Jan (or others), I'm seeing an issue with the EHCI debug port functionality, that I'm wondering whether it is a known limitation of the hardware, or if this is a bug. On a particular system that I had limited debug capabilities in the past (Dell Inspiron 15 i3) I am seeing the debug messages abruptly stop during boot, followed by garbage that looks like an uninitialized serial port: ... [ 3.816669] Freeing SMP alternatives: 20k freed [ 3.819784] ACPI: Core revision 20121018 x�~x�~P This always stops with the same ACPI message. I'm wondering if some bus level reset is going on, that might be confusing the debug port, and perhaps it is not recovering from this? Any thoughts would be appreciated. Ben _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org http://lists.xen.org/xen-devel
>>> On 24.06.13 at 15:18, Ben Guthro <ben@guthro.net> wrote: > I''m seeing an issue with the EHCI debug port functionality, that I''m > wondering whether it is a known limitation of the hardware, or if this > is a bug. > > On a particular system that I had limited debug capabilities in the > past (Dell Inspiron 15 i3) I am seeing the debug messages abruptly > stop during boot, followed by garbage that looks like an uninitialized > serial port: > > > ... > [ 3.816669] Freeing SMP alternatives: 20k freed > [ 3.819784] ACPI: Core revision 20121018 > x~x~P > > > This always stops with the same ACPI message. > I''m wondering if some bus level reset is going on, that might be > confusing the debug port, and perhaps it is not recovering from this?Possibly - who knows what firmware does... Does the BIOS setup have any option to turn off as much of legacy emulation as possible (which otherwise may get turned off as ACPI gets enabled)? If so, I''d suggest trying this. If not, only the machine or BIOS vendors may know what''s going on there... The reset that the code is prepared to deal with happens in any case much later than where you show output stops, and I never saw any garbage coming through (albeit there is a certain percentage of cases where recovery doesn''t work, yet according to my testing that''s also the case for the original Linux code). Jan
On Mon, Jun 24, 2013 at 9:53 AM, Jan Beulich <JBeulich@suse.com> wrote:>>>> On 24.06.13 at 15:18, Ben Guthro <ben@guthro.net> wrote: >> I''m seeing an issue with the EHCI debug port functionality, that I''m >> wondering whether it is a known limitation of the hardware, or if this >> is a bug. >> >> On a particular system that I had limited debug capabilities in the >> past (Dell Inspiron 15 i3) I am seeing the debug messages abruptly >> stop during boot, followed by garbage that looks like an uninitialized >> serial port: >> >> >> ... >> [ 3.816669] Freeing SMP alternatives: 20k freed >> [ 3.819784] ACPI: Core revision 20121018 >> x~x~P >> >> >> This always stops with the same ACPI message. >> I''m wondering if some bus level reset is going on, that might be >> confusing the debug port, and perhaps it is not recovering from this? > > Possibly - who knows what firmware does... Does the BIOS setup > have any option to turn off as much of legacy emulation as > possible (which otherwise may get turned off as ACPI gets enabled)? > If so, I''d suggest trying this.Thanks. Disabling the "USB emulation" setting seemed to do it. Much appreciated. Ben