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

Bug#1000208: ITP: pcmemtest -- stand-alone memory tester



Hi Adam,

thanks for all your comments.

Am Montag, dem 22.11.2021 um 09:27 +0100 schrieb Adam Borowski:
> On Sun, Nov 21, 2021 at 04:46:50PM +0100, Felix Zielcke wrote:
> > Am Sonntag, dem 21.11.2021 um 10:15 -0500 schrieb Antoine Beaupré:
> > > i would suggest not blocking on the grub deployment. you can
> > > probably
> > > just deploy whatever architecture the package was built from, in
> > > any
> > > case.
> 
> > I just added now the GRUB 2 integration and a README.Debian.
> 
> The integration isn't triggered on the package's install, but it is
> picked
> up the next time something else bumps grub.

I forgot that grub2 isn't using triggers.
Added now a postinst based on the one from memtest86+. Except that I
removed the lilo parts.

> 
> 
> > Note that I mention in README.Debian one nasty bug/issue:
> > 
> > In EFI mode the keyboard only works if you have the CSM aka legacy
> > boot
> > also enabled:
> > 
> > https://github.com/martinwhitaker/pcmemtest/issues/2
> 
> This is a nasty one.  It seems most new machines lack CSM; no one wants
> to
> support and validate 16-bit stuff -- it's effectively expending
> resources
> to have two BIOSes instead of one, and the 8086 one has no practical
> usage.
> 

I opened now upstream a new issue about this:

https://github.com/martinwhitaker/pcmemtest/issues/13

But in the closed one he said, it takes a while to write a usb keyboard
driver for EFI.

> > Upstream doestn't say anything there that this will change. Issue was
> > closed with the hint it's documented. And it will just run with
> > default
> > settings.
> 
> It does, but using just a single thread.  There's not exactly many x86
> machines with only a single hardware thread that are still in use.
> 
> What about defaulting to SMP when there's no user input?  The UP mode
> has little purpose for existing -- if concurrent accesses to memory
> break, they'll also break when running the actual productive task the
> machine is supposed to do.
> 

I also made an issue for this:

https://github.com/martinwhitaker/pcmemtest/issues/14

I don't know the reasons why it's not enabled by default. So I woudn't
change that for the first upload to Debian on my own.

> Meow!

Cheers!
Felix


Reply to: