New F11 for the XO-1 build 5-swap

[Paul Fox]

mikus wrote:
 > > root at OLPC:~# tail /media/olpcmtd0/var/log/dmesg
 > > [   25.269277] ip used greatest stack depth: 1732 bytes left
 > > [   27.747985] Adding 258040k swap on /dev/mmcblk0p2.  Priority:-1 extents:1 
 > ac
 > > ross:258040k SS
 > > 
 > > _note the "SS" at the end_ normally not present when booting is OK
 > 
 > I still think you are mistaken when you blame bootup handling of the 
 > swap partition.  For instance, I normally have the "SS" at the end 
 > of that line in /var/log/dmesg -- yet for me bootup works correctly.
 > [Except (see postscript) I do believe in 'internal timing' hurdles.]
 > 
 > My understanding of the way /var/log/dmesg works is that it is 
 > "buffered" - the system fills one "buffer" to a certain point, and 
 > on the XO [on Fedora?] it then stops writing to /var/log/dmesg (even 
 > though if from the command line one issues 'dmesg', what gets shown 
 > is additional lines beyond the one where /var/log/dmesg stops).  So 

the dmesg mechanism is even simpler than this.  the kernel stores
its log in an internal buffer.  the dmesg command shows you the
contents of that buffer at any point in time.  this buffer will
eventually fill, and the oldest entries will be overwritten
(i.e., it's a circular buffer).  at some point during boot, the
dmesg command is run, and the output is stored in /var/log/dmesg. 
the idea is to save the early boot messages, so that they don't
get lost when the kernel's internal buffer fills up.

paul 

 > the LAST line the system writes to the XO /var/log/dmesg file seems 
 > to be the one about adding the swap partition -- but my systems then 
 > continue bootup (without more lines being added to /var/log/dmesg). 
 >   That is why I'm not alarmed with you finding that particular line 
 > to be the last line in /var/log/dmesg
 > 
 > 
 > Turns out that the same line is also copied into /var/log/messages. 
 >   Would you please do 'rm /var/log/messages' (this gives you a clean 
 > file - /var/log/messages is otherwise not erased between reboots), 
 > and then reboot.  After your bootup has stalled, use Ubuntu (or 
 > something) to look at /var/log/messages (not just /var/log/dmesg). 
 > _IF__ the "time stamp sequence" in /var/log/messages continues 
 > beyond the "Adding 25804k swap ..." line - then hooking up the swap 
 > partition was *not* the last thing the system tried -- and whatever 
 > it did later is also a candidate for "why did bootup stall".
 > 
 > mikus
 > 
 > 
 > p.s.
 > Note that for me, boot-up sometimes stalls with the last line 
 > showing on the screen being "loading ... initrd.img" (or whatever it 
 > loads).  [This is while the screen still has a white background -- I 
 > normally press 'check' while booting, so the screen switches to a 
 > black background as the bootup progresses.]
 > 
 > My XO systems have *lots* of external (USB) devices.  I interpret 
 > the bootup stalling before the screen can switch to black to be an 
 > indication that some sort of 'internal timing' did not work 
 > correctly, resulting in something or other not being recognized. 
 > What I do in case of such a stall is to pull out one or more USB 
 > cables (this seems to affect the 'internal timing') and repeat the 
 > bootup from power-on.  Sooner or later bootup progresses to the 
 > black background screen - then I'm "in like Flynn".
 > 
 > _______________________________________________
 > Fedora-olpc-list mailing list
 > Fedora-olpc-list at redhat.com
 > https://www.redhat.com/mailman/listinfo/fedora-olpc-list

=---------------------
 paul fox, pgf at laptop.org