Date: Fri, 29 May 2015 04:30:05 -0500 From: "Zane C. B-H." <v.velox@vvelox.net> To: Trevor Roydhouse <trev@sentry.org> Cc: freebsd-arm@freebsd.org Subject: Re: Raspberry Pi 2 - Xorg issues Message-ID: <20150529043005.1ffe79f3@vixen42.vulpes.vvelox.net> In-Reply-To: <20150528115259.44dbd337@vixen42.vulpes.vvelox.net> References: <555E80F4.7030300@sentry.org> <20150526090259.30c74175@vixen42.vulpes.vvelox.net> <55651788.2040406@sentry.org> <20150528115259.44dbd337@vixen42.vulpes.vvelox.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 28 May 2015 11:52:59 -0500
"Zane C. B-H." <v.velox@vvelox.net> wrote:
> On Wed, 27 May 2015 11:02:00 +1000
> Trevor Roydhouse <trev@sentry.org> wrote:
>
> > Zane C. B-H. wrote:
> > > On Fri, 22 May 2015 11:05:56 +1000
> > > Trevor Roydhouse <trev@sentry.org> wrote:
> > [...]
> > >> Solution to get X up:
> > >>
> > >> Leave out DefaultDepth nn in the Screen section, but this
> > >> yields very slow motion screen draws and long cursor tails.
> > >> However, if I run a find / in an xterm, screen draws return to
> > >> normal until find finishes... weird.
> > >>
> > >> Any other solutions?
> > >
> > > I am seeing the exact same thing using a image I built using a
> > > fresh source tree on the 23rd.
> > >
> > > It appears to be load related in some manner. For example I can
> > > also get it to behave nicely. If I start compiling something
> > > like firefox or the like.
> > >
> > > Make any headway with this on your end?
> >
> > I wondered about whether it was caused by powerd, but disabling it
> > had no effect.
> >
> > In the meantime I've reverted to my RPi B+ which works nicely as
> > far as X goes.
>
> On another note, I've just tried it with out any of the debugging
> stuff compiled in and that did not make a difference. Now seeing if
> hz makes a difference.
>
> On a interesting note, I did find that with out invariants, the
> noise on the +5V GPIO line is different. It takes awhile for longer
> for each of the additional 3 sections to form and all 4 of them are
> not as consistent.
So far I've manage to rule out SMP(turned it off), the
scheduler(switched to 4BSD as that is what that kern conf used), and
various hz settings(just spit balling there).
Also decided to see what
would happen when I used the timer device for the RPI-B and that just
resulted in no timer being found(I though this was suppose to be
backwards compatible?).
Hmm... decided to see what truss is seeing and I am seeing this come
up occasionally. This happens when it gets slow and the screen is
just more or less static.
SIGNAL 23 (SIGIO)
sigprocmask(SIG_SETMASK,SIGIO,0x0) = 0 (0x0)
SIGNAL 14 (SIGALRM)
sigprocmask(SIG_SETMASK,SIGALRM|SIGIO,0x0) = 0 (0x0)
sigreturn(0xbfbff524,0x2027f9d4,0x20c03300,0x1,0x0,0x0) = 0 (0x0)
select(8,{7},0x0,0x0,{0.000000 }) = 0 (0x0)
sigreturn(0xbfbff7f4,0x2027f9d4,0x20c03300,0x1,0x0,0x0) = 0 (0x0)
read(10,"&\^A\^B\0\^W\0@\0",4096) = 8 (0x8)
writev(0xa,0xbfbffa74,0x1,0x20268fa4,0x1,0x0) = 64 (0x40)
setitimer(0,{0.000000, 0.000000 },0x0) = 0 (0x0)
SIGNAL 14 (SIGALRM)
sigprocmask(SIG_SETMASK,SIGALRM,0x0) = 0 (0x0)
sigreturn(0xbfbff4d4,0x2027f9d4,0x20c03300,0x1,0x0,0x0) = 0 (0x0)
clock_gettime(4,{956.416096102 }) = 0 (0x0)
select(256,{1 3 4 6 8 9 10 11 12 13},0x0,0x0,0xffffffff) = 1 (0x1)
clock_gettime(4,{956.420401571 }) = 0 (0x0)
setitimer(0,{0.020000, 0.020000 },0x0) = 0 (0x0)
clock_gettime(4,{956.424181987 }) = 0 (0x0)
read(10,"&\^A\^B\0\^W\0@\0",4096) = 8 (0x8)
clock_gettime(4,{956.428843914 }) = 0 (0x0)
read(10,0x20f4a000,4096) ERR#35 'Resource temporarily unavailable'
writev(0xa,0xbfbffa74,0x1,0x20268fa4,0x1,0x0) = 32 (0x20)
setitimer(0,{0.000000, 0.000000 },0x0) = 0 (0x0)
clock_gettime(4,{956.437341154 }) = 0 (0x0)
select(256,{1 3 4 6 8 9 10 11 12 13},0x0,0x0,0xffffffff) ERR#4 'Interrupted system call'
SIGNAL 23 (SIGIO)
sigprocmask(SIG_SETMASK,SIGIO,0x0) = 0 (0x0)
select(8,{7},0x0,0x0,{0.000000 }) = 1 (0x1)
read(7,"\M^G\0\0\0\0\0\0\^?",64) = 8 (0x8)
clock_gettime(4,{957.157854644 }) = 0 (0x0)
clock_gettime(4,{957.160702404 }) = 0 (0x0)
select(1024,{7},0x0,0x0,{-382906384.-498216932 }) = 0 (0x0)
sigreturn(0xbfbff4d4,0x2027f9d4,0x20c03300,0x1,0x0,0x0) ERR#4 'Interrupted system call'
writev(0xa,0xbfbffa74,0x1,0x20268fa4,0x1,0x0) = 32 (0x20)
clock_gettime(4,{957.173545633 }) = 0 (0x0)
select(256,{1 3 4 6 8 9 10 11 12 13},0x0,0x0,0xffffffff) = 1 (0x1)
clock_gettime(4,{957.178277612 }) = 0 (0x0)
setitimer(0,{0.020000, 0.020000 },0x0) = 0 (0x0)
clock_gettime(4,{957.182889383 }) = 0 (0x0)
read(10,"\n\^A\^B\0\^V\0@\0\n\0\^B\0\^W\0"...,4096) = 20 (0x14)
clock_gettime(4,{957.188955008 }) = 0 (0x0)
clock_gettime(4,{957.193565581 }) = 0 (0x0)
clock_gettime(4,{957.196813029 }) = 0 (0x0)
read(10,0x20f4a000,4096) ERR#35 'Resource temporarily unavailable'
writev(0xa,0xbfbffa74,0x1,0x20268fa4,0x1,0x0) = 32 (0x20)
SIGNAL 14 (SIGALRM)
So yeah any one have any thoughts on this?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150529043005.1ffe79f3>