Groups | Search | Server Info | Login | Register


Groups > alt.os.linux.mint > #47139

Re: Sick HTPC

From Paul <nospam@needed.invalid>
Newsgroups alt.os.linux.mint
Subject Re: Sick HTPC
Date 2026-04-19 17:03 -0400
Organization A noiseless patient Spider
Message-ID <10s3fva$92el$1@dont-email.me> (permalink)
References (1 earlier) <10rnq54$r1cv$1@dont-email.me> <20260416201429.90b2771896e16cf82cabf286@gmail.com> <10rrj3h$1vvpn$1@dont-email.me> <20260418220233.ac08e2cf5a106306e4459f7a@gmail.com> <20260419214446.800c80cd1824f87887a56f4f@gmail.com>

Show all headers | View raw


On Sun, 4/19/2026 4:44 PM, pinnerite wrote:

> 
> It is isn't. Despite the new kernel, the mouse cursor suddenly went
> slow and the MythTV couldfn't get live TV. A reboot fixed it. For a
> while. This has been going on for a fe years. I could nevedr pin down
> the cause. But is now pretty old so a rebuild is imminent.

Did you take a look at "top" ?

There was one Promise brand chip, years ago, that could jam the
interrupt line and any running OS would then rely on its interrupt
limiter behavior, to have an CPU left to respond to any other activity
on the machine. Not many hardware devices have had defects like that.
Not all the Promise chips for that part number had the defect,
so there might have been a timing element to the behavior.

I don't know what tool would give a picture of interrupt activity.
On Windows, using Process Explorer, there is a line labeled
"Hardware Interrupts and DPCs", where a DPC is a Delay Procedure Call
which is the Ring 3 portion of interrupt handling, while Ring 0
is the Hardware Interrupt component. Minimal time is spent in the
Hardware Interrupt, whereas the DPC execution time might be 10x longer
as it is doing the actual service response (adding elements to a ring buffer
or similar). You would have to look for the similar elements on
Linux and what tool can read that portion out.

There aren't too many cases in hardware, where the actual clock rate
of the CPU was defective. There might have been a couple Dell laptops
where that was going on. I would think something interrupt related
(like an issue with the video card interrupts) is a more likely
kind of scenario. Oh, yeah, the Dell one was called "ThrottleGate"
because the clock would slow down, but it would not speed up
later. The document prepared by the person suffering from the
problem, was removed from the Internet, which is why you might
have trouble finding the details of that.

   Paul

Back to alt.os.linux.mint | Previous | NextPrevious in thread | Next in thread | Find similar


Thread

Sick HTPC pinnerite <pinnerite@gmail.com> - 2026-04-14 21:08 +0100
  Re: Sick HTPC Paul <nospam@needed.invalid> - 2026-04-15 06:43 -0400
    Re: Sick HTPC pinnerite <pinnerite@gmail.com> - 2026-04-16 20:14 +0100
      Re: Sick HTPC Paul <nospam@needed.invalid> - 2026-04-16 17:08 -0400
        Re: Sick HTPC pinnerite <pinnerite@gmail.com> - 2026-04-18 22:02 +0100
          Re: Sick HTPC pinnerite <pinnerite@gmail.com> - 2026-04-19 21:44 +0100
            Re: Sick HTPC Paul <nospam@needed.invalid> - 2026-04-19 17:03 -0400
      Re: Sick HTPC Mike Easter <MikeE@ster.invalid> - 2026-04-16 22:58 -0700

csiph-web