From: Timothy Miller <miller@techsource.com>
To: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Compressed kernel messages from an actual extraction from kernel source
Date: Tue, 15 Apr 2003 19:13:50 -0400 [thread overview]
Message-ID: <3E9C922E.9030108@techsource.com> (raw)
I decided it was time to see if I could compress REAL kernel messages,
extracted from real kernel source. There was too much junk in the
output of 'strings' from the image. I tried using 2.5.67, but I
couldn't get it to build. After finally giving up, I grabbed 2.4.20. I
hacked up Rules.make to produce .i files as well as .o files and wrote a
parser to scan all source files looking for valid printk statements.
After dealing with some issues with __FUNCTION__, etc., I was able to
get a good list of all strings found in a build of 2.4.20 with all
default parameters (I ran 'make config' and held down Enter). If
someone would like to help me build a better list, please email me.
As you'll see, the actual number of kernel messages only come to about
128280 bytes, which is much less than my previous experiment, but it's
also, by proportion, more compressible. That shrinks to 59849 bytes.
That's 53% smaller.
My next step after this is to write a filter which locates format
strings a little more intelligently than the extractor, replaces them
properly, and readies them for compilation. I'm going to need perhaps a
bit more help with the Makefiles. One thing I tried was making %.o
depend on %.i, hoping it would use the existing %.i rule, but it didn't.
One way to do the filtering is to have .q (or whatever) files depend on
.i files which depend on .c files, and .o files depend on .q, where .q
files have the strings replaced. Oh, and, of course, I have to rewrite
printk() which I've not even bothered to LOOK at yet. :)
If you're wondering why I'm expending so much energy on shrinking the
kernel memory usage by a mere 68431 bytes, it's because I'm a geek, and
I'm having a blast. Isn't the open source movement great? :) Now,
it's time to go home to see my wife. Yes, I have one of those. :)
Here's the code table. Codes are in octal.
code shrink count word
200 5160 179 [ OUT_RING( 0X%08X ) AT 0X%X]
201 3602 902 [>[DRM]
202 2224 2227 [ %]
203 1638 1641 [: ]
204 1506 755 [ING]
205 1463 1466 [D ]
206 1399 468 [<7\201:]
(<7>[DRM:)
207 1390 175 [] *ERROR*]
210 1264 423 [LOCK]
211 1188 71 [KERNEL\203ASSERTION (]
(KERNEL: ASSERTION ()
212 1164 1167 [RE]
213 1107 1110 [T ]
214 1051 352 [ ]
215 1001 1004 [ER]
216 926 33 [ ADVANCE_R\204() WR=0X%06X TAIL=0]
( ADVANCE_RING() WR=0X%06X TAIL=0)
217 851 854 [E ]
220 814 205 [FAILE]
221 772 259 [<6\201]]
(<6>[DRM])
222 745 84 []\214OUT_R\204\202X]
(] OUT_RING %X)
223 742 745 [IN]
224 664 334 [<3>]
225 655 658 [S ]
226 650 653 [OR]
227 636 23 [7>\212SET_XMIT_TIM\215 SK=%P\202\205WHEN=0]
(7>RESET_XMIT_TIMER SK=%P %D WHEN=0)
230 614 617 [AL]
231 613 42 [\210\203FH NO\213V\215IFIED!]
(LOCK: FH NOT VERIFIED!)
232 598 76 [FLOPPY%D\203]
(FLOPPY%D: )
233 578 21 [COPY\203COPY\204\202S/%S, \230\212ADY V\215IFIED]
(COPY: COPYING %S/%S, ALREADY VERIFIED)
234 577 580 [\202D]
( %D)
235 557 560 [AT]
236 544 547 [ON]
237 520 175 [<3\201:]
(<3>[DRM:)
240 510 33 [\221 BEG\223_R\204(\202\205) \223\202S]
(<6>[DRM] BEGIN_RING( %D ) IN %S)
241 504 23 [<7>BUG\203UNKNOWN TIM\215 V\230UE]
(<7>BUG: UNKNOWN TIMER VALUE)
242 502 253 [<6>]
243 492 248 [0X%]
244 489 492 [DE]
245 478 481 [TE]
246 470 60 [UNABL\217TO ]
(UNABLE TO )
247 450 227 [NFS]
250 448 451 [%D]
251 438 21 [<4>FH_\210:\202S/%\225\230\212ADY \210ED!]
