| 00:03.20 | brailsmt | so we will be XORing the bits 32, 26, 23, 22, 16, 12, 11, 10, 8, 7, 5, 4, 2, 1 and 0 for crc-32 |
| 00:03.45 | brailsmt | s/32// |
| 00:04.10 | brailsmt | the 32nd bit being the bit that is XORed in the other positions... |
| 00:05.00 | brailsmt | han_solo: did he say in class to actually use a shift register? or just to use isBitSet function? |
| 00:05.22 | han_solo | i don't really know what he was suggesting |
| 00:05.36 | han_solo | I always turn off when profs start going into how to do it |
| 00:05.47 | han_solo | they usually try to dumb it down, and it's harder for me to grasp than the 'real way' |
| 00:05.57 | brailsmt | me too... i those cases i think i was playing nethack... :) |
| 00:06.00 | han_solo | so I just go by the book, or think it through myself |
| 00:06.06 | han_solo | saves me time |
| 00:09.07 | brailsmt | <PROTECTED> |
| 00:09.29 | brailsmt | <PROTECTED> |
| 00:09.29 | brailsmt | Mem: 191340 185432 5908 0 13948 80936 |
| 00:09.29 | brailsmt | -/+ buffers/cache: 90548 100792 |
| 00:09.29 | brailsmt | Swap: 488368 52068 436300 |
| 00:09.39 | brailsmt | i figured it out... for irssi at least :) |
| 00:09.54 | brailsmt | <PROTECTED> |
| 00:25.20 | Bradipo_ | Man, compiling KDE really chews up RAM in some points. |
| 00:25.36 | Bradipo_ | I've had to add two additional temporary swap files just to get to it not lockup my system. |
| 00:34.32 | brailsmt | han_solo: we crc the whole stuffed frame, excluding the sentinels? |
| 00:34.44 | brailsmt | or just the 'payload'? |
| 00:34.44 | han_solo | everything but the sentinels. the payload |
| 00:35.28 | brailsmt | k, thx for clarifying that |
| 00:49.19 | *** join/#utah Jayce^ (~jayce@65.204.214.20) |
| 01:03.22 | *** join/#utah fuzz (~whatever@cpe-66-87-27-164.ut.sprintbbd.net) |
| 01:52.57 | *** join/#utah fuzz (~whatever@cpe-66-87-27-164.ut.sprintbbd.net) |
| 03:16.36 | *** join/#utah brailsmt (~michael@12-254-226-160.client.attbi.com) |
| 03:17.17 | brailsmt | <PROTECTED> |
| 03:17.32 | brailsmt | installing new kernels ruins uptime... |
| 03:41.48 | han_solo | brailsmt: do we crc first, or bitstuff first? |
| 03:45.23 | brailsmt | crc, then stuff |
| 03:47.10 | brailsmt | i started hacking away on my files... then i realized that I still haven't passed the off ethernet lab... so I messed up my timestamps... luckily I hadn't run make yet, so my .o and executable files still had the right timestamps... |
| 03:48.02 | brailsmt | so I tar'd up the directory with --atim-preserve, and i guess, I 'll have to convince the TA's, that I really did finish thursday... |
| 03:48.13 | han_solo | btw, we do more than payload, we do the ethernet header too. not the sentinels though |
| 03:48.29 | brailsmt | yeah, everything but the sentinels... |
| 03:48.52 | brailsmt | I'm about ready to run my test to see that my crc function is reversible |
| 03:49.05 | brailsmt | tests, that is... |
| 03:49.21 | brailsmt | after I have repeatability down, I will go for correctness :) |
| 03:49.56 | brailsmt | btw, i think the pseudo-code on the webpage has an error... |
| 03:51.12 | brailsmt | it says to XOR the polynomial if the MSB of the CRC is 1 and if the MSB of the shift-register is 1 you recycle that one to the LSB... |
| 03:51.35 | brailsmt | i think, it should be XOR the polynomial if the MSB of the shift register is one |
| 03:51.54 | brailsmt | and wrap around the MSB of the CRC to LSB if it is a one... |
