March 2001 - Packet Writing - openSUSE Mailing Lists

Patch 0.0.2f test results
by Cory 24 Mar '01

24 Mar '01

OK, I've thoroughly tested this patch, and have come to the following conclusions: 1) After applying the fix to the patch, it compiles and installs into the kernel perfectly. 2) Using the most recent UDF tools from CVS, I'm able to format and setup a CDRW disc without problem. 3) Mounting and unmounting the disc work perfectly in almost all cases (no data written) 4) Writing to the drive works to a point. It will allow me to write any sized file I want without hitch, but it doesn't seem to like when I copy a large amount of files on at once. For example, if I try to copy the Linux source tree onto it, it'll get about halfway done, then crash, leaving the FS corrupt. 5) Reading works fine as well, but I'm concerned it isn't running as fast as it could. I've got a 8x8x32 writer, and the packet driver claims it reads/writes at 6x/4x. Is this right? 6) For some strange reason, with the packet driver loaded it slows my entire kernel down. IDE access to all drives is most prominently affected, resulting in less than half my normal access speeds in most/all cases. Is this normal? I even tried back porting the driver to 2.4.1, a kernel I know ran just fine. With the driver in, I slowed to a crawl. 7) It seems to work best with SCSI emulation enabled. During the course of my writing to the disc, sometimes I get errors: __refile_buffer: rdev is packet, should move? bh timed out: sector 16556 dev/rdev 61:00/61:00, count 4, state 1c, end_io c01bb91c, private 00000000, list 0 Are these errors normal? I've also noticed I get a lot of I/O errors when writing to discs. I know they have a few small scratches here and there, but is it really enough to through the drive out? It eventually does write correctly, but only after about 2-3 minutes of accessing. My main concern about this patch is the general slowdown that occurs when the driver is loaded. My harddrive shouldn't be running at 5.04mb/sec sequential throughput. Anyway, great job with this patch. Despite these problems, I'm starting to use it regularly. My system specs: ---------------- AMD K6-2 500mHz processor running on a Tyan Trinity ATX motherboard VIA MVP2 vt82c686a IDE chipset Samsung 30gb ATA-66 harddrive Ricoh RW7083A 8x8x32 IDE CD-RW drive w/ SCSI emulation enabled hda = 30gig harddrive hdb = absent hdc = CD-RW drive hdd = 56x CD-ROM ide0 = 80 wire cable ide1 = 40 wire cable ----------------- That's my 2 cents. Cory (ssaver(a)crosswinds.net)

3 3

Where to get that patch?
by Johnathan Hicks 24 Mar '01

24 Mar '01

So where can I get that patch? If I knew of a list archive I'd search it, but I don't. -John

2 1

New patch release soon?
by Johnathan Hicks 24 Mar '01

24 Mar '01

Hi I was wondering if there will be a new patch release anytime soon. I did a "patch --dry-run" with the current release of the patch on a 2.4.3-pre6 kernel tree and got more an 60 lines or errors, and only a few successes. And yes I did invoke "patch" correctly... --John

2 1

.
by lk89lk99＠yahoo.com 22 Mar '01

22 Mar '01

ADD 2 - 4 INCHES, WHERE YOU NEED AND WANT THEM!! Want to Enlarge your Penis? Most men do. Add up to four inches. Make your penis harder, longer and thicker. Dont be embarrassed by the size of penis. Dont leave your lover unfulfilled. Works for all ages. You can rejuvenate your penis. Three out of four women say size DOES matter. Discover the very latest, proven techniques. Get answers and solutions that really work. Stop being ashamed and get into action! Based on doctor interviews as well as consultation with the leading industry porn stars. Results are guaranteed! Mailed in plain wrapper so no one will know but you, but we recommend that you receive our packet by email to your own personal email address. Once your order is filled we delete your address. Indicate below if you prefer delivery by private email. ORDER NOW!!! This important information is just $29.95 [shipping and handling included if mailed]. We accept checks, money orders, even cash. Print the following form and mail it to: WA Consultants 3104 E. Camelback Road, #754 Phoenix, AZ 85016 Ship to: _________________________________________________________ Shipping Address: _________________________________________________ City: ___________________________________________________________ State: ___________________________________________________________ Zip Code: ________________________________________________________ Yes, deliver to me by email at: _______________________________________________________ Make your check payable to: WA Consultants All orders ship within 24 hours. @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ This mailing is done by an independent marketing co. We apologize if this message has reached you in error. Save the Planet, Save the Trees! Advertise via E mail. No wasted paper! Delete with one simple keystroke! Less refuse in our Dumps! This is the new way of the new millennium To be removed please email g14toy(a)yahoo.com with the word "remove" in the subject line. @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

