/7/21 12:16 PM, drago01 wrote:
>  > >     It can only be an alternative, not a replacement, since it is
>  > dropping features:
>  > >
>  > >     «In particular, no attempt is made to measure the cache and main
>  > memory speed,
>  > >     or to identify and report the DRAM type.»
>  > >
>  > >
>  > >   Which is nice to have but not really the point of a memory tester ...
>  >
>  > I'm pointing out that memtest86+ is not just a memory tester.
>  > Is there any other tool covering that functionality? 
Memory type and speed information is available through dmi. In systemd we
recently merged a patch [1] to make this information available to unprivileged
users, so that gnome can display it in a pretty way:

$ udevadm info /sys/class/dmi/id/
P: /devices/virtual/dmi/id
...
E: MODALIAS=dmi:bvnLENOVO:bvrN1FET68W(1.42):bd03/13/2019...
...
E: MEMORY_ARRAY_LOCATION=System Board Or Motherboard
E: MEMORY_ARRAY_NUM_DEVICES=2
...
E: MEMORY_DEVICE_0_SIZE=4294967296
E: MEMORY_DEVICE_0_FORM_FACTOR=Chip
E: MEMORY_DEVICE_0_TYPE=LPDDR3
E: MEMORY_DEVICE_0_SPEED_MTS=1867
E: MEMORY_DEVICE_0_MANUFACTURER=Samsung
E: MEMORY_DEVICE_0_PART_NUMBER=K4E6E304EE-EGCF
...
E: MEMORY_DEVICE_1_SIZE=4294967296
E: MEMORY_DEVICE_1_FORM_FACTOR=SODIMM
E: MEMORY_DEVICE_1_LOCATOR=ChannelB
E: MEMORY_DEVICE_1_BANK_LOCATOR=BANK 2
E: MEMORY_DEVICE_1_TYPE=LPDDR3
...

[1] https://github.com/systemd/systemd/pull/17837

This is just repeated from what the firmware gives us, not measured,
but might be good enough.

Zbyszek

[John Reiser]

>      «In particular, no attempt is made to measure the cache and main
> memory speed,
>      or to identify and report the DRAM type.»
>  
>  
>    Which is nice to have but not really the point of a memory tester ... 
When the tester reports errors then it is handy to know as much as possible
about where the errors lie, including the manufacturer and part number
if an error can be isolated to an address range.  For a system which
added memory in mid-life using a different brand of RAM, it is helpful
to confirm the brand which is failing.  /usr/sbin/dmidecode reports
what it knows, but that requires cross-referencing etc.

[Chris Murphy]

On Sun, Feb 7, 2021 at 2:08 AM Neal Gompa <ngompa13(a)gmail.com&gt; wrote:
>  Hey all,
>
>  I discovered today that there's a new replacement for memtest86+ that
>  appears to even have UEFI support called PCMemTest[0].
>
>  The main reason I call out to this is because we don't have a memory
>  tester offering in our UEFI boot variant for the Fedora live media,
>  and this is actively maintained (unlike memtest86+, which we currently
>  use...).
>
>  Mageia is shipping this starting with Mageia 8[1], and we should
>  consider shipping this with Fedora 34. 

* A listed limitation: "When booted on a UEFI system, keyboard input
will only be seen if the CSM is enabled in the BIOS. Without this, the
test will run, but you will be unable to alter the configuration."

- How does a CSM provide keyboard input to an EFI application? Or does
this mean with CSM enabled, we'd use the BIOS version of the memory
tester; and with CSM disabled, we'd use the UEFI version of the memory
tester?

- As far as I'm aware, enabling CSM requires disabling UEFI Secure Boot.

* Any UEFI memory tester needs to be signed with Fedora's signing key,
same as grubx64.efi so that it works with UEFI Secure boot enabled.


-- 
Chris Murphy

[Vít Ondruch]

Being devils advocate, but should we have the memtest86 or similar by 
default? I have certainly not used this feature in my 10+ yeas with Fedora.


