J-Soft Blog

Tuesday, July 21, 2009

This issue is no doubt very esoteric, but it caught me, so there must be someone else out there scratching their head about it. This is a problem that plagues the Linux kernel for the avr32 architecture version 2.6.22.atmel.6, and likely other versions. It is caused by a bug in gcc. I do not remember the exact version numbers off the top of my head, but I believe it was avr32-linux-gcc 4.2 which has a bug in it that causes a few problems, one of which is expressed in n_tty.c and causes a kernel oops whenever a character is sent to a pty. In other words, anything you type in a terminal session crashes the machine. Definitely not good! The solution is to patch n_tty.c as described below, or just use an older gcc. What follows is from a posting on the avr32 Linux kernel mailing list.

-----------------------------------------------------

I found what looks like another symptom of this gcc bug in
drivers/char/n_tty.c from kernel 2.6.22.atmel.6. It causes a kernel Oops
when you send a character to a pty (ie. type anything at all in a telnet
session).

Here's the Oops:
---------------------------------------------------------------------- #
Unable to handle kernel NULL pointer dereference at virtual address
00000000
ptbr = 91ceb800 pgd = 11d6bc66 pte = 00000000
Oops: Kernel access of bad area, sig: 11 [#1]
FRAME_POINTER chip: 0x01f:0x1e82 rev 2
Modules linked in:
PC is at n_tty_receive_buf+0xd4/0xb80
LR is at pty_write+0x2a/0x38
pc : [<900c41e0>]    lr : [<900c5d36>]    Not tainted
sp : 91d8fe2c  r12: 914fd000  r11: 914fd800
r10: 00000000  r9 : 00000000  r8 : 00000000
r7 : 91d8febc  r6 : 900c410c  r5 : 00000001  r4 : 0007f03c
r3 : 914fd000  r2 : 00000001  r1 : 91e5fb40  r0 : 914fd800
Flags: qvnzc
Mode bits: hrje....g
CPU Mode: Supervisor
Process: telnetd [246] (task: 91c6ea80 thread: 91d8e000)
Stack: (0x91d8fe2c to 0x91d90000)
fe20:                            914fd800 00000000 00000000 00000001
91d8ff8c fe40: 00000038 00000020 00000000 00000000 00000020 00000000
00000fa0 91fdbc80 fe60: 900102d8 00000000 91d8fe6c 00000000 00000000
91d8ff38 00000001 00000000 fe80: 00000000 00000040 91d8feb0 90042a9c
91d8ff40 900e1ea8 91d0bc80 0007f03c fea0: 91d8feb0 00000000 91d8ff88
91d8e000 0007ffdc 900c5f8e 00000000 900c5d36 fec0: 91d8fee0 900c410c
00000001 0007f03c 00000001 914fd800 91e5fb40 914fd800 fee0: 900c3fa6
91d8ff0c 91c6ea80 91d0cc00 0007f03c 91d8e000 00000000 91c6ea80 ff00:
900162a8 91d0cd20 91d0cd20 900c2184 91d8ff40 00000001 91d0cfa8 0007f03c
ff20: 00000400 91d0cc00 00000001 00000000 0007f03c 91e5fb40 900c3dc6
91d0cc0c ff40: 90042eba 91d8ff64 91e5fb40 0007f03c 0007f03c 91d8ff88
00000000 0007f03c ff60: 91d8e000 90043126 91d8ff94 91e5fb40 00000001
0007f03c 00000000 00000000 ff80: 0007f03c 91d8e000 00000000 00000000
00000000 90010132 00000000 2aafb074 ffa0: 00000001 0007f03c 00000005
0007f03d 80000000 91d8e000 00000001 2aaf616a

---------------------------------------------------------------------

This is resolved by sticking in a call to barrier() as follows:

--- drivers/char/n_tty.c        2008-08-20 14:25:02.000000000 -0400 +++
drivers/char/n_tty-fixed.c  2008-08-20 14:58:01.000000000 -0400 @@ -929,9
+929,10 @@ static void n_tty_receive_buf(struct tty
               spin_unlock_irqrestore(&tty->read_lock, cpuflags);
       } else {
               for (i=count, p = cp, f = fp; i; i--, p++) {
-                 if (f)
+                 if (f) {
+                   barrier();
                   flags = *f++;
-
+                 }
                       switch (flags) {
                       case TTY_NORMAL:
                               n_tty_receive_char(tty, *p);


.....or use an older avr32-linux-gcc.



