Hi Andy, On Tue, Oct 28, 2014 at 12:10 AM, fugang.duan@xxxxxxxxxxxxx <fugang.duan@xxxxxxxxxxxxx> wrote: > Hi, Fabio, > > After the patch, uart wake up feature may be break. (So I always don't do like this.) > > Since the current driver enable AWAKE before system suspend in default, after suspend, if user click console keys, it generates AWAKE interrupt, > in the irq handler, there have register access that cause system hang since uart module/ipg clocks are disabled, and system cannot resume back. I have tested suspend/resume on a mx6qsabresd with this patch applied: root@freescale /$ echo enabled > /sys/class/tty/ttymxc0/power/wakeup root@freescale /$ echo mem > /sys/power/state [ 61.322648] PM: Syncing filesystems ... done. [ 61.342263] Freezing user space processes ... (elapsed 0.001 seconds) done. [ 61.351159] Freezing remaining freezable tasks ... (elapsed 0.002 seconds) d. [ 61.360875] Suspending console(s) (use no_console_suspend to debug) (then sent a char in console) [ 61.416329] PM: suspend of devices complete after 45.778 msecs [ 61.416337] PM: suspend devices took 0.050 seconds [ 61.420745] PM: late suspend of devices complete after 4.393 msecs [ 61.425547] PM: noirq suspend of devices complete after 4.787 msecs [ 61.425555] Disabling non-boot CPUs ... [ 61.426971] CPU1: shutdown [ 61.429095] CPU2: shutdown [ 61.431190] CPU3: shutdown [ 61.431992] Enabling non-boot CPUs ... [ 61.432635] CPU1: Booted secondary processor [ 61.433006] CPU1 is up [ 61.433627] CPU2: Booted secondary processor [ 61.433998] CPU2 is up [ 61.434622] CPU3: Booted secondary processor [ 61.435021] CPU3 is up [ 61.437511] PM: noirq resume of devices complete after 2.472 msecs [ 61.441288] PM: early resume of devices complete after 2.909 msecs [ 61.602251] PM: resume of devices complete after 160.946 msecs [ 61.606860] PM: resume devices took 0.170 seconds [ 61.688676] Restarting tasks ... done. [ 61.795155] ata1: SATA link down (SStatus 0 SControl 300) and the system wakes up normally. This patch does not change anything regarding enabling/disabling of wakeup irqs. Instead of requesting the irqs everytime when the uart port will transmit (inside imx_startup() ) it allocates the irq only once in the probe function, so the suspend/resume behaviour is unchanged. -- To unsubscribe from this list: send the line "unsubscribe linux-serial" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
