Re: Very bad regression from 2.6.33 as of 1600f9def

Previous thread: Re: [PATCH] arm/mach-msm: Fix wrong include path. by Daniel Walker on Wednesday, April 28, 2010 - 6:48 pm. (1 message)

Next thread: linux-next: manual merge of the mtd tree with Linus' tree by Stephen Rothwell on Wednesday, April 28, 2010 - 7:58 pm. (1 message)
From: Alex Elsayed
Date: Wednesday, April 28, 2010 - 7:28 pm

I'm being bit by an extremely troublesome regression on HEAD (when the kernel
was built, at least) of Linus' tree. The symptoms are that the kernel boots,
blinks the cursor a few times, then reboots without showing any messages or
switching to the uvesafb mode specified on the commandline. Trying to boot the
kernel in qemu causes it to hang with a blank screen and a non-blinking cursor,
not even printing anything to a serial terminal. .config is below.

I'm on an MSI-1651-ID2 laptop, core2duo T5550, 4g ram, compiled with GCC 4.4.3.

#
# Automatically generated make config: don't edit
# Linux kernel version: 2.6.34-rc5
# Tue Apr 27 18:12:53 2010
#
CONFIG_64BIT=y
# CONFIG_X86_32 is not set
CONFIG_X86_64=y
CONFIG_X86=y
CONFIG_OUTPUT_FORMAT="elf64-x86-64"
CONFIG_ARCH_DEFCONFIG="arch/x86/configs/x86_64_defconfig"
CONFIG_GENERIC_TIME=y
CONFIG_GENERIC_CMOS_UPDATE=y
CONFIG_CLOCKSOURCE_WATCHDOG=y
CONFIG_GENERIC_CLOCKEVENTS=y
CONFIG_GENERIC_CLOCKEVENTS_BROADCAST=y
CONFIG_LOCKDEP_SUPPORT=y
CONFIG_STACKTRACE_SUPPORT=y
CONFIG_HAVE_LATENCYTOP_SUPPORT=y
CONFIG_MMU=y
CONFIG_ZONE_DMA=y
CONFIG_NEED_DMA_MAP_STATE=y
CONFIG_GENERIC_ISA_DMA=y
CONFIG_GENERIC_IOMAP=y
CONFIG_GENERIC_BUG=y
CONFIG_GENERIC_BUG_RELATIVE_POINTERS=y
CONFIG_GENERIC_HWEIGHT=y
CONFIG_GENERIC_GPIO=y
CONFIG_ARCH_MAY_HAVE_PC_FDC=y
# CONFIG_RWSEM_GENERIC_SPINLOCK is not ...
From: Justin P. Mattock
Date: Wednesday, April 28, 2010 - 7:52 pm

I'm being bit by an extremely troublesome regression on HEAD (when the 
kernel
was built, at least) of Linus' tree. The symptoms are that the kernel boots,
blinks the cursor a few times, then reboots without showing any messages or
switching to the uvesafb mode specified on the commandline. Trying to 
boot the
kernel in qemu causes it to hang with a blank screen and a non-blinking 
cursor,
not even printing anything to a serial terminal. .config is below.

I'm on an MSI-1651-ID2 laptop, core2duo T5550, 4g ram, compiled with GCC 
4.4.3.

<----  clip  ----->

what is the last good kernel?
can you bisect this?

Justin P. Mattock
--

From: Alex Elsayed
Date: Wednesday, April 28, 2010 - 8:05 pm

Last good kernel was 2.6.33 release, right after tagging.
I'll try and bisect. Probably be a while before I have anything,
since the CPU is rather slow. Hopefully last good will work in qemu.

--

From: Justin P. Mattock
Date: Wednesday, April 28, 2010 - 8:12 pm

yep.. luckily 33 is not that far behind
so it shouldn't take too long.
As for the CPU patience my friend..