Jesse Malone

-------------------------------------------

The original thread is here: http://www.avr32linux.org/archives/kernel/2008-August/001467.html

posted by Jesse at 3:29 AM #




I'm here to offer a possible solution to a common and likely frustrating problem. Here it is: your wireless connection keeps on saying "limited or no connectivity" and when you look closer it turns out that it has been unable to find an IP address to use. Well, it could be that the clock in your wireless router is not set properly, most likely in the past; possibly by several years. When your computer requests an address from the router it gets one, but the expiry time for that address is in the past. Your computer, obedient thing that it is, doesn't use that address since it expired long ago. Now, I can't go into detail on how to set the time for your particular router, but at least now you should know what you're looking for.
posted by Jesse at 2:35 AM #


Sunday, February 17, 2008

Firefox Cursor Bug

There is a bug in firefox that prevents it from showing a blinking cursor in a text box under certain conditions. It happens when you have a text input element overlapping another element that is scrollable. The fix that worked for me was to put a <div>> around the text input with the property overflow:auto. There are other possible solutions as well, some of which involve cumbersom javascript.

Labels: bug, css, design, Firefox, javascript, programming, web

posted by Jesse at 8:03 PM #


Thursday, February 07, 2008

This space has been much neglected for a while now. I've pretty much resigned myself to the fact that I just can't keep a personal blog going for any length of time, so I have decided to change the mandate of this blog. As part of my job as an IT type I frequently run accross problems that have me scouring the net for others who have encountered the same one and have either solved it or can provide some insight to point me in the right direction. In this regard I am very much a taker and not a giver and feel indebted to the people who sit around and post solutions on message boards rather than visiting a given board once, finding a solution and never returning (as I do). To rectify this situation I've decided that whenever I come accross a problem that others may also encounter I should detail the problem and the solution here, be it the product of a google search or my own ingenuity. So, without further ado, the first installment...

Creating an image of a damaged hard drive

I recently had to replace a hard drive that was nearing death. S.M.A.R.T. caught the trouble early, so the drive was still functional but with large swaths of unreadable sectors (large as in about 1GB). I surmise that this was largely occupied by free space because nothing critical seems to be missing from the copy I made of the drive. In any case...on with the show.

When I need to image a drive my first step is usually G4u(www.feyrer.de/g4u/), which is itself an indispensable tool. In the case where G4u doesn't boot on the given hardware I reach for my trusty knoppix live cd(http://www.konpper.net) and use dd to copy an image of the drive either to a second drive of to some nfs share. (Depending on the filesystem of the machine you are writing the image to you will have problems with the maximum file size).

In this particular case G4u chugged along for a while until it reached an unreadable block at which point it gave up. At this point I started looking for alternatives. I had heard of dd_rescue, which is a modified dd that doesn't quit when it hits bad blocks, it just skips ahead and allows you to tell it where to try next. I had tried it in the past and knew that it takes an ungodly amount of time to get through rough patches, so I started searching around the net for other options. Well, I hit gold: a little script called dd_rhelp. dd_rhelp is just a shell script that directs dd_rescue in a very clever manor, skipping from the start of a patch of bad blocks to some point past it, working backwards until it reaches the end of that patch and then continuing. Once it reaches the end of the disk or partition it will return to sections it has skipped looking for readable sections. I used it on the failing drive (copying to a new drive, not a network share because of file size limits) and it worked like a charm.

You can find dd_rescue and dd_rhelp in this archive. Before you can use dd_rhelp you need to compile dd_rescue and copy the binary to /usr/local/bin (or somewhere in your PATH). Go the the dd_rescue-1.59 directory and do a 'make' and then copy dd_rescue to an appropriate place. Read the documentation in the dd_rhelp directory and be VERY CAREFUL to get the parameters in the right order to avoid writing to the partition you want to read from.

