Home | History | Annotate | Line # | Download | only in doc
TODO revision 1.34 
      1 /* $NetBSD: TODO,v 1.34 2021/02/22 10:30:57 palle Exp $ */
      2 
      3 Things to be done:
      4 
      5 common:
      6 - make %g6 point to curcpu
      7 - make %g7 point to curlwp
      8 - change run-time checks for cpu type to function pointers
      9 
     10 sun4u:
     11 - GENERIC.UP kernel hangs on v445 (missing interrupt?)
     12 
     13 sun4v:
     14  - current status
     15      T5 ldom with 2 VCPU and 4GB:
     16        The kernel boots and starts userland when booting miniroot.fs.
     17        The sysinst tool starts properly and requests terminal type.
     18 	   Upon entering characters on the console, a crash occurs inside
     19 	   OpenBoot (properly trashed registers).
     20 	 T2000 ldom with 8 VCPU and 4GB:
     21 	   On this platform it crashes in /sbin/init doing an access() call where %o0 is corrupted (zero)
     22 - 64-bit kernel support
     23 - 32-bit kernel support
     24 - libkvm
     25 - ofwboot: tlb_init_sun4v() hardcodes number of slots to 64
     26 - locore.s: sun4v_datatrap missing implementation for trap level 1
     27 - check build without SUN4V defined
     28 - replace relevant references to %ver with GET_MAXCWP
     29 - pmap_mp_init(): sun4v missing handling
     30 - replace constructs like "wrpr %g0, PSTATE_KERN, %pstate" with NORMAL_GLOBALS
     31 - replace constructs line "wrpr %g0, PSTATE_INTR, %pstate" with ALTERNATE_GOBALS
     32 - sun4v tsb no need to lock... per cpu... anyway...
     33 - ci_tsb_desc->td_ctxidx: -1 or 1?
     34 - MP support - currently bypassed in pmap_bootstrap() for sun4v
     35 - vpci.c/vpcivar.h: cleanup FIXMEs
     36 - interrupts not handled properly (com at ebus only...)
     37 - mpt(4) complains: mpt0: Phy 0: Link Status Unknown
     38 - man pages for drivers imported from OpenBSD lke vpci, vbus, cbus, vdsk, ldc etc.
     39 - vdsk and ldc drivers: code maked with OPENBSD_BUSDMA - make the bus_dma stuff work properly
     40 - vbus.c: handle prom_getprop() memory leaks
     41 - locore.s: rft_user (sun4v specific manaul fill) - seems to work, but is it good enough (compared to openbsds rft_user?
     42  - platforms tested so far:
     43  -- qemu sun4v/niagara - reaches the point where the kernel needs to mount a root fs
     44  -- T2000 - mmu fault somewhere
     45  -- T5 - stickcmpr is not properly initialized ???
     46