Greg
animals brewing
food and drink gardening
general health
history language
music multimedia
opinion photography
politics Stones Road house
technology
Greg's diary
recent entries
Translate this page
Select day in February 2018:
Su Mo Tu We Th Fr Sa
1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28
Select month:
2017 May Jun Jul Aug
2017 Sep Oct Nov Dec
2018 Jan Feb Mar Apr
Today's diary entry
Diary index
About this diary
Greg's home page
Greg's photos
Network link stats
Greg's other links
Copyright information
    
Groogle

Monday, 19 February 2018 Dereel
Top of page
next day
last day

Oil low, change it
Topic: general, opinion Link here

Last Thursday I received a voice mail from Lisa Sperber of Ballarat Automotive, telling me that my car was ready, but that <mumble> about the oil. I dropped in on Friday, but at 15:30 they were shut.

Today Lisa called again, during breakfast. The oil in the engine was dirty, and the level was low, so it needed changing. OK, more details please. No, that's enough. But I had the oil changed only a few 1000 km ago. In any case, I was in the middle of breakfast, can I call back? No, it'll only take a moment to check. I was in the middle of breakfast, can I call back? Line went quiet, so I hung up.

Called back later. It seems that the last service was in May 2016. How many kilometres have I done since then? Don't know, but that's enough. Car must be serviced. I pointed out that low oil level isn't a reason for a service: I can just top it up. OK, pick up your car and top it up yourself. <click>. And she hung up on me.

So is a service due? I still don't know. It seems that the last one was done at 165,000 km. What's the current mileage? Lisa didn't check. How often are the service intervals? Lisa says “less than 15,000”, but she didn't give any further details. Since she's so uncooperative, I'll have to find out for myself. Looking online showed a surprisingly complicated web page which, after many questions, told me that I needed an oil change every 7,500 miles.

That's because it's US American, and they still use archaic units. It corresponds to 12,000 km, inconveniently half way between the more plausible 10,000 or 15,000 km. But despite much searching I couldn't find a corresponding Australian page. I'll have to ask CJ Ellis, who has the same model with an instruction manual. And I'll have to check the current kilometres on Wednesday when I go in.

This isn't the first time I've had issues with Lisa. I've already commented on how she can be obstructive, though this is by far the worst I have experienced. They're a strange couple: Paul is competent, friendly and helpful, but just about every time I interact with Lisa, it's negative. It's also clear that the business is not doing well. When he bought it from Vic England, over 8 years ago, it was doing well, and there were two other mechanics working there. Now they're down to Paul and Lisa (who does the accounts and customer interface, fiercely insulating Paul from them), working only 4 days a week.

Why? It can't be the quality or the cost of the work; that's as good as I have had, and it's not like me not to complain about workmanship. My guess is that Lisa has driven them away. Under the present situation, it would be better for her to go home and stay home, and let Paul take over that side of things as well until business picks up. Then he could hire somebody of a more pleasant disposition.


More playing with HDR Express
Topic: photography, technology Link here

My previous experience with HDR Express 3 was not the most positive, but then, it is a Microsoft space program. Maybe I'm trying to do too much at once. Let's take one step at a time.

Fired it up again, this time remotely from euroa, and confirmed once again that I got a black screen. Started it locally and migrated to rdesktop, again with no problems. This time I took the normal interface. And once again it insisted on looking in C:\Users\grog\Pictures. But this time the Preferences screen looked completely different, and in particular didn't allow me to set a default folder.

Tried things anyway. It seems that when you have the contents of the directory displayed on the screen, you're not done yet: you have to select the ones you want (just the opposite of what the batch window does, it seems). And then it will merge things. I tried it with the sequence e-from-house-4-* and got what looked like a quite acceptable result:


This should be HDREx-1.png.  Is it missing?
Image title: HDREx 1          Dimensions:          2522 x 1416, 3076 kB
Make a single page with this image Hide this image
Make this image a thumbnail Make thumbnails of all images on this page
Make this image small again Display small version of all images on this page
All images taken on Saturday, 17 February 2018, thumbnails          All images taken on Saturday, 17 February 2018, small
Diary entry for Saturday, 17 February 2018 Complete exposure details

 

The horizontal bar towards the bottom seems to be due to incorrect screen refresh after I finally got rid of an irritating “Preset” bar. But the image looked acceptable, though possibly still a little too contrasty. In particular it was far ahead of my enfuse solution (on the left) regarding ghosting:

 
This should be e-from-house-0-detail.jpeg.  Is it missing?
Image title: e from house 0 detail
Display location on map
Complete exposure details
Dimensions: 424 x 638, 67 kB
Dimensions of original: 258 x 388, 54 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Saturday, 17 February 2018:
thumbnails    small images    diary entry
 
This should be HDREx-2-detail.png.  Is it missing?
Image title: HDREx 2 detail
Complete exposure details
Dimensions: 385 x 701, 418 kB
Dimensions of original: 201 x 366, 88 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Saturday, 17 February 2018:
thumbnails    small images    diary entry

In particular, Amber's left foot is completely missing in the enfuse version.

