Sleep/resume problems, etc.
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 7 lines · [email protected] said:
The other more serious problem (which seems to only be related to
resuming from deep sleep) is that the entire OS seems to become unresponsive (processes will not run) after a resume and I'll see these messages on the console:waiting for chip to be ready timed out in read. WSM status =
e0021985Error reading node from0x0030fa84: -5
The power cycle has reset the flash chip into read mode. The driver still thinks it's in status mode.
The chip driver's resume function needs to make sure the chip is in a known state, rather than assuming it'll still be in the same state it was before the sleep.
I'm sure I'd done this before.
cfi_cmdset_0001.c +3 -1
@@ -1582,7 +1582,9 @@spin_lock_bh (chip->mutex);if (chip->state ==FL_PM_SUSPENDED ) { - chip->state = chip->oldstate; +/* We need to force it back to a known state. */ +cfi_write (map,CMD (0xff ), cmd_addr); + chip->state =FL_READY ;wake_up (&chip->wq); }
--
dwmw2
Quoted 13 lines
The power cycle has reset the flash chip into read mode. The driver still thinks it's in status mode.
The chip driver's resume function needs to make sure the chip is in a known state, rather than assuming it'll still be in the same state it was before the sleep.
I'm sure I'd done this before.
diff -u -r1.72
cfi_cmdset_0001.c
cfi_cmdset_0001.c line 1586
Quoted 6 lines
cfi_cmdset_0001.c +2
-1@@ -1582,7 +1582,9 @@spin_lock_bh (chip->mutex);if (chip->state ==FL_PM_SUSPENDED ) { - chip->state = chip->oldstate; +/* We need to force it back to a known state. */ +cfi_write (map,CMD (0xff ), cmd_addr);
^^^^^^^^
What should cmd_addr be here? It's not defined.
--
Regards, Devin.