Hi!
There's extremely stupid bug in sleep.c -- it will only alow user to
enter *unsupported* states. What's even worse that I remember fixing
that once before, and *it was merged to mainline*.
Please, really merge it to all copies so it is not reintroduced again.
[Oh, at it also makes S4 transition fail when SWSUSP support is not
there.]
Pavel
--- clean/drivers/acpi/sleep.c 2002-09-22 23:46:56.000000000 +0200
+++ linux-swsusp/drivers/acpi/sleep.c 2002-10-01 23:38:48.000000000 +0200
@@ -329,8 +329,8 @@
state_string[count] = '\0';
state = simple_strtoul(state_string, NULL, 0);
-
- if (sleep_states[state])
+
+ if (!sleep_states[state])
return_VALUE(-ENODEV);
#ifdef CONFIG_SOFTWARE_SUSPEND
@@ -338,7 +338,10 @@
software_suspend();
return_VALUE(count);
}
+#else
+ return_VALUE(-ENODEV);
#endif
+
status = acpi_suspend(state);
if (ACPI_FAILURE(status))
--
Worst form of spam? Adding advertisment signatures ala sourceforge.net.
What goes next? Inserting advertisment *into* email?
> From: Pavel Machek [mailto:[email protected]]
> There's extremely stupid bug in sleep.c -- it will only alow user to
> enter *unsupported* states. What's even worse that I remember fixing
> that once before, and *it was merged to mainline*.
Cough cough Pat... ;-)
It's (re-)fixed in my bk tree already. Didn't have this bit at the end so
I'll apply that.
Regards -- Andy
> +#else
> + return_VALUE(-ENODEV);
> #endif