Well, we are towards the end of device removal at this point, having called bus_remove_device(dev) for example, but still we've got it on dpm_active ... This may not be technically wrong (ie. we should be able to recover from that), but it seems conceptually wrong and with pm_sleep_rwsem in place it can be avoided. Rafael --

Subject: track and print last unloaded module in the oops trace From: Arjan van de Ven <arjan@linux.intel.com> CC: rusty@rustcorp.com.au CC: ak@suse.de CC: mingo@elte.hu Based on a suggestion from Andi: In various cases, the unload of a module may leave some bad state around that causes a kernel crash AFTER a module is unloaded; and it's then hard to find which module caused that. This patch tracks the last unloaded module, and prints this as part of the module list in the oops trace...

Kernel is 2.6.24-rc6 + linuxpps patches, which are all to the serial port driver. 2.6.23 was known stable. I haven't tested earlier 2.6.24 releases. I think it happened once before; I got a black-screen lockup with keyboard LEDs blinking, but that was with X running so I couldn't see a console oops. But given that I installed 2.6.24-rc6 about 24 hours ago, that's a disturbing pattern. (N.B. I was pretty careful, but the following was transcribed by hand.) BUG: unable to handle kernel paging ...

It's been two weeks since rc6, but let's face it, with xmas and new years (and birthdays) in between, there hasn't actually been a lot of working days, and the incremental patch from -rc6 is about half the size of the one from rc5->rc6. And I'll be charitable and claim it's because it's all stabilizing, and not because we've all been in a drunken stupor over the holidays. The shortlog (appended below) is short and fairly informative. It's all really just a lot of rather small changes. T...

Hello folks, 2.6.24-rc6 regresses on the 10000 network interface creation test relative to 2.6.23. The cause appears to be the new code in sysctl_check_lookup(), which shows up as the #1 item while profiling. Is a revert of this new code possible until its scaling issues are fixed? 2.6.23 can do more than 100 new network interfaces per second for the first few thousand devices, but with 2.6.24-rc6 the results drop off rather dramatically to less than 10 interfaces per second. The 10000 in...

That could well be the root cause as the full log shows: Jan 6 01:21:02 faramir kernel: ppdev0: registered pardevice Jan 6 01:21:03 faramir kernel: sysctl table check failed: /dev/parport/parport0/devices/ppdev0/timeslice Sysctl already exists Jan 6 01:21:03 faramir kernel: Pid: 11078, comm: hpijs Not tainted 2.6.24-rc6 #1 Jan 6 01:21:03 faramir kernel: Jan 6 01:21:03 faramir kernel: Call Trace: Jan 6 01:21:03 faramir kernel: [<ffffffff8024c104>] set_fail+0x3f/0x47 Jan 6 01:21:03 f...

Hi Andrew, Thanks for the email back. Ever written the previous year on your checks? Well when I was googling about for folks with similar woes, the posts I had thought were from this January were actually from Jan 2007. O Well! Not to mention it's In this case, not needed. I spent friday taking a tour of the codepath and basically narrowed it down to a device driver passing bum data. So if this ever happens in the future and someone somewhere seems the message about trying to load binfm...

I'm afraid this is pretty much wrong. The scsi.h being referred to here is the local scsi.h in the build directory, so it should be included as a string. For #include, "..." means begin the search in the local directory (what is wanted here) and <...> means begin the search starting with the predefined include paths. The rest of the patch looks like a spurious (unrelated and undescribed) downcasing of TRUE and FALSE. James --

On Sun, 6 Jan 2008 17:27:38 +0100 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0libata.dm= I think that libata.dma=3D1 is designed to behave like=20 CONFIG_IDEDMA_ONLYDISK used to. "hdx=3Dnodma" is more fine-grained than Alan's original patch took care of PATA. SATA support was added =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0libata.dm= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0libata.dm= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0...

Hi, I just switched to libata (pata) on my laptop and the immediate panic made it impossible to figure out why my boot partition wasn't available. After applying this little patch I could check boot printk output and then saw everything was properly recognized and only scsi-disk support was missing. Signed-off-by: Bernd Schubert <bernd-schubert@gmx.de> Index: zd1211rw.git.beno/init/do_mounts.c =================================================================== --- zd1211rw.git.beno.o...

Hello Ingo, There is no dump_stack() here, but disc detection is relatively early in boot process and on all these information are already scrolled off screen when the panic is done. For this and any other panic it would be optimal if scrolling still would work, but scrolling also requires kernel code, so I see there's a reason not to this for all panics. However, for this boot problem I tend to say there's no need to panic at all... Btw, not all stack straces are useless, *most* of them...

Here's a couple of small additions to BUG-HUNTING. 1. point out that you can list code in gdb with only one command (gdb) l *(<symbol> + offset) 2. give a very brief hint how to decode module symbols in call traces Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk> ---- It would have saved me some time if I'd known both of these things earlier, so I thought it might just be of help to anyone else new to the kernel. Richard diff --git a/Documentation/BUG-HUNTING b/Do...

* Convert legacy host drivers to use struct ide_port_info + ide_device_add() instead of open-coding initialization of IDE port data structures. * Add ->cable_detect method for cable detection + use it in PCI host drivers. * Fix some bugs in the process (handling of "idex=ata66" in atiixp, cs5535 and scc_pata and handling I/O 32-bit setting in dtc2278 and au1xxx-ide). diffstat: drivers/ide/arm/icside.c | 25 ++++--- drivers/ide/arm/ide_arm.c | 2 drivers/ide/ar...