Vít



Dne 07. 02. 21 v 10:07 Neal Gompa napsal(a):
>  Hey all,
>
>  I discovered today that there's a new replacement for memtest86+ that
>  appears to even have UEFI support called PCMemTest[0].
>
>  The main reason I call out to this is because we don't have a memory
>  tester offering in our UEFI boot variant for the Fedora live media,
>  and this is actively maintained (unlike memtest86+, which we currently
>  use...).
>
>  Mageia is shipping this starting with Mageia 8[1], and we should
>  consider shipping this with Fedora 34.
>
>  [0]: https://github.com/martinwhitaker/pcmemtest
>  [1]: https://wiki.mageia.org/en/Mageia_8_Release_Notes#PCMemTest
>
>  --
>  真実はいつも一つ！/ Always, there's only one truth!
>  _______________________________________________
>  devel mailing list -- devel(a)lists.fedoraproject.org
>  To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives:
> https://lists.fedoraproject.org/archives/list/[e-mail]

[Chris Murphy]

On Mon, Feb 8, 2021 at 2:46 AM Vít Ondruch <vondruch(a)redhat.com&gt; wrote:
>  Being devils advocate, but should we have the memtest86 or similar by
>  default? I have certainly not used this feature in my 10+ yeas with Fedora.
>  
You mean get rid of it (from media and installations)? Because we
install it unconditionally already, even though it's a BIOS only
utility and there isn't a boot entry for it in the bootloader.

It's a bit obscure how to use it, given there's no menu entry for it.


-- 
Chris Murphy

[Brandon Nielsen]

On 2/8/21 2:44 PM, Chris Murphy wrote:
>  On Mon, Feb 8, 2021 at 2:46 AM Vít Ondruch
> <vondruch(a)redhat.com&gt; wrote:
> >
> > Being devils advocate, but should we have the memtest86 or similar by
> > default? I have certainly not used this feature in my 10+ yeas with Fedora.
> >
>  
>  You mean get rid of it (from media and installations)? Because we
>  install it unconditionally already, even though it's a BIOS only
>  utility and there isn't a boot entry for it in the bootloader.
>  
>  It's a bit obscure how to use it, given there's no menu entry for it.
>  
>   
There's also the fact it doesn't seem to work real well with modern 
hardware[0][1].

[0] - https://bugzilla.redhat.com/show_bug.cgi?id=1815742
[1] - https://bugzilla.redhat.com/show_bug.cgi?id=1869211

[Vít Ondruch]

Dne 08. 02. 21 v 21:44 Chris Murphy napsal(a):
>  On Mon, Feb 8, 2021 at 2:46 AM Vít Ondruch
> <vondruch(a)redhat.com&gt; wrote:
> > Being devils advocate, but should we have the memtest86 or similar by
> > default? I have certainly not used this feature in my 10+ yeas with Fedora.
> >
>  You mean get rid of it (from media and installations)? 

Yes, that is my proposal.


>    Because we
>  install it unconditionally already, even though it's a BIOS only
>  utility and there isn't a boot entry for it in the bootloader.
>
>  It's a bit obscure how to use it, given there's no menu entry for it.
>
>  
Ah, good to know that I have it on my system. I'm going to remove it 
right now.

