> port-arm32/8564[open sw-bug] Kernel forgets that it is ptracing a process
> AND (these two are medium, serious)
> port-arm32/7122[open sw-bug] Breakpoints lost under heavy swapping
> AND

I saw these while trying to build (and debug) Mozilla on a SHARK.  I gave 
up after a while, since I was unable to make progress, and I don't have 
the source downloaded any more to check.  I am fairly sure there are some 
weirdnesses in the arm32 VM system somewhere, but I've no idea where.

> port-arm32/8156[open sw-bug] step/next in gdb do not work

The PR cited in this one (bin/7565) has now been closed.  Any remaining 
problems here are due to the fact that gcc (in the NetBSD tree) does not 
pass -x to the linker to force it to strip out local labels after linking 
(most assemblers do this during assembly, but for various reasons the arm 
assembler does not -- maybe cannot).  The local labels confuse gdb when it 
tries to single-step (it appears to the debugger that the function has 
changed).  Similarly, -x should be passed when building shared libs (or 
strip -x should be run afterwards).

> Priority: medium Severity: non-critical
> port-arm32/8484[open change-request] rewrite of Acorn ASC SCSI driver
> Submit? or has richard got a newer version?

I'm working on this one.

> port-arm32/12155[open change-request] Add libarch support for flushing the 
> read buffer
> Submit?

I thought the patch was in the PR.  If not, I'm fairly sure it was posted 
to this list.  We need to ensure that such a call only tries to issue the 
CP15 instruction on processors that have that feature.

> port-arm32/12363[open change-request] Minor performance tweak to bcopy_page.S
> I thought this went in?  or was it only part of it?

Some of it went in.  But other bits didn't.

R.