OK, save it, and got an output file e-from-house-4-1EV-0003-Tone&nbsp;Map&nbsp;Linear.tif, which lost much of the Exif data. It also blows the mind of my standard viewer xv. OK, use ImageMagick to convert to JPEG:

=== grog@eureka (/dev/pts/1) /Photos/6-HDR4-grog 64 -> convert e-from-house-4-1EV-0003-Tone-Map-Linear.tif e-from-house-4-1EV-0003-Tone-Map-Linear.jpeg
convert: Wrong data type 5 for "ExposureBiasValue"; tag ignored. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.
convert: Wrong data type 3 for "PixelXDimension"; tag ignored. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.
convert: Wrong data type 3 for "PixelYDimension"; tag ignored. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.
convert: Incompatible type for "FileSource"; tag ignored. `TIFFFetchNormalTag' @ warning/tiff.c/TIFFWarnings/896.
convert: Wrong data type 3 for "GainControl"; tag ignored. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.
convert: Unknown field with tag 42033 (0xa431) encountered. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.
convert: Unknown field with tag 42034 (0xa432) encountered. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.
convert: Unknown field with tag 42036 (0xa434) encountered. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.
convert: Unknown field with tag 42037 (0xa435) encountered. `TIFFReadCustomDirectory' @ warning/tiff.c/TIFFWarnings/896.

The result? Not what I was expecting. Here again the enfuse output on the left, the HDR Express output on the right:

 
This should be e-from-house-0.jpeg.  Is it missing?
Image title: e from house 0
Display location on map
Complete exposure details
Dimensions: 600 x 450, 78 kB
Dimensions of original: 5184 x 3888, 2974 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Saturday, 17 February 2018:
thumbnails    small images    diary entry
 
This should be e-from-house-0-1EV-HDREx.jpeg.  Is it missing?
Image title: e from house 0 1EV HDREx
Display location on map
Complete exposure details
Dimensions: 600 x 450, 124 kB
Dimensions of original: 5184 x 3888, 7096 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Saturday, 17 February 2018:
thumbnails    small images    diary entry

Why is this such a mess? The output image looks nothing like what I saw on the screen. And, like HDR Projects 4, it's nothing like what you could consider reasonable. Once again I'm left with the feeling that I have to outwit the software to get any useful results.


Tuesday, 20 February 2018 Dereel
Top of page
previous day

Panic!
Topic: technology, opinion Link here

Updated the system on teevee today, as I do about once a month. The last time was:

FreeBSD teevee.lemis.com 11.1-STABLE FreeBSD 11.1-STABLE #2 r327971: Mon Jan 15 10:55:53 AEDT 2018     grog@teevee.lemis.com:/home/obj/eureka/home/src/FreeBSD/svn/stable/11/sys/GENERIC  amd64

Nothing very interesting there, which is why I almost never mention it. Today it finished, I rebooted, went away briefly, and came back in time to see the system displaying:

reboot after panic: page fault
writing core to /var/crash/vmcore.1

Huh? I can't recall when I last had a panic with a FreeBSD -STABLE kernel. OK, watch what it does next. Finish rebooting normally, (automatically) start X and...

Fatal trap 12: page fault while in kernel mode
cpuid = 2; apic id = 02
fault virtual address    = 0x8
fault code               = supervisor read data, page not present
instruction pointer      = 0x20:0xffffffff82798f0b
stack pointer            = 0x28:0xfffffe011773b570
frame pointer            = 0x28:0xfffffe011773b640
code segment             = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags = interrupt enabled, resume, IOPL = 3
current process          = 1140 (Xorg)
trap number              = 12
panic: page fault
cpuid = 2
KDB: stack backtrace:
#0 0xffffffff80ac1637 at kdb_backtrace+0x67
#1 0xffffffff80a7b5a6 at vpanic+0x186
#2 0xffffffff80a7b413 at panic+0x43
#3 0xffffffff80ef7f22 at trap_fatal+0x352
#4 0xffffffff80ef7f79 at trap_pfault+0x49
#5 0xffffffff80ef77e6 at trap+0x2c6
#6 0xffffffff80ed7c80 at calltrap+0x8
#7 0xffffffff827a65f3 at nvidia_dev_dtor+0x23
#8 0xffffffff80946885 at devfs_fpdrop+0xc5
#9 0xffffffff80949592 at devfs_open+0x142
#10 0xffffffff8106754c at VOP_OPEN_APV+0x7c
#11 0xffffffff80b4d003 at vn_open_vnode+0x203
#12 0xffffffff80b4cbdd at vn_open_cred+0x34d
#13 0xffffffff80b45ec2 at kern_openat+0x212
#14 0xffffffff80ef8fa8 at amd64_syscall+0xa38
#15 0xffffffff80ed84f1 at fast_syscall_common+0x105
Uptime: 1m43s

