Re: [linux-pm] [PATCH 1/5] ACPI / Sleep: Do not allocate memory for saving NVS in advance

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Friday, July 02, 2010, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rjw@sisk.pl>
> 
> We only can try to allocate memory for saving the NVS region if it
> is known that the target system sleep state is valid.  Otherwise
> we will leak memory if suspend_nvs_alloc() is successful and the
> target state is invalid, because suspend_nvs_free() will not be
> called in that case.
> 
> Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>

Well, the $subject patch doesn't cover a real problem where the suspending
of devices fails (or suspend is tested in the "devices" mode) and
acpi_pm_finish() is not called at all.  So, the appended patch should be used
instead.

Rafael

---
From: Rafael J. Wysocki <rjw@sisk.pl>
Subject: ACPI / Sleep: Free NVS copy if suspending of devices fails

If suspending of devices fails or system suspend is tested in the
"devices" mode, the memory allocated for storing a copy of the ACPI
NVS area will not be freed, because acpi_pm_finish() is not called
in that case.  Fix this by moving the suspend_nvs_free() call to
acpi_pm_end().

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
---
 drivers/acpi/sleep.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

Index: linux-2.6/drivers/acpi/sleep.c
===================================================================
--- linux-2.6.orig/drivers/acpi/sleep.c
+++ linux-2.6/drivers/acpi/sleep.c
@@ -145,7 +145,6 @@ static void acpi_pm_finish(void)
 {
 	u32 acpi_state = acpi_target_sleep_state;
 
-	suspend_nvs_free();
 	acpi_ec_unblock_transactions();
 
 	if (acpi_state == ACPI_STATE_S0)
@@ -167,6 +166,7 @@ static void acpi_pm_finish(void)
  */
 static void acpi_pm_end(void)
 {
+	suspend_nvs_free();
 	/*
 	 * This is necessary in case acpi_pm_finish() is not called during a
 	 * failing transition to a sleep state.