NetBSD-Bugs archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: port-macppc/59014: Shutdown -r now often freezes on macppc b&w G3
The following reply was made to PR port-macppc/59014; it has been noted by GNATS.
From: Chris Tucker <capa150%gmail.com@localhost>
To: Taylor R Campbell <riastradh%netbsd.org@localhost>, gnats-bugs%netbsd.org@localhost
Cc:
Subject: Re: port-macppc/59014: Shutdown -r now often freezes on macppc b&w G3
Date: Tue, 21 Jan 2025 21:58:08 -0800
--0000000000002e18b8062c452e69
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
On Tue, Jan 21, 2025 at 11:44=E2=80=AFAM Taylor R Campbell <riastradh@netbs=
d.org>
wrote:
> > Date: Tue, 21 Jan 2025 11:25:31 -0800
> > From: Chris Tucker <capa150%gmail.com@localhost>
> >
> > netbsd# rndctl -S seed2 (works,
> > netbsd# rndctl -S seed2 (works)
> > netbsd# rndctl -S seed2 (works)
> > netbsd# rndctl -S seed2 (freezes, and I press ctrl-t)
> > [ 139.8510292] load: 0.42 cmd: rndctl 633 [biowait] 0.01u 0.00s 0% 142=
4k
>
> Interesting. What kind of disk do you have?
It's an old 12.0GB IBM Travelstar DJSA-220 that I grabbed out of my box o'
parts to test NetBSD 10.1 with. I have been using NetBSD 9.1 on my
"official" system (dual-boot with MacOS 8.6), running off a new-old-stock
30GB Maxtor Diamondmax I bought off ebay a year ago. Both IDE.
I wanted to run through the 10.1 install process without disturbing my
functional 9.1 system, hence the use of the old drive.
> In crash(8) or ddb, can you get the kernel stack trace of the rndctl
> process? In this case, the pid is 633, so it would be:
>
> crash> bt/t 0t633
>
> (Note the `0t' prefix for decimal.)
>
> I bet it's going to come up in sys_fsync_range, which is going to
> reveal a problem with fsync_range(FDISKSYNC) on your disk controller
> or hard disk.
>
I ran the rndctl command in the background to get it to fail.
While in crash on console:
PID LID S CPU FLAGS STRUCT LWP * NAME WAI=
T
4941 4941 3 0 0 fb3a680
rndctl biowait
>From an ssh session on another machine at the same time:
netbsd# crash
Crash version 10.1, image version 10.1.
Output from a running system is unreliable.
crash> bt/t 0t4941
trace: pid 4941 lid 4941 at 0x172cb00
0x0172cb60: at cpu_switchto+0x28
0x0172cb70: at mi_switch+0x140
0x0172cbb0: at sleepq_block+0xe0
0x0172cbd0: at cv_wait+0xc0
0x0172cc00: at bwrite+0x150
0x0172cc20: at ffs_update.part.0+0x3cc
0x0172cc60: at ufs_gro_rename+0xec
0x0172ccd0: at genfs_sane_rename+0x220
0x0172cd40: at ufs_sane_rename+0x3c
0x0172cd80: at genfs_insane_rename+0x9c
0x0172cdb0: at VOP_RENAME+0x80
0x0172cdf0: at do_sys_renameat.isra.0+0x558
0x0172cec0: at syscall+0x294
0x0172cf20: user SC trap #128 by 0xfdcac858: srr1=3D0xd032
r1=3D0xffffe0c0 cr=3D0x44000482 xer=3D0x20000000 ctr=3D0xfdcac8=
50
I installed NetBSD 10.1 on a third IDE drive, and the error persists.
I tried it on my 9.1 machine, running randctl -L seed 16 times in a row
before it froze on the 16th time; hit ctrl-t and saw "biowait" again. But I
suspect randctl wasn't meant to be run 16 times in rapid succession, if
that matters.
Guess it could be a hardware problem, IDE controller mebbe. I'll try all
the usual hw fixes (cuda, pram, etc) and mebbe look into a PCI IDE
controller card, or see if I can dig up a suitable SCSI drive.