1 0

packet-writing for 2.4.x, take 2
by Ben Fennema 22 Mar '01

22 Mar '01

Ok, so take 2 of packet-writing for 2.4.2 =) I've made some actual changes this time now that I understand what the heck is going on (or trying to go on) - at least a little better =) The old code merged everything in the same packet in the elevator. Now, I merge requests the are in the same packet in the request function. (and only if there are no requests outside the packet occuring before) Hopefully the elevator will stick requests in the same packet next to each other (I havn't done extensive testing to make sure this is usually the case) I also made it so the blocks that have to be read in get a chance to be merged by submitting all the read requests, then waiting for them to complete insteading of submitting them one at a time, and waiting for each to complete before submitting the next. I also allocated the bh's data in 1 call so that hopefully sequential bh's would occur in the same segment. Of course, I didn't test to make sure the bh's were allocated in the same order, so theres a chance their allocation is backwards =) Anyways, this means pktcdvd tries to allocate 64k w/ kmalloc, which might fail after running for a while, or on a machine with very little memory. Anyways, let me know how/if it works for you =) Ben Patch attached -- Linux UDF - http://linux-udf.sourceforge.net Latest Is - udf-0.9.3 (http://www.csc.calpoly.edu/~bfennema/udf.html)

2 1

Compilation error
by Adam Huffman 22 Mar '01

22 Mar '01

Got the following compilation error with the latest diff: make[2]: Entering directory `/home/adam/Build/linux-marcel/drivers/block' gcc -D__KERNEL__ -I/home/adam/Build/linux-marcel/include -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -fno-strict-aliasing -pipe -mpreferred-stack-boundary=2 -march=i586 -DMODULE -DMODVERSIONS -include /home/adam/Build/linux-marcel/include/linux/modversions.h -c -o pktcdvd.o pktcdvd.c pktcdvd.c: In function `pkt_elevator_merge': pktcdvd.c:1714: warning: implicit declaration of function `bh_rq_in_between_R44b22191' pktcdvd.c: In function `pkt_new_dev': pktcdvd.c:1794: structure has no member named `bdev' pktcdvd.c: In function `pkt_media_change': pktcdvd.c:1907: warning: unused variable `cdi' pktcdvd.c: In function `pkt_ioctl': pktcdvd.c:1972: structure has no member named `bdev' make[2]: *** [pktcdvd.o] Error 1 make[2]: Leaving directory `/home/adam/Build/linux-marcel/drivers/block' make[1]: *** [_modsubdir_block] Error 2 make[1]: Leaving directory `/home/adam/Build/linux-marcel/drivers' make: *** [_mod_drivers] Error 2 The diff was applied without errors, although there were some offsets, probably because it is 2.4.2-pre4 (for various other reasons). Is anyone else seeing this? Adam

3 4

Results with new disc
by Adam Huffman 22 Mar '01

22 Mar '01

From what I remember, the sorts of messages I get here with a newly-formatted disc are similar to what I got with the last version released by Jens. Basically, copying small-ish files seems to work, but errors start appearing with larger files. I got these errors when starting to copy a ~250Mb tar file to the UDF disc. The log-file is attached, bzipped owing to size. Also attached is the dmesg output, which includes some other errors. Adam

1 0

Web Site Traffic Building
by sales＠ecastmarketing.com 17 Mar '01

17 Mar '01

We have ranked the top ten traffic building sites on the web, these sites have been ranked by our experts, our clients and by you our visitors. Come to www.greatfreesite.net/bestten One of the top ten sites, can even guarantee top placement on the top ten search engines. Our site has been developed to help you find the most cost-effective traffic on the web. Please feel free to contact us with any questions.

