On Fri, 2023-03-03 at 13:38 +0100, Thomas Huth wrote: > On 21/02/2023 18.45, jro...@linux.ibm.com wrote: > > From: Jared Rossi <jro...@linux.ibm.com> > > > > Check for a List Directed IPL Boot Record, which would supersede > > the CCW type > > entries. If the record is valid, proceed to use the new style > > pointers > > and perform LD-IPL. Each block pointer is interpreted as either an > > LD-IPL > > pointer or a legacy CCW pointer depending on the type of IPL > > initiated. > > > > In either case CCW- or LD-IPL is transparent to the user and will > > boot the same > > image regardless of which set of pointers is used. Because the > > interactive boot > > menu is only written with the old style pointers, the menu will be > > disabled for > > List Directed IPL from ECKD DASD. > > > > If the LD-IPL fails, retry the IPL using the CCW type pointers. > > > > If no LD-IPL boot record is found, simply perform CCW type IPL as > > usual. > > > > Signed-off-by: Jared Rossi <jro...@linux.ibm.com> > > > > --- > > pc-bios/s390-ccw/bootmap.c | 157 ++++++++++++++++++++++++++++---- > > ----- > > pc-bios/s390-ccw/bootmap.h | 30 ++++++- > > 2 files changed, 148 insertions(+), 39 deletions(-) > > Janosch, Jason, Eric, Christian, > > could you please help with reviewing + testing this patch?
Thanks for the heads up, Thomas. I've been using this on a z16 system with SE guests (with newer zipl to actually write out LD-IPL data; something that should have been mentioned here), and it seems to behave fine. I gave a quick glance through the code, and while nothing jumped out at me this code is rather opaque. I realize this is already queued by now but I'll still go through it later this week so it gets another set of eyeballs. Eric