| 03:52.11 | brailsmt | lots of acronyms there :) |
| 03:52.49 | brailsmt | that sound right to you? |
| 04:04.01 | han_solo | shift register = CRC |
| 04:04.02 | han_solo | what you're calling the shift register is just the current byte of the data |
| 04:04.12 | han_solo | <PROTECTED> |
| 04:04.14 | han_solo | <PROTECTED> |
| 04:04.16 | han_solo | <PROTECTED> |
| 04:04.18 | han_solo | <PROTECTED> |
| 04:04.20 | han_solo | <PROTECTED> |
| 04:04.22 | han_solo | <PROTECTED> |
| 04:04.24 | han_solo | <PROTECTED> |
| 04:04.26 | han_solo | <PROTECTED> |
| 04:04.28 | han_solo | <PROTECTED> |
| 04:04.30 | han_solo | <PROTECTED> |
| 04:04.32 | han_solo | <PROTECTED> |
| 04:04.34 | han_solo | <PROTECTED> |
| 04:04.36 | han_solo | <PROTECTED> |
| 04:07.33 | brailsmt | right... |
| 04:10.14 | brailsmt | well, I am doing my crc as an int... |
| 04:11.01 | han_solo | unsigned I hope |
| 04:13.41 | brailsmt | i don't get the CRC |= msg >> 7... you are just ORing the msb of the msg byte to the CRC? |
| 04:15.46 | brailsmt | | 504 int idx = 0; |
| 04:15.46 | brailsmt | | 505 int cycle = 0; |
| 04:15.46 | brailsmt | | 506 for(; idx < num_frame_bits; idx++) { |
| 04:15.46 | brailsmt | | 507 if(crc & CRC_MSB_MASK) |
| 04:15.46 | brailsmt | | 508 cycle = 1; |
| 04:15.48 | brailsmt | | 509 else |
| 04:15.51 | brailsmt | | 510 cycle = 0; |
| 04:15.53 | brailsmt | | 511 |
| 04:15.56 | brailsmt | | 512 crc = crc << 1 |
| 04:15.59 | brailsmt | | 513 if(isBitSet(&frame[idx / BITS_PER_BYTE], idx % 8)) { |
| 04:16.01 | brailsmt | | 514 crc ^= CRC_POLY; |
| 04:16.04 | brailsmt | | 515 } |
| 04:16.07 | brailsmt | | 516 |
| 04:16.09 | brailsmt | | 517 crc ^= cycle; |
| 04:16.11 | brailsmt | | 518 } |
| 04:17.03 | han_solo | yes, and it is then XOR'ed with the lsb of CRC_POLY |
| 04:18.00 | han_solo | the LSB of the CRC is MSB of the unshifted CRC ^ next bit coming into the shift register |
| 04:18.05 | brailsmt | i thought the msb of the crc got wrapped around... and not the msb of the msg get or'd to the lsb of the crc... |
| 04:18.52 | brailsmt | oh, so both... |
| 04:19.00 | han_solo | book page 102, see the xor gate and its inputs |
| 04:22.45 | brailsmt | i think your |= should be ^= |
| 04:26.36 | han_solo | it doesn't matter since the lsb will always be 0 (after shifting). ^= is misleading though since we don't need to xor, it's just a convenient place to hold that bit until we get the chance to xor it with the CRC_POLY |
| 04:29.44 | han_solo | well, it passes my unit tests (extracted from the web site) and runs perfectly against itself and as the server against the ta code |
| 04:30.53 | han_solo | as usual, I can't figure out if the ta passoff is just screwed up in places or if it matters |
| 04:36.56 | brailsmt | good point as usual... |
| 06:56.51 | *** join/#utah Bradipo_ (bradipo@sicilia.bradfords.org) |
| 08:43.06 | *** join/#utah fozz (~fozz@12-255-4-85.client.attbi.com) |
| 14:36.17 | *** join/#utah levi__ (~levi@ash.lnxi.com) |
| 18:20.27 | *** join/#utah levi (~levi@dpc6682174249.direcpc.com) |
| 19:59.33 | *** join/#utah TimRiker (timr@rikers.org) |
| 19:59.33 | *** join/#utah neybar (~jalance@dsl233.sisna.com) [NETSPLIT VICTIM] |
| 19:59.33 | *** join/#utah levi_ (~levi@dpc6682174249.direcpc.com) [NETSPLIT VICTIM] |
| 19:59.34 | *** join/#utah Jayce^ (~jayce@65.204.214.20) [NETSPLIT VICTIM] |
| 19:59.34 | *** mode/#Utah [+o TimRiker] by calvino.freenode.net |