1 0

test with hp8100
by Mark Vytlacil 17 Mar '01

17 Mar '01

Hi, I tested the 2.4.2 driver with an hp8100 and new r/w disc with the following results: # cdrwtool -d /dev/hdb -q -t 2 using device /dev/hdb setting speed to 2 1015KB internal buffer setting write speed to 2x Settings for /dev/hdb: Fixed packets, size 32 Mode-2 disc I'm going to do a quick setup of /dev/hdb. The disc is going to be blanked and formatted with one big track. All data on the device will be lost!! Press CTRL-C to cancel now. Initiating quick disc blank Disc capacity is 275616 blocks (551232KB/538MB) Formatting track Writing UDF structures to disc -18000 (EST)/300 Quick setup complete! Mar 15 21:07:01 theDesk kernel: hdb: ATAPI 24X CD-ROM CD-R/RW drive, 1024kB Cache Mar 15 21:07:01 theDesk kernel: Uniform CD-ROM driver Revision: 3.12 Mar 15 21:07:01 theDesk kernel: hdd: ATAPI 48X CD-ROM drive, 128kB Cache Mar 15 21:07:01 theDesk kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error } Mar 15 21:07:01 theDesk kernel: hdb: packet command error: error=0x54 Mar 15 21:07:01 theDesk kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error } Mar 15 21:07:01 theDesk kernel: hdb: packet command error: error=0x54 Mar 15 21:07:01 theDesk kernel: hdb: packet command error: status=0x51 { DriveReady SeekComplete Error } Mar 15 21:07:01 theDesk kernel: hdb: packet command error: error=0x54 Mar 15 21:07:01 theDesk kernel: cdrom: This disc doesn't have any tracks I recognize! Mar 15 21:07:01 theDesk insmod: Note: /etc/modules.conf is more recent than /lib/modules/2.4.2/modules.dep # pktsetup /dev/pktcdvd0 /dev/hdb Mar 15 22:15:24 theDesk kernel: pktcdvd: v0.0.2e 10/03/2000 Jens Axboe (axboe(a)suse.de) Mar 15 22:15:24 theDesk insmod: Note: /etc/modules.conf is more recent than /lib/modules/2.4.2/modules.dep Mar 15 22:15:25 theDesk kernel: pktcdvd: writer hdb sucessfully registered Mar 15 22:15:25 theDesk kernel: pktcdvd: kernel thread pktcdvd0 started # mount /dev/pktcdvd0 /mnt Mar 15 22:15:57 theDesk kernel: pktcdvd: speed (R/W) 6/4 Mar 15 22:15:59 theDesk kernel: udf: registering filesystem Mar 15 22:15:59 theDesk insmod: Note: /etc/modules.conf is more recent than /lib/modules/2.4.2/modules.dep Mar 15 22:15:59 theDesk kernel: pktcdvd: inserted media is CD-RW Mar 15 22:15:59 theDesk kernel: pktcdvd: Fixed packets, 32 blocks / packet, Mode-2 disc Mar 15 22:15:59 theDesk kernel: pktcdvd: disabled write caching on pktcdvd0 Mar 15 22:16:00 theDesk kernel: pktcdvd: speed (R/W) 6/4 Mar 15 22:16:00 theDesk kernel: pktcdvd: 551232KB available on disc Mar 15 22:16:02 theDesk kernel: UDF-fs INFO UDF 0.9.1 (2000/02/29) Mounting volume '', timestamp 2001/03/15 21:54 (1ed4) # ls /mnt lost+found At this point I tried to revert an emacs buffer with the kernel log. Emacs responded but the system then froze. No further kernel messages were available on reboot. I don't have a serial link set up but I thought I would post these results since, unlike others, I hadn't even written to the file system yet. Note the speeds which don't match my drive which is 2x. Maybe I should try to set up a serial link if I am going to try to work on kernel modules. Mark Vytlacil -- There are trivial truths and the great truths. The opposite of a trivial truth is plainly false. The opposite of a great truth is also true. --Niels Bohr

2 2

Hit the Lotto+Make $ Guaranteed #741B
by Luis E. Milen 14 Mar '01

14 Mar '01

You are GUARANTEED to Win a National Lottery or theyll pay you $10,000. Thats because as a player/agent, you are participating in the largest pool of Lotto players on the planet!The average player/ agent is winning about once every 2 - 3 weeks! The average odds to win a US Lottery are 13 Million to one! According to this Global companies National Better Business Bureau, the odds of winning their lottery are 3500 to one! This opportunity has just hit the US! You are one of the first to hear about it. We are looking for a limited number of entrepreneurs who want to play, win and earn some very large commissions by owning their own Lotto Agency online! Refer other players and earn $3,500-$50,000 monthly in commissions! This offer will only be made once. Act NOW!! Please reply to: mailto:33mtk@verizonmail.com?subject=info with your: email address Name Phone number Best time to reach you and we will contact you within 24 hours with additional information!! Reply Now!!! ///////////////////////////////////////////////////////// PLEASE REMOVE AT: mailto:wxpp@whomail.net?subject=remove /////////////////////////////////////////////////////////

1 0