Linux on ARM based TI OMAP SoCs
 help / color / mirror / Atom feed
From: Andreas Kemnade <andreas@kemnade.info>
To: "Kevin Hilman (TI)" <khilman@baylibre.com>
Cc: linux-omap@vger.kernel.org, Aaro Koskinen <aaro.koskinen@iki.fi>,
	Roger Quadros <rogerq@kernel.org>,
	linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH] ARM: OMAP2+: pm34xx: add NEON wakeup dependency before clockdomain idle
Date: Thu, 19 Nov 2026 20:04:50 +0100	[thread overview]
Message-ID: <20261119200450.038bcc12@kemnade.info> (raw)
In-Reply-To: <20261113-wip-omap3-vfp-neon-mismatch-v1-1-3f6f67b25251@baylibre.com>

On Fri, 13 Nov 2026 10:11:05 -0800
"Kevin Hilman (TI)" <khilman@baylibre.com> wrote:

> On OMAP3630 (BeagleBoard-xM), about one boot in three dies early when
> a context switch touches the VFP, which traps as undefined.
> 
> The VFP is unreachable because NEON is powered off while the MPU runs.
> omap3_pm_init() sets NEON's next state to OFF/RET and lets all
> clockdomains idle before it adds the NEON -> MPU wakeup dependency, so
> NEON can power down immediately and stay off until the MPU next idles.
> This was visible in pm_debug, which showed neon_pwrdm OFF:1 while
> mpu_pwrdm OFF:0.
> 

I haved played around during debugging with this:

@ -687,7 +695,17 @@ static int vfp_kmode_exception(struct pt_regs *regs, unsigned int instr)
         * rebuilding and running with CONFIG_DEBUG_ATOMIC_SLEEP enabled should
         * be helpful in localizing the problem.
         */
+       u32 access = get_copro_access();
+
+
+       WARN_ON(1);
+       if ((CPACC_FULL(10) | CPACC_FULL(11)) != (access & (CPACC_FULL(10) | CPACC_FULL(11)))) {
+               pr_err("no access to copro\n");
+               WARN_ON(1);
+               vfp_enable(NULL);
+       }
        if (fmrx(FPEXC) & FPEXC_EN)

That error comes. but the endless exception  recursion of course still comes.
I am wondering if I should submit something like that with a BUG() instead
to have some more sane error message.
So the visibile symptom is that these CPI15 bits cannot be set.

Regards,
Andreas

      parent reply	other threads:[~2026-11-19 19:04 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-13 18:11 [PATCH] ARM: OMAP2+: pm34xx: add NEON wakeup dependency before clockdomain idle Kevin Hilman (TI)
2026-11-13 22:06 ` Andreas Kemnade
2026-11-16 16:16 ` Kevin Hilman (TI)
2026-11-19 19:04 ` Andreas Kemnade [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20261119200450.038bcc12@kemnade.info \
    --to=andreas@kemnade.info \
    --cc=aaro.koskinen@iki.fi \
    --cc=khilman@baylibre.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-omap@vger.kernel.org \
    --cc=rogerq@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox