Core 2 Errata -- problematic or overblown?

By: Rob Thorpe (rthorpe.delete@this.realworldtech.com), June 29, 2007 6:15 am
Room: Moderated Discussions
Linus Torvalds (torvalds@osdl.org) on 6/28/07 wrote:
---------------------------
>Rob Thorpe (rthorpe@realworldtech.com) on 6/28/07 wrote:
>>
>>Well, most other CPUs are embedded ones.
>
>Well, I'd call those "commodity", wouldn't you? Even more
>so than x86.

Yes, that was my point. The CPUs that are most common generally are commodity microcontrollers, which have fewer errata than x86s, or at least about the same.

>That said, a lot of them are quite buggy. The way the
>embedded people define it, though, any bugs becomes just
>"specifications".

Yes. In a smart-card chip one of my friends encountered there was supposed to be a single-wire serial port with a FIFO. As it turned out there was only the single-wire part on the deal, a pad on the chip that could be read from and turned on and off. Anyone wanting the FIFO had to do it in software.

He said "I'm not going to tell the manufacturer because they'll just document it as an errata". That was quite a low volume thing though.

>For example, the ARM cache control has always been pretty
>damn buggy. What is the solution? Document it as a bug,
>and tell people not to use it.

In the particular case of cache behaviour are you sure you're not mistaking eccentricity for bugs. ARMs have strange caches.

>Or grep for "errata" in arch/arm in the Linux kernel. Trust
>me, they exist.

Oh, I don't doubt that.

>But yeah, you can make a microcontroller CPU that is based
>on some decade-old core that is fairly bug-free. Not that
>it probably really is, but over the years the bugs have all
>become "behavior" rather than "bugs".
>
>>Whether other higher-end CPUs have more errata than x86s
>>I don't know.
>
>They tend to fix the bugs that are user-visible, and then
>not fix the bugs that can be worked around on an OS
>level.
>
>Also, boutique vendors tend to not talk about them,
>because it's all internal to their own stuff. Of course, if
>they don't catch it in time, they'll have to release OS
>upgrades, but if they find an errata early, they can just
>work around it and need never tell anybody, exactly like
>the random embedded ones.
>
>It's really simple: when you count your CPU's in thousands
>rather than millions, you generally don't want to do a whole
>new mask set that costs you months and a few megabucks. It's
>much cheaper to just ship the buggy crud.
>
>Yeah, x86 errata get more attention. But those things are
>pretty damn well tested. Better than most. And since the OS
>is outside the control of the vendors, they get fixed too.

That sounds like my experience in communications electronics. The protocols that work well are the ones that are either simple or widely used.

So GSM, bluetooth and WCDMA chipsets/equipment for mobile phones and WLAN cards for PCs work well because they are sold in reasonable volume. These things work well despite being complex because they are well used. Things like DECT and obscure single-purpose digital comms systems also work well, generally though because they're simpler so missing bugs is less forgivable, they're easier to test.

Systems that are complex and not well used are worse. Things like WLAN chips for things other than PCs, GSM equipment for data transfer purposes and Bluetooth systems for particular applications.