I don't think this is utility, which is targets typical Fedora user. 
Here is PR proposing the removal:


https://pagure.io/fedora-kickstarts/pull-request/754



Vít

[Nikolay Nikolov]

On 2/9/21 10:53 AM, Vít Ondruch wrote:
>  Dne 08. 02. 21 v 21:44 Chris Murphy napsal(a):
> > On Mon, Feb 8, 2021 at 2:46 AM Vít Ondruch <vondruch(a)redhat.com&gt; wrote:
> >> Being devils advocate, but should we have the memtest86 or similar by
> >> default? I have certainly not used this feature in my 10+ yeas with 
> >> Fedora.
> >>
> > You mean get rid of it (from media and installations)?
>
>
>  Yes, that is my proposal.
>
>
> >   Because we
> > install it unconditionally already, even though it's a BIOS only
> > utility and there isn't a boot entry for it in the bootloader.
> >
> > It's a bit obscure how to use it, given there's no menu entry for it.
> >
> >
>
>  Ah, good to know that I have it on my system. I'm going to remove it 
>  right now. 
Enjoy the 357 KB space savings. :)
>  I don't think this is utility, which is targets typical Fedora user. 
>  Here is PR proposing the removal:
>
>
>  https://pagure.io/fedora-kickstarts/pull-request/754 
A memory testing utility is certainly something useful in ruling out 
hardware reasons for system crashes. Even Microsoft has included one in 
recent versions of Windows. Being BIOS only is a problem, yes, but that 
makes it even more useful to have on the installation media, because I 
can enable the legacy BIOS CSM and boot the installation DVD on my UEFI 
systems, while I can't really install it on my hard drive and enable it 
in the GRUB menu if I have an UEFI install. Now, I would need to use a 
different installation media to run this tool for the sake of reducing 
the >2GB installation image by several kilobytes. Doesn't sound like a 
great idea to me.

I agree it should not be installed by default on UEFI systems and I 
think it should probably have been enabled by default in the GRUB menu 
on BIOS systems. I also agree an alternative like PCMemTest is needed 
for UEFI systems, because not all systems have legacy BIOS CSMs.

Nikolay

[Roberto Ragusa]

On 2/8/21 10:46 AM, Vít Ondruch wrote:
>  Being devils advocate, but should we have the memtest86 or similar by
> default? I have certainly not used this feature in my 10+ yeas with Fedora. 
I've
not used GNOME in my 18 years with Fedora (plus 5 pre-Fedora).
Can we consider removing it?

Having memtest86 saved my life just a few months ago: random crashes, apparently caused by
pressing
in the middle of the keyboard.

It was able to demonstrate _live_ that the pressure was causing bit flips.
Press, errors, don't press, no more errors (including appearance/disappearance of
vertical bands
in the display, because of integrated chipset, VRAM=RAM).

Opened the laptop, reseated two DIMMs, TESTED AGAIN, TESTED AGAIN, shaken the laptop,
TESTED AGAIN, no problem anymore since then.

Regards.

-- 
    Roberto Ragusa    mail at robertoragusa.it

[Vít Ondruch]

Dne 09. 02. 21 v 10:08 Roberto Ragusa napsal(a):
>  On 2/8/21 10:46 AM, Vít Ondruch wrote:
> > Being devils advocate, but should we have the memtest86 or similar by 
> > default? I have certainly not used this feature in my 10+ yeas with 
> > Fedora.
>  I've not used GNOME in my 18 years with Fedora (plus 5 pre-Fedora).
>  Can we consider removing it? 

Have you installed Gnome by default? You probably went with KDE spin or 
some other spins so Gnome was never part of your system.

Better example would be if you argued that KDE should be installed for 
every Gnome user even though they don't use it.


Vít


(Apologies if you are using some other desktop or if you are not using 
desktop at all. I just picked up KDE as the second most popular desktop 
choice).


>  Having memtest86 saved my life just a few months ago: random crashes, 
>  apparently caused by pressing
>  in the middle of the keyboard.
>
>  It was able to demonstrate _live_ that the pressure was causing bit 
>  flips.
>  Press, errors, don't press, no more errors (including 
>  appearance/disappearance of vertical bands
>  in the display, because of integrated chipset, VRAM=RAM).
>
>  Opened the laptop, reseated two DIMMs, TESTED AGAIN, TESTED AGAIN, 
>  shaken the laptop,
>  TESTED AGAIN, no problem anymore since then.
>
>  Regards.
>