Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed from November 25 through December 2. Thank you for your patience.

  • State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 102 subscribers
  • Views 2595 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

processors affected by ARM errata 430973

Peter Meerwald
Prodigy 70 points
Other Parts Discussed in Thread: DM3730, DM3725

Linux kernel has a workaround for ARM errata 430973 (Stale prediction on replaced interworking branch)

on a recent DM3730, the workaround must be applied to resolve issues; I want to check against the errata list of the processor -- can you make it available?

which processors are affected by ARM errata 430973?

Processor       : ARMv7 Processor rev 2 (v7l)
BogoMIPS        : 313.29
Features        : swp half thumb fastmult vfp edsp thumbee neon vfpv3 tls
CPU implementer : 0x41
CPU architecture: 7
CPU variant     : 0x3
CPU part        : 0xc08
CPU revision    : 2

  • 0 in reply to Maciej Kucia1
    TI__Guru* 75890 points

    Peter, you will have to get the full errata directly from ARM.  I can tell you that the DM3730 uses the r3p2 version of the ARM core, in which errata 430973 has been fixed, so you should not need to implement the workaround.

    Regards,

    James

  • 0 in reply to JJD
    Prodigy 70 points

    Hello,

    I have been told by ARM support that I should get the document from TI;

    TI so far pointed me to a document called "ARM Core Cortex-A8 (AT400/AT401) errata" that is available via local FAE

    the document "DM3730, DM3725 Digital Media Processors
    Silicon Errata (Revs 1.2, 1.1 & 1.0) (Rev. E)", advisory 1.10 (page 39), points to a separate document from TI

    I am aware that the DM3730 uses r3p2 of the ARM core and reportedly the errata should have been addressed (but I want to check the errata document)

    I am seeing very reproducible segfaults/illegal instruction on a DM3730 when the code for the errata workaround is NOT applied; i.e. I strongly believe that the DM3730 has an issue related to errata 430973

    the code in question is public, see http://ssvb.github.com/2012/04/10/cpuburn-arm-cortex-a9.html and try ssvb-cpuburn-a8.S concurrently with other processes; the code is stable with errata 430973 applied in the kernel, but crashes frequently without the errate code applied

    thanks, p.

  • 0 in reply to Peter Meerwald
    TI__Guru* 75890 points

    Peter, i think you have the right doc from ARM concerning the errata (AT400/AT401).  I'm sorry i'm not sure i can help much here as I am not familiar with the workaround that was implemented.  The ARM errata clearly states that this bug was fixed back in the r2p1 core. 

    This errata is present in our OMAP35x processor family, which uses an older ARM core.

    Regards,

    James