(<4>FH_LOCK: %S/%S ALREADY LOCKED!)
252 436 439 [%S]
253 423 426 [, ]
254 421 424 [LE]
255 419 422 [\204 ]
(ING )
256 416 210 [RPC]
257 409 412 [ST]
260 398 401 [OU]
261 379 382 [SE]
262 376 190 [NO\213]
(NOT )
263 352 355 [AN]
264 349 118 [\244VIC]
(DEVIC)
265 348 71 [) \220\205A\213]
() FAILED AT )
266 346 13 [:\252\207 EXCESS
F\212ES:\202\205F\212ES,\202\205\230LOCS]
(:%S] *ERROR* EXCESS FREES: %D FREES, %D ALLOCS)
267 346 175 [<4>]
270 342 87 [.C(\250)]
(.C(%D))
271 329 332 [EN]
272 328 23 [C\230\254\205WITH\260\213\210 HELD]
(CALLED WITHOUT LOCK HELD)
273 325 110 [08LX]
274 317 320 [RO]
275 302 305 [] ]
276 298 76 [BUFF\215]
(BUFFER)
277 298 21 [\247D\203FH_\210(\252) \210E\205=\234]
(NFSD: FH_LOCK(%S) LOCKED = %D)
300 287 290 [:\202]
(: %)
301 284 23 [<\227X%LX\253C\230L\215=%P]
(<7>RESET_XMIT_TIMER SK=%P %D WHEN=0X%LX, CALLER=%P)
302 284 144 [TO ]
303 266 68 [APIC ]
304 264 267 [ C]
305 258 131 [<7>]
306 250 127 [02X]
307 242 245 [AR]
310 241 244 [TI]
311 238 81 [SCSI]
312 235 238 [AD]
313 235 80 [\230LOC]
(ALLOC)
314 228 231 [IS]
315 227 230 [UN]
316 227 230 [\202\205]
( %D )
317 223 226 [CH]
320 223 16 [RTNL\203ASS\215\310\236 \220\205A\213]
(RTNL: ASSERTION FAILED AT )
321 222 113 [...]
322 222 113 [PCI]
323 219 222 [US]
324 219 14 [\275BEG\223_LP_R\204(\250) \223\202S]
(] BEGIN_LP_RING(%D) IN %S)
325 218 45 [P\274CESS]
(PROCESS)
326 206 209 [ ]
327 206 209 [Y ]
330 206 9 [\257\307\213= \243LX\253\271\205=
\243LX\253OFF\261\213= \243LX]
(START = 0X%LX, END = 0X%LX, OFFSET = 0X%LX)
331 202 103 [\210D\203]
(LOCKD: )
332 199 68 [\224FH_]
(<3>FH_)
333 197 200 [MA]
334 190 193 [=%]
335 190 97 [\215R\226]
(ERROR)
336 190 33 [\221\216X%06X]
(<6>[DRM] ADVANCE_RING() WR=0X%06X TAIL=0X%06X)
337 188 191 [\202S]
( %S)
340 187 190 [TH]
341 186 48 [SOCKE]
342 184 32 [\224HUB.C\203]
(<3>HUB.C: )
343 182 185 [RI]
344 180 183 [P ]
345 179 182 [NO]
346 179 182 [\252\203]
(%S: )
347 178 91 [QUE]
350 178 19 [\224\223IT_MODU\254\203]
(<3>INIT_MODULE: )
351 178 61 [\256\3004\205]
(RPC: %4D )
352 176 179 [\221\200]
(<6>[DRM] OUT_RING( 0X%08X ) AT 0X%X)
353 173 176 [LD]
354 172 175 [AG]
355 172 30 [UHCI.C\203]
(UHCI.C: )
