Date: Sat, 19 Jun 2010 16:39:20 +0200 From: Ulrich =?utf-8?B?U3DDtnJsZWlu?= <uqs@spoerlein.net> To: Andriy Gapon <avg@freebsd.org> Cc: freebsd-arch@freebsd.org Subject: Re: KDB_TRACE in GENERIC Message-ID: <20100619143920.GM3632@acme.spoerlein.net> In-Reply-To: <4C1C7686.3050206@freebsd.org> References: <4C1C7686.3050206@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 19.06.2010 at 10:49:26 +0300, Andriy Gapon wrote: > > I propose to add the following options into GENERIC (perhaps only i386 and amd64): > > options KDB > options KDB_TRACE > options KDB_UNATTENDED > options SC_DISABLE_KDBKEY # <- not sure about his one > > This would increase kernel size only by a fraction, would not result in any POLA > violations (KDB_UNATTENDED, SC_DISABLE_KDBKEY) and would provide a useful service. > > Currently we get many panic reports that are almost impossible to analyze > because there is no stack trace. Especially this is true for panics at > installation time, when people use GENERIC from release media and do not have > ability to try a custom kernel. > > What do you think? I'd rather propose we turn on textdumps, as many people will be in X11 when disaster strikes and there's no easy way to get the crashdump then. But we would also need to bump the ddb capture buffer size significantly, as it cannot hold information for more than a few dozen processes right now. Uli
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100619143920.GM3632>