diff options
author | Hiago De Franco <hiago.franco@toradex.com> | 2025-08-07 17:15:06 -0300 |
---|---|---|
committer | Hiago De Franco <hiago.franco@toradex.com> | 2025-08-07 17:24:55 -0300 |
commit | 30f40c9d81341e60e8f17141f18a15e2869db04f (patch) | |
tree | 256eaa284f4ee7e48a35b42c9ad1635d2b463c78 /tools/lib/string.c | |
parent | acec5e45c2ad3f62d1e69de4a91c03fc917e12e5 (diff) |
Revert "remoteproc: k3-m4: Move suspend to suspend_late"
This reverts commit 49b236a4e51e804a1363289c6bdb6d7fb1037e78.
After this change, system suspend fails with the following error:
k3-m4-rproc 5000000.m4fss: PM mbox_send_message failed: -22
k3-m4-rproc 5000000.m4fss: PM: dpm_run_callback(): k3_m4_suspend_late+0x0/0x1a8 returns -22
k3-m4-rproc 5000000.m4fss: PM: failed to suspend late: error -22
The suspend callback was moved from the prepare phase to the suspend
late phase to make sure that the TI SCI prepare sleep command runs
before the remote processor is suspended. However, at that point in the
suspend sequence, the mailbox channel used to send the suspend message
to the remote processor is already unavailable (shut down), causing
mbox_send_message() to fail with -EINVAL.
To debug this a printk() was added in mbox_send_message() to verify the
state of the mailbox channel:
if (!chan || !chan->cl) {
printk("hfranco: channel is null! returning");
return -EINVAL;
}
This confirmed that the channel was NULL during the suspend_late phase:
hfranco: channel is null! returning
k3-m4-rproc 5000000.m4fss: PM mbox_send_message failed: -22
This commit is broken, therefore, to restore system suspend
functionality, revert this patch.
This patch is downstream only, this was never sent mainline and it is
not available on TI 6.12 downstream kernel.
Upstream-Status: Inappropriate
Signed-off-by: Hiago De Franco <hiago.franco@toradex.com>
Diffstat (limited to 'tools/lib/string.c')
0 files changed, 0 insertions, 0 deletions