Date: Mon, 23 Dec 2013 09:00:20 -0700 From: Warner Losh <imp@bsdimp.com> To: John-Mark Gurney <jmg@funkthat.com> Cc: svn-src-releng@FreeBSD.org, svn-src-all@FreeBSD.org, src-committers@FreeBSD.org, Warner Losh <imp@FreeBSD.org> Subject: Re: svn commit: r259748 - releng/10.0/sys/arm/at91 Message-ID: <E711D79C-A0D5-4F05-8868-846CCEAC1E36@bsdimp.com> In-Reply-To: <20131222224718.GO99167@funkthat.com> References: <201312222231.rBMMVeql061111@svn.freebsd.org> <20131222224718.GO99167@funkthat.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Dec 22, 2013, at 3:47 PM, John-Mark Gurney wrote: > Warner Losh wrote this message on Sun, Dec 22, 2013 at 22:31 +0000: >> The real fix is to use cngrab/cnungrab function pointers to disable >> RXRDY interrupts while grabbed. However, that touches the MI uart >> code, so was disallowed for 10.0 due to the lateness of the hour = this >> fix was proposed. It works for mountroot, the most common atmel = kernel >> prompt use cases, but wouldn't work for GELI since it prompts later = in >> the boot process. >=20 > Hmmm, good point about geli, there have been complaints that the geli > prompt gets garbled by other output (and I'm anoyed by it too)... >=20 > Can we solve this w/ grab? How would the console code know the other > writers don't have the console grabed? Maybe by thread? cngrab/cnungrab was meant to toggle between userland and kernel use of = the console for input in rare cases. Contention doesn't seem to be = covered at all, but could be... Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E711D79C-A0D5-4F05-8868-846CCEAC1E36>