--0000000000002e18b8062c452e69
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div dir=3D"ltr"><div dir=3D"ltr"><div dir=3D"ltr"><div dir=3D"ltr"><br></d=
iv><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On =
Tue, Jan 21, 2025 at 11:44=E2=80=AFAM Taylor R Campbell <<a href=3D"mail=
to:riastradh%netbsd.org@localhost" target=3D"_blank">riastradh%netbsd.org@localhost</a>> wro=
te:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px =
0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(2=
04,204,204);padding-left:1ex">> Date: Tue, 21 Jan 2025 11:25:31 -0800<br=
>
> From: Chris Tucker <<a href=3D"mailto:capa150%gmail.com@localhost"; target=3D"=
_blank">capa150%gmail.com@localhost</a>><br>
> <br>
> netbsd# rndctl -S seed2=C2=A0 =C2=A0(works,=C2=A0<br>
> netbsd# rndctl -S seed2=C2=A0 =C2=A0(works)<br>
> netbsd# rndctl -S seed2=C2=A0 =C2=A0(works)<br>
> netbsd# rndctl -S seed2=C2=A0 =C2=A0(freezes, and I press ctrl-t)<br>
> [ 139.8510292] load: 0.42=C2=A0 cmd: rndctl 633 [biowait] 0.01u 0.00s =
0% 1424k<br>
<br>
Interesting.=C2=A0 What kind of disk do you have?</blockquote><div><br></di=
v><div>It's an old 12.0GB IBM Travelstar DJSA-220 that I grabbed out of=
my box o' parts to test NetBSD 10.1 with. I have been using NetBSD 9.1=
on my "official" system (dual-boot with MacOS 8.6), running off =
a new-old-stock 30GB Maxtor Diamondmax I bought off ebay a year ago. Both I=
DE.</div><div><br></div><div>I wanted to run through the 10.1 install proce=
ss without disturbing my functional 9.1 system, hence the use of the old dr=
ive.</div><div><br></div><blockquote class=3D"gmail_quote" style=3D"margin:=
0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left=
-color:rgb(204,204,204);padding-left:1ex">=C2=A0</blockquote><blockquote cl=
ass=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px=
;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1e=
x">
<br>
In crash(8) or ddb, can you get the kernel stack trace of the rndctl<br>
process?=C2=A0 In this case, the pid is 633, so it would be:<br>
<br>
crash> bt/t 0t633<br>
<br>
(Note the `0t' prefix for decimal.)<br>
<br>
I bet it's going to come up in sys_fsync_range, which is going to<br>
reveal a problem with fsync_range(FDISKSYNC) on your disk controller<br>
or hard disk.<br></blockquote><div><br></div><div>I ran the rndctl=C2=A0com=
mand in the background to get it to fail.</div><div><br></div><div>While in=
crash on console:</div><div>PID =C2=A0 =C2=A0LID =C2=A0 =C2=A0 S =C2=A0 =
=C2=A0CPU =C2=A0 =C2=A0 FLAGS =C2=A0 =C2=A0 =C2=A0 =C2=A0 STRUCT LWP =C2=A0=
* =C2=A0 =C2=A0 =C2=A0 NAME =C2=A0 =C2=A0 =C2=A0WAIT</div><div>4941 =C2=A04=
941 =C2=A0 3 =C2=A0 =C2=A0 0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A00 =C2=
=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 fb3a680 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rndctl =C2=A0 =C2=
=A0 =C2=A0biowait</div><div><br></div><div>From an=C2=A0ssh session on anot=
her machine at the same time:</div><div>netbsd# crash<br>Crash version 10.1=
, image version 10.1.<br>Output from a running system is unreliable.<br>cra=
sh> bt/t 0t4941<br>trace: pid 4941 lid 4941 at 0x172cb00<br>0x0172cb60: =
at cpu_switchto+0x28<br>0x0172cb70: at mi_switch+0x140<br>0x0172cbb0: at sl=
eepq_block+0xe0<br>0x0172cbd0: at cv_wait+0xc0<br>0x0172cc00: at bwrite+0x1=
50<br>0x0172cc20: at ffs_update.part.0+0x3cc<br>0x0172cc60: at ufs_gro_rena=
me+0xec<br>0x0172ccd0: at genfs_sane_rename+0x220<br>0x0172cd40: at ufs_san=
e_rename+0x3c<br>0x0172cd80: at genfs_insane_rename+0x9c<br>0x0172cdb0: at =
VOP_RENAME+0x80<br>0x0172cdf0: at do_sys_renameat.isra.0+0x558<br>0x0172cec=
0: at syscall+0x294<br>0x0172cf20: user SC trap #128 by 0xfdcac858: srr1=3D=
0xd032<br>=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 r1=3D0xffffe0c0 cr=3D0x=
44000482 xer=3D0x20000000 ctr=3D0xfdcac850<br></div><div><br></div><div>I i=
nstalled NetBSD 10.1 on a third IDE drive, and the error persists.</div><di=
v><br></div><div>I tried it on my 9.1 machine, running randctl -L seed 16 t=
imes in a row before it froze on the 16th time; hit ctrl-t and saw "bi=
owait" again. But I suspect randctl=C2=A0wasn't meant to be run 16=
times in rapid succession, if that matters.</div><div><br></div><div>Guess=
it could be a hardware problem, IDE controller mebbe. I'll try all the=
usual hw fixes (cuda, pram, etc) and mebbe look into a PCI IDE controller =
card, or see if I can dig up a suitable SCSI drive.</div></div></div>
</div>
</div>
--0000000000002e18b8062c452e69--
Home |
Main Index |
Thread Index |
Old Index