Justin P. Mattock
--

From: Alex Elsayed
Date: Wednesday, April 28, 2010 - 9:50 pm

Argh, the default resume partition thing was a red herring.
Here's the text of the diff, if it's useful to anyone, working kernel is
zen-sources:

--- working_but_nonstandard_kernel	2010-04-28 21:35:02.321863681 -0700
+++ .config	2010-04-28 20:18:36.192271752 -0700
@@ -1,7 +1,7 @@
 #
 # Automatically generated make config: don't edit
-# Linux kernel version: 2.6.33-zen0 "Eat It Raw"
-# Wed Feb 24 20:40:05 2010
+# Linux kernel version: 2.6.33
+# Wed Apr 28 20:18:36 2010
 #
 CONFIG_64BIT=y
 # CONFIG_X86_32 is not set
@@ -61,7 +61,6 @@
 #
 # General setup
 #
-# CONFIG_SCHED_BFS is not set
 CONFIG_EXPERIMENTAL=y
 CONFIG_LOCK_KERNEL=y
 CONFIG_INIT_ENV_ARG_LIMIT=32
@@ -107,8 +106,8 @@
 CONFIG_GROUP_SCHED=y
 CONFIG_FAIR_GROUP_SCHED=y
 CONFIG_RT_GROUP_SCHED=y
-CONFIG_USER_SCHED=y
-# CONFIG_CGROUP_SCHED is not set
+# CONFIG_USER_SCHED is not set
+CONFIG_CGROUP_SCHED=y
 CONFIG_CGROUPS=y
 # CONFIG_CGROUP_DEBUG is not set
 CONFIG_CGROUP_NS=y
@@ -143,12 +142,6 @@
 CONFIG_INITRAMFS_COMPRESSION_LZMA=y
 # CONFIG_INITRAMFS_COMPRESSION_LZO is not set
 CONFIG_CC_OPTIMIZE_FOR_SIZE=y
-
-#
-# Custom Build Flags
-#
-CONFIG_CUSTOM_CFLAGS=""
-CONFIG_CUSTOM_MAKEFLAGS=""
 CONFIG_SYSCTL=y
 CONFIG_ANON_INODES=y
 # CONFIG_EMBEDDED is not set
@@ -185,7 +178,6 @@
 CONFIG_COMPAT_BRK=y
 # CONFIG_SLAB is not set
 CONFIG_SLUB=y
-# CONFIG_SLQB is not set
 # CONFIG_SLOB is not set
 CONFIG_PROFILING=y
 CONFIG_TRACEPOINTS=y
@@ -238,10 +230,8 @@
 CONFIG_IOSCHED_CFQ=y
 CONFIG_CFQ_GROUP_IOSCHED=y
 # CONFIG_DEBUG_CFQ_IOSCHED is not set
-CONFIG_IOSCHED_SIO=m
 # CONFIG_DEFAULT_DEADLINE is not set
 CONFIG_DEFAULT_CFQ=y
-# CONFIG_DEFAULT_SIO is not set
 # CONFIG_DEFAULT_NOOP is not set
 CONFIG_DEFAULT_IOSCHED="cfq"
 CONFIG_PREEMPT_NOTIFIERS=y
@@ -315,7 +305,6 @@
 # CONFIG_MK6 is not set
 # CONFIG_MK7 is not set
 # CONFIG_MK8 is not set
-# CONFIG_MK10 is not set
 # CONFIG_MCRUSOE is not set
 # CONFIG_MEFFICEON is not set
 # CONFIG_MWINCHIPC6 is not set
@@ -329,10 +318,8 @@
 ...
From: Alex Elsayed
Date: Wednesday, April 28, 2010 - 9:36 pm

On further examination (diff'ing /proc/config.gz and the config that isn't
working) it appears likely that the answer is simply that I am an idiot, having
set the default resume partition to an LVM volume that isn't present at the
beginning of boot. Recompiling and checking now.

--

From: Alex Elsayed
Date: Wednesday, April 28, 2010 - 9:28 pm

Uh oh. 2.6.33 compiled with this config (taking all defaults for make oldconfig)
doesn't boot either. Totally lost now.

--

Previous thread: Re: [PATCH] arm/mach-msm: Fix wrong include path. by Daniel Walker on Wednesday, April 28, 2010 - 6:48 pm. (1 message)

Next thread: linux-next: manual merge of the mtd tree with Linus' tree by Stephen Rothwell on Wednesday, April 28, 2010 - 7:58 pm. (1 message)