Hi Marc, > I have tested your version and it works fine. I saw that you enforced > the NO_WAY_OUT bit and implemented the KEEPALIVEPING and MAGICCLOSE > features properly. Many thanks. No problem. > Regarding the timer, the driver can not read the timeout value set in > the BIOS, so there is still a risk of reboot if the BIOS value is lower > than driver one. I assume that other drivers may be in the same case, > and that is required for user compatibility anyway. The driver assumes 1 second as heartbeat of the watchdog device. It thus takes half of this time do actually 'tickle' the watchdog. (or 500ms). We should set this 1 second value to the smallest time that is possible (either the smallest value that can be set in the BIOS or the smallest value that is listed in the datasheets). > Here is the comment block that I propose to insert regarding register > programming issues: > > /* > * Caveat: the only known way to activate the watchdog timer or to set > * its period is through BIOS configuration, when the hardware supports it. > * It seems that the watchdog control registers are locked and disabled if > * the watchdog is not exposed or enabled by BIOS, and that writing them is > * useless. No public documentation is available regarding this limitation. > */ Looks good. I will add this. Just let me know what the smallest value is for the timeout in the BIOS and/or datasheet and I will do the necessary for inclusion in the watchdog tree. Kind regards, Wim. -- To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