Usage:

If you're running dd_rhelp from a ramdisk, as is the case if you are using a linux live cd then dd_rhelp will complain that it can't find a suitable location for its log file, which it uses to keep track of what parts of the disk it has skipped over. If you have another hard drive available (ie. not the drive you are copying, or the drive you are copying to) you can mount that as /tmp. So, if your extra drive is /dev/hdc1 you would do:

mount /dev/hdc1 /tmp

If you don't have a spare drive then you'll need to use a network drive. If you're on a windows network here's how you do it. If you're not on a windows network the odds are good that you know what you're doing.

mount -t smbfs //*computer-name*/*share-name* /tmp -o "username=*windows-domain-name*/*user*"

(items with '*' around them should be replaced with acutal values)

Now dd_rhelp should run without issue. I'd recommend reading the dd_rhelp documentation to get to know it before doing anything. To get dd_rhelp started you need two parameters: the source device/file and the output device/file. You have to be ABSOLUTELY CERTAIN which is which. To copy /dev/hda to /dev/hda execute this from the dd_rescue-1.59 directory:

./dd_rhelp /dev/hda /dev/hdb

Everything on /dev/hdb will be CLOBBERED(To be exact, the partition table will be overwritten) the instant you run this command so be ABSOLUTELY SURE that you have the parameters right. I usually check by mounting a partition on what I think is the source drive and making sure the contents and size are correct. You can check the size of mounted partitions using the df command. If you mount /dev/hda1(the first partition on /dev/hda) to some location , look for that location in the list df returns.

Labels: backup, hard drive, hardware, image, recovery, storage

posted by Jesse at 12:46 AM #


Sunday, September 10, 2006

I'm looking for a flugel horn at the moment. For now I'm renting a Bach Strad 183. I will be trying out a Yamaha 631G soon. I've been playing an old 631 and like it better than any other horns I've tried, and I'm hoping the new 631G hasn't changed much. In anycase I was browsing for flugel horns on ebay out of curiosity and found this amusing advertisement. For starters, its advertised as a flugel horn, but from the pictures is clearly a cornet.

You are seeing a good item now, the product is do very good, you can see some beautiful designs of the appearance of the product. They are vivid and lovely. The product is very beautiful, there are many very good designs on its body, very astonishing, each detail is make a high technique of inhale the craftsman and are of intelligence. There is the costliness value of the collections, if you have a collection house of Chinese handicraft product, please give more attention, if you want to own item, you will see get a more detail concerning item, please don't blunder away an opportunity!?

posted by Jesse at 1:12 AM #


Saturday, August 12, 2006

So Pluto is a planet after all

"If they have the power to make or unmake planets, why can't they do something really useful with Pluto, such as decree it to be a really huge scoop of chocolate cookie-dough ice cream?"

..."So, you're saying that anything that's only 5.5 times smaller than a planet should also be considered a planet?

Earth/5.5 = 2300km => Pluto is a planet

Pluto/5.5 = 420km => Vesta asteroid (450km) is a planet

Vesta/5.5 = 82km => 61 Danaë asteroid (82km) is a planet

61 Danaë/5.5 = 15km => 2685 Masursky asteroid is a planet

2685 Masursky/5.5 = 2.5km => 2002 JF56 asteroid is a planet

2002 JF56/5.5 = 455 meters => CN Tower (550m) is a planet

CN Tower/5.5 = 100m => Eyeglass orbital telescope would be a planet

Eyeglass/5.5 = 18m => The Titan Missle silo was a planet

Titan/5.5 = 10.7 feet => Your bathroom is a planet

Bathroom/5.5 = 2 feet => Your Mom is a planet! Oh, snap!"

....."Your Mom/5.5 = 4.36 inches => Oh yeah? Well my.... nevermind."

posted by Jesse at 12:21 AM #


Thursday, March 16, 2006

F*@# Internet Explorer, It's Time for CSS

This site is well overdue for a redesign. The cobwebs are everywhere, in the form of <table>'s. It's about time I stopped using tables for layout. My latest project, a website for Golfi Rustproofing, marks the end of tables based design for me. Starting with the Golfi website, I am abandoning tables. Naturally, Internet Explorer mangles the CSS layout that works like a charm in every other browser. I'll deal with it, as everone must, and find workarounds to accomodate it. What else can one do? Internet Explorer is a terrible browser that nevertheless, most people use.

In other news, this site has been moved onto a new server. After serving reliably for over two years, the 9 year old P166 that hosted this website was showing signs of possible impending hardware failure. I have replaced the old workhorse with a somewhat less old PIII 550MHz 128MB.
posted by Jesse at 7:36 PM #