Rob
Speaking only for myself.
< Previous Post in ThreadNext Post in Thread >
TopicPosted ByDate
Core 2 Errata -- problematic or overblown?Matt Sayler2007/06/27 11:39 AM
  Core 2 Errata -- problematic or overblown?anonymous2007/06/27 02:08 PM
  Core 2 Errata -- problematic or overblown?David Kanter2007/06/27 02:34 PM
    Core 2 Errata -- problematic or overblown?Joe Chang2007/06/27 03:01 PM
    Core 2 Errata -- problematic or overblown?A.T. Ho2007/06/27 04:33 PM
      Core 2 Errata -- problematic or overblown?David Kanter2007/06/27 06:31 PM
      Core 2 Errata -- problematic or overblown?Andi Kleen2007/06/28 12:35 AM
        Core 2 Errata -- problematic or overblown?anonymous2007/06/28 04:16 AM
    Core 2 Errata -- problematic or overblown?Jack A.2007/06/29 07:43 PM
      Core 2 Errata -- problematic or overblown?anon2007/07/01 10:13 AM
        AMD DocumentationDavid Kanter2007/07/01 11:03 AM
  Core 2 Errata -- problematic or overblown?Andi Kleen2007/06/27 04:02 PM
    Core 2 Errata -- problematic or overblown?Paul2007/06/27 05:06 PM
    Core 2 Errata -- problematic or overblown?Groo2007/06/28 01:10 AM
  Core 2 Errata -- problematic or overblown?Linus Torvalds2007/06/27 05:10 PM
    Core 2 Errata -- problematic or overblown?Rob Thorpe2007/06/28 01:11 AM
      Core 2 Errata -- problematic or overblown?JS2007/06/28 04:32 AM
        Core 2 Errata -- problematic or overblown?Rob Thorpe2007/06/28 05:15 AM
          Core 2 Errata -- problematic or overblown?JS2007/06/28 09:55 PM
            Core 2 Errata -- problematic or overblown?Rob Thorpe2007/06/29 12:44 AM
      Core 2 Errata -- problematic or overblown?Linus Torvalds2007/06/28 08:30 AM
        Core 2 Errata -- problematic or overblown?philt2007/06/28 04:28 PM
          Core 2 Errata -- problematic or overblown?Linus Torvalds2007/06/28 07:31 PM
            Core 2 Errata -- problematic or overblown?philt2007/06/29 01:58 AM
              Core 2 Errata -- problematic or overblown?IlleglWpns2007/06/29 02:23 AM
                Core 2 Errata -- problematic or overblown?philt2007/06/29 08:16 PM
              Core 2 Errata -- problematic or overblown?Andi Kleen2007/06/29 04:36 AM
                Core 2 Errata -- problematic or overblown?philt2007/06/29 08:09 PM
                  Core 2 Errata -- problematic or overblown?Brendan2007/06/29 11:04 PM
                    Core 2 Errata -- problematic or overblown?philt2007/06/30 01:52 AM
                      Core 2 Errata -- problematic or overblown?Brendan2007/06/30 09:25 AM
                  Core 2 Errata -- problematic or overblown?Jukka Larja2007/06/29 11:22 PM
                  Core 2 Errata -- problematic or overblown?dan2007/07/05 12:36 PM
              Core 2 Errata -- problematic or overblown?Dean Kent2007/06/29 07:31 AM
                Core 2 Errata -- problematic or overblown?philt2007/06/29 07:57 PM
                  Core 2 Errata -- problematic or overblown?Dean Kent2007/06/29 08:50 PM
                    Core 2 Errata -- problematic or overblown?philt2007/06/30 01:57 AM
                      Core 2 Errata -- problematic or overblown?Dean Kent2007/06/30 05:34 AM
                        Core 2 Errata -- problematic or overblown?philt2007/07/02 03:33 AM
                          Core 2 Errata -- problematic or overblown?Dean Kent2007/07/02 05:36 AM
                            Core 2 Errata -- problematic or overblown?philt2007/07/02 03:02 PM
                              Core 2 Errata -- problematic or overblown?JS2007/07/02 10:02 PM
                      Core 2 Errata -- problematic or overblown?David Kanter2007/07/02 03:20 PM
            Core 2 Errata -- problematic or overblown?Max2007/06/29 11:23 AM
        Core 2 Errata -- problematic or overblown?Rob Thorpe2007/06/29 06:15 AM
          Core 2 Errata -- problematic or overblown?Michael S2007/06/29 06:57 AM
    Core 2 Errata -- problematic or overblown?jozzete2007/06/29 12:54 AM
  Core 2 Errata -- problematic or overblown?Brendan2007/06/28 07:15 PM
    Core 2 Errata -- problematic or overblown?france2007/06/28 10:38 PM
  Core 2 Errata -- problematic or overblown?rwessel2007/06/29 03:20 AM
    Core 2 Errata -- problematic or overblown?jschlossel2007/06/29 10:16 AM
      Core 2 Errata -- problematic or overblown?Dean Kent2007/06/29 10:25 AM
        Core 2 Errata -- problematic or overblown?Howard Chu2007/06/30 05:09 AM
          Core 2 Errata -- problematic or overblown?JS2007/07/02 04:58 AM
      baby and bath waterrels2007/06/30 07:00 AM
      Planes, trains and cars...Anna N.2007/07/01 06:28 PM
        Planes, trains and cars...David W. Hess2007/07/01 07:40 PM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell green?