* Remove 'struct pci_dev *dev' argument from ide_hwif_setup_dma(). * Un-static ide_hwif_setup_dma() and add CONFIG_BLK_DEV_IDEDMA_PCI=n version. * Add 'const struct ide_port_info *d' argument to ide_device_add[_all](). * Factor out generic ports init from ide_pci_setup_ports() to ide_init_port(), move it to ide-probe.c and call it in in ide_device_add_all() instead of ide_pci_setup_ports(). * Move ->mate setup to ide_device_add_all() from ide_port_init(). * Add IDE_HFLAG_NO_AUTOTUNE ...

Always set DMA masks in ide_pci_setup_ports() to make sure that the valid masks for a host are set. Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com> --- +34 bytes drivers/ide/setup-pci.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) Index: b/drivers/ide/setup-pci.c =================================================================== --- a/drivers/ide/setup-pci.c +++ b/drivers/ide/setup-pci.c @@ -556,10 +556,15 @@ void ide_pci_setup_ports(struct pci_d...

Don't override the cable type if the "idex=ata66" parameter was used. While at it: * atiixp.c: factor out cable detection to atiixp_cable_detect() from init_hwif_atiixp(). * cs5535.c: pass 'ide_hwif_t *hwif' instead of 'struct pci_dev *dev' to cs5535_cable_detect(). * scc_pata.c: factor out cable detection to scc_cable_detect() from init_hwif_scc() and remove incorrect comment. Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com> --- +16 bytes drivers/ide/pci/atiix...

On DTC2278 32-bit I/O has to be enabled for both devices on the port so always enable it during init time and disallow further changes. Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com> --- drivers/ide/ide.c | 4 ---- drivers/ide/legacy/dtc2278.c | 12 ++++++------ 2 files changed, 6 insertions(+), 10 deletions(-) Index: b/drivers/ide/ide.c =================================================================== --- a/...

Steven asked for per group periods in order to get closer to RMA or EDF scheduling. Use the fancy new hrtimers to provide a per group period Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> --- include/linux/sched.h | 2 kernel/sched.c | 229 ++++++++++++++++++++++++++++++++++++++++++----- kernel/sched_rt.c | 61 ++++++------ kernel/sysctl.c | 2 kernel/time/tick-sched.c | 5 - 5 files changed, 237 insertions(+), 62 deletions(-) Index:...

Currently all highres=off timers are run from softirq context, but HRTIMER_CB_IRQSAFE_NO_SOFTIRQ timers expect to run from irq context. Fix this up by splitting it similar to the highres=on case. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> --- include/linux/hrtimer.h | 5 - kernel/hrtimer.c | 232 +++++++++++++++++++++++++----------------------- kernel/timer.c | 3 3 files changed, 125 insertions(+), 115 deletions(-) Index: linux-2.6/kernel/hrtimer.c ...

hrtimer_wakeup creates a base->lock rq->lock lock dependancy. Avoid this by switching to HRTIMER_CB_IRQSAFE_NO_SOFTIRQ which doesn't hold base->lock. This fully untangles hrtimer locks from the scheduler locks, and allows hrtimer usage in the scheduler proper. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> --- kernel/hrtimer.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) Index: linux-2.6/kernel/hrtimer.c =====================================...

this time compile tested on all 16 combinations of: CONFIG_SMP CONFIG_FAIR_GROUP_SCHED CONFIG_HIGH_RES_TIMERS CONFIG_NO_HZ ran some but not all combinations -- --

Steven mentioned the fun case where a lock holding task will be throttled. Simple fix: allow groups that have boosted tasks to run anyway. This is ofcourse not quite correct. Needs more tricks. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> --- kernel/sched.c | 3 +++ kernel/sched_rt.c | 48 ++++++++++++++++++++++++++++++++++++++++-------- 2 files changed, 43 insertions(+), 8 deletions(-) Index: linux-2.6/kernel/sched.c ==================================================...

don't start the load_balance_monitor when there is only a single cpu. rename the kthread because its currently longer than TASK_COMM_LEN Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> --- kernel/sched.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) Index: linux-2.6/kernel/sched.c =================================================================== --- linux-2.6.orig/kernel/sched.c +++ linux-2.6/kernel/sched.c @@ -7070,8 +7070,11 @@ void __init sched_init_smp(void) ...

From: Julia Lawall <julia@diku.dk> A call to cpufreq_cpu_get should be matched by a call to cpufreq_cpu_put. The problem was fixed using the following semantic patch. (http://www.emn.fr/x-info/coccinelle/) // <smpl> @@ type T; identifier E; expression E1, E2; int ret; statement S; @@ T E; ... when any ( E = cpufreq_cpu_get(...); if (E == NULL) S | if ((E = cpufreq_cpu_get(...)) == NULL) S ) ... when != E2 = E when strict ( cpufreq_cpu_put(E); ...

ioport_{32|64}.c unification. This patch unifies the code from the ioport_32.c and ioport_64.c files. Tested and working fine with i386 and x86_64 kernels. Signed-off-by: Miguel Botón <mboton@gmail.com> diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 0903bbf..92d9ac6 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -10,7 +10,7 @@ CFLAGS_vsyscall_64.o := $(PROFILING) -g0 obj-y := process_$(BITS).o signal_$(BITS).o entry_$(BITS).o o...