I couldn't read all that, of course, so in to look at the dump. After all, when it comes to FreeBSD dumps, I wrote the book. OK:

=== root@teevee (/dev/pts/5) /var/crash 2 -> gdb -k /boot/kernel/kernel vmcore.1
gdb: unrecognized option `-k'
Use `gdb --help' for a complete list of options.
=== root@teevee (/dev/pts/5) /var/crash 3 -> kgdb /boot/kernel/kernel vmcore.1
GNU gdb 6.1.1 [FreeBSD]
...
This GDB was configured as "amd64-marcel-freebsd"...(no debugging symbols found)...
Attempt to extract a component of a value that is not a structure pointer.
Attempt to extract a component of a value that is not a structure pointer.
#0  0xffffffff80a7b38b in doadump ()
(kgdb) bt
#0  0xffffffff80a7b38b in doadump ()
#1  0xffffffff80a7b3b4 in doadump ()
#2  0xfffffe011773b200 in ?? ()
#3  0xffffffff80a7b10e in kern_reboot ()
Previous frame identical to this frame (corrupt stack?)

What went wrong there? Really corrupt kernel stack? What appeared on the screen seems plausible enough. And then I found a new file that never used to be there:

=== root@teevee (/dev/pts/5) /var/crash 4 -> l
total 802
...
-rw-------  1 root  wheel      146,739 20 Feb 15:54 core.txt.0
-rw-------  1 root  wheel          473 20 Feb 15:59 info.last
-rw-------  1 root  wheel  313,876,480 20 Feb 15:59 vmcore.1

That seemed to be related to the previous panic, but presumably the panic was the same. Looking inside showed all kinds of goodies, far beyond what I needed, but including a proper gdb trace:

#0  doadump (textdump=<value optimized out>) at pcpu.h:229
229     pcpu.h: No such file or directory.
        in pcpu.h
#0  doadump (textdump=<value optimized out>) at pcpu.h:229
#1  0xffffffff80a7b10e in kern_reboot (howto=260)
    at /eureka/home/src/FreeBSD/svn/stable/11/sys/kern/kern_shutdown.c:366
#2  0xffffffff80a7b5e0 in vpanic (fmt=<value optimized out>,
    ap=<value optimized out>)
    at /eureka/home/src/FreeBSD/svn/stable/11/sys/kern/kern_shutdown.c:759
#3  0xffffffff80a7b413 in panic (fmt=<value optimized out>)
    at /eureka/home/src/FreeBSD/svn/stable/11/sys/kern/kern_shutdown.c:690
#4  0xffffffff80ef7f22 in trap_fatal (frame=0xfffffe01176c04b0, eva=8)
    at /eureka/home/src/FreeBSD/svn/stable/11/sys/amd64/amd64/trap.c:817
...
#8  0xffffffff82799f0b in rm_free_unused_clients ()
   from /boot/modules/nvidia.ko
#9  0xfffffe01176c05d8 in ?? ()
#10 0xfffff80004bd0a00 in ?? ()
#11 0xfffffe01176c05c0 in ?? ()
#12 0xffffffff827a89fe in os_pci_read_dword () from /boot/modules/nvidia.ko
#13 0xffffffff827a75f3 in nvidia_dev_dtor () from /boot/modules/nvidia.ko
#14 0xffffffff80946885 in devfs_fpdrop (fp=<value optimized out>)
    at /eureka/home/src/FreeBSD/svn/stable/11/sys/fs/devfs/devfs_vnops.c:193

So it seems that since I last debugged a kernel, people have changed the method Yet Again. I had half expected that—that's why I tried two different incantations, but they're now both useless. Time for another RTFM.

   Debugging a crash dump
     By default, crash dumps are stored in the directory /var/crash.
     Investigate them from the kernel build directory with:

           gdb -k kernel.debug /var/crash/vmcore.29

That doesn't help much. It looks more like a case of WTFM., or possibly the issue is that this was the version of kgdb relating to the old kernel.

In any case, core.txt.0 (why isn't there a core.txt.1?) told me everything I needed to know: it's related to the nvidia X driver. Whose fault? Difficult to say, but my outstanding bug report means that I'm running an old version of the driver. Does it occur with the newest driver? That'll take a bit of testing. In the meantime the workaround was to revert to the old kernel.


This page contains (roughly) yesterday's and today's entries. I have a horror of reverse chronological documents, so all my diary entries are chronological. This page normally contains the last two days, but if I fall behind it may contain more. You can find older entries in the archive. Note that I often update a diary entry a day or two after I write it.     Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.


Greg's home page This month Greg's photos Greg's links

RSS 2.0 Valid XHTML 1.0!