356 168 35 [EXT2-F]
357 166 169 [AB]
360 166 169 [NE]
361 165 168 [ME]
362 165 168 [OF]
363 163 166 [ F]
364 163 34 [\223T\215RUP]
(INTERRUP)
365 160 82 [IRQ]
366 160 55 [NULL]
367 158 161 [\215 ]
(ER )
370 157 160 [\260N]
(OUN)
371 154 79 [08X]
372 152 155 [AC]
373 150 153 [AS]
374 150 20 [SYM53C8XX]
375 148 76 [SVC]
376 148 151 [TR]
377 148 31 [\235\245MP\213\302]
(ATTEMPT TO )
001 143 146 [ B]
002 142 17 [U.%U.%U.%U]
003 142 9 [\223C\226REC\213\261GM\271\213C\370\213A\213\243P]
(INCORRECT SEGMENT COUNT AT 0X%P)
004 142 37 [\223V\230I\205]
(INVALID )
005 142 37 [\312D\212SS]
(ADDRESS)
006 141 14 [\275\312V\263CE_LP_R\204]
(] ADVANCE_LP_RING)
007 139 142 [SI]
010 138 36 [SUPP\226]
(SUPPOR)
013 132 68 [C\230L]
(CALL)
014 132 68 [\247D\203]
(NFSD: )
015 128 131 [IO]
016 128 131 [LO]
017 127 130 [ \212]
( RE)
020 124 64 [TCP]
021 124 10 [\211\257\307\213<= \362F\261T+\254N\265]
(KERNEL: ASSERTION (START <= OFFSET+LEN) FAILED AT )
022 122 125 [->]
023 122 63 [\235I\236]
(ATION)
024 121 124 [=\250]
(=%D)
025 120 123 [ID]
026 118 25 [\314APNP\203]
(ISAPNP: )
027 116 60 [BA\205]
(BAD )
030 113 116 [E\205]
(ED )
031 112 115 [C\236]
(CON)
032 112 58 [S\215V]
(SERV)
033 112 115 [X ]
034 112 115 [\214 ]
( )
035 112 115 [\257\235]
(STAT)
036 111 114 [EX]
037 110 113 [(%]
177 110 29 [\315K\345WN]
(UNKNOWN)
141 109 112 [EC]
142 108 6 [\275PI\205=\234\253\264\217=
\243X\253OP\271_C\370\213=\234]
(] PID = %D, DEVICE = 0X%X, OPEN_COUNT = %D)
143 106 55 [MOD]
144 104 54 [<5>]
145 104 107 [MP]
146 104 107 [\226 ]
(OR )
147 103 22 [C\260\353N'\213]
(COULDN'T )
150 103 22 [\223I\310\230IZ]
(INITIALIZ)
151 102 105 [ED]
152 102 105 [_\212]
(_RE)
153 102 9 [\207\304\263\345\213C\212\235\217/P\274C/]
(] *ERROR* CANNOT CREATE /PROC/)
154 100 103 [CO]
155 100 103 [\207 ]
(] *ERROR* )
156 100 35 [\244\245C\245]
(DETECTE)
157 99 102 [\202P]
( %P)
160 98 101 [V\215]
(VER)
161 98 6 [\346\025\217\031T\274LL\367\236 \322\001\323\202\306
\244V\202\306]
(%S: IDE CONTROLLER ON PCI BUS %02X DEV %02X)
162 96 99 [ P]
163 96 99 [IC]
164 94 17 [ES1371\203]
(ES1371: )
165 94 97 [L\217]
(LE )
166 94 33 [SWAP]
167 94 49 [\361M\226]
(MEMOR)
170 93 