Note: Without this patch, one particular laptop is failing after 12 S4 iterations. With the patch we don't get the memcpy() oops anymore.
I've also soak tested this patch on another machine with 250 S4 cycles. I also threw in 250 S3 cycles to make sure we don't get any regressions. ** Description changed: Ubuntu-2.6.35-25, Maverick. After repeated S4 cycles, a memcpy() Oops occurs in suspend_nvs_save. This error occurs because the ioremap of the NVS region fails, returning NULL and the subsequent memcpy() tries to copy to a NULL pointer, hence the oops. - We running out of mappable regions because suspend_nvs_save() is called - twice during hibernation, and yet the unmapping via suspend_nvs_free() - is only called once, hence we have a leak. + We are running out of mappable regions because suspend_nvs_save() is + called twice during hibernation, and yet the unmapping via + suspend_nvs_free() is only called once, hence we have a leak. Looking at create_image() in kernel/power/hibernate.c, we can see that we call the __acpi_pm_prepare() via arch_prepare_suspend() and a little later we call platform_pre_snapshot() which calls acpi_hibernation_pre_snapshot() which does the duplication of the suspend_nvs_save(). We need to drop on these. This has been fixed in upstream commit: commit c5f7a1bb65bca03253c189e946b3ca79669f08af Author: Rafael J. Wysocki <r...@sisk.pl> Date: Fri Jul 2 00:14:09 2010 +0200 - ACPI / Sleep: Consolidate suspend and hibernation routines - - Some hibernation and suspend routines can be merged, which reduces - the complexity of code a bit. - - Signed-off-by: Rafael J. Wysocki <r...@sisk.pl> - Signed-off-by: Len Brown <len.br...@intel.com> + ACPI / Sleep: Consolidate suspend and hibernation routines + + Some hibernation and suspend routines can be merged, which reduces + the complexity of code a bit. + + Signed-off-by: Rafael J. Wysocki <r...@sisk.pl> + Signed-off-by: Len Brown <len.br...@intel.com> ..where the extraneous suspend_nvs_save() has been dropped from function __acpi_pm_prepare() ** Description changed: Ubuntu-2.6.35-25, Maverick. After repeated S4 cycles, a memcpy() Oops occurs in suspend_nvs_save. This error occurs because the ioremap of the NVS region fails, returning NULL and the subsequent memcpy() tries to copy to a NULL pointer, hence the oops. We are running out of mappable regions because suspend_nvs_save() is called twice during hibernation, and yet the unmapping via suspend_nvs_free() is only called once, hence we have a leak. Looking at create_image() in kernel/power/hibernate.c, we can see that we call the __acpi_pm_prepare() via arch_prepare_suspend() and a little later we call platform_pre_snapshot() which calls acpi_hibernation_pre_snapshot() which does the duplication of the - suspend_nvs_save(). We need to drop on these. + suspend_nvs_save(). We need to drop one of these. This has been fixed in upstream commit: commit c5f7a1bb65bca03253c189e946b3ca79669f08af Author: Rafael J. Wysocki <r...@sisk.pl> Date: Fri Jul 2 00:14:09 2010 +0200 ACPI / Sleep: Consolidate suspend and hibernation routines Some hibernation and suspend routines can be merged, which reduces the complexity of code a bit. Signed-off-by: Rafael J. Wysocki <r...@sisk.pl> Signed-off-by: Len Brown <len.br...@intel.com> ..where the extraneous suspend_nvs_save() has been dropped from function __acpi_pm_prepare() -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/703228 Title: memcpy Oops in suspend_nvs_save after multiple S4 cycles -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
