Date: Wed, 6 Feb 2002 20:43:10 +1100 (EST) From: Bruce Evans <bde@zeta.org.au> To: John Baldwin <jhb@FreeBSD.org> Cc: <cvs-committers@FreeBSD.org>, <cvs-all@FreeBSD.org>, Bruce Evans <bde@FreeBSD.org> Subject: RE: cvs commit: src/sys/i386/i386 db_interface.c Message-ID: <20020206201716.B2295-100000@gamplex.bde.org> In-Reply-To: <XFMail.020205215801.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 5 Feb 2002, John Baldwin wrote: > On 05-Feb-02 Bruce Evans wrote: > > On Tue, 5 Feb 2002, John Baldwin wrote: > >> Err, have you tested this? In my tests locally this still doesn't work > >> properly which is why I haven't committed it. :( > > > > PS: perhaps you are thinking of the flag in Debugger(). That is > > ... > > Hmm, the problem I was having is that interrupts were still firing while I was > in ddb. I could tell because new KTR entries due to clock interrutps kept > showing up. This can't happen :-). Except for bugs which would affect most forms of interrupt disablement. Perhaps there is a path through trap() which enables interrupts even for debugger traps, but only when they are enabled when the trap occurs. Ah, I see a related broken path, not for debugger traps but for pagefaults. Interrupts are enabled for pagefaults almost unconditionally, so a pagefault in ddb would cause problems. I think this causes the "Context switches not allowed in the Debugger" message. I thought that this was caused by a more fundamental bug. BTW, why does the spinlocking in kern_clock.c use MTX_QUIET? nanotime() for witness timestamps should work normally there. I noticed this when I uninlined mtx_*_spin_*(). The `flags' versions are not used anywhere else. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020206201716.B2295-100000>