Re: disable a memory block from swapping to hard disk

"behzad" <bn at yahoo.com> wrote in message
news:u2BMKEC4IHA.4856@xxxxxxxxxxxxxxxxxxxxxxx
Thanks for the answer.
(When I disabled hyper threading in bios, I did not see BSOD any more, at
least in last few days).
That was the reason I am looking if it is possible to get a memory block
that always be in RAM.

Clearly this is a driver bug, so you should report it to the vendor along
with the dump file generated at the time of BSOD. You should probably set
the system failure dump file settings to "Kernel Memory Dump" (generated as
%SystemRoot%\MEMORY.DMP) rather than "Complete memory dump" since that'd be
quite a large file (several Gigs I would imagine), and probably unnecessary.
Most vendors will get back to you within one week depending on the severity
of the bug. NVIDIA got back to me within 3 or 4 days following a similar bug
report (BSOD in xxx.dll where xxx.dll was the video card driver - i forget
the actual name now). I'm not sure about emailing an attachement since you
might get filtered away as spam. Perhaps save the file off and simply offer
to send it to them should they require it.

As a side note, clearly this is a synchronization issue and probably due to
some section of the driver code assuming the impossibility of a context
switch "mid-instruction" (ie. hyperthreading). This was a reasonable
assumption for many years, so my guess is the bug would likely be found in
some ancient (ie. > 2 years old) section of their code, perhaps making it
fairly easy for them to locate the problem with the information you provide.

!IMPORTANT!
Also, and I'm sure helped a lot in my case, build them an executable which
reproduces the bug with reasonable regularity (say within an hour or so).
That way they can debug the problem in real-time rather than trying to
decifer the dump file.

- Alan Carre


.

Relevant Pages

  • Re: [PATCH 1/2] LogFS proper
    ... Do enums have a significant ... Reason enough to use it for the simple cases. ... why are function and line included in BUG() then? ... More than in other filesystems simply because LogFS is younger. ...
    (Linux-Kernel)
  • Re: Checking for A Blank String
    ... Yes and in this case you pointed out a bug. ... as I do with other programming languages. ... | screwdriver to screw in a nail although there is a hammer in the toolbox ... function, regardless of their actual reason, why would I want to prefer it ...
    (microsoft.public.dotnet.languages.vb)
  • Re: 8.0 rc.d scripts less verbose?
    ... bug, but it's a bug that we've come to depend on. ... This late in the release cycle I'm comfortable adding a knob, ... pre-command routine or the command itself failed. ... whatever reason), it is nice to have an approximate idea which one ...
    (freebsd-stable)
  • Re: Security researchers organization
    ... > The Sardonix.org security auditing web site was designed to ... Sardonix provides: ... prevent last year's Chunked Encoding bug? ... -> this provides a reason for individual team members to share their ...
    (NT-Bugtraq)
  • Re: Vista SP1 does not stop INTERNAL_POWER_ERROR Blue Screen
    ... I have installed Ubuntu on dozens and dozens of laptops and desktops and didn't encounter one single bug. ... You can go back to staring at your BSOD now and think you're doing some in depth computing. ...
    (microsoft.public.windows.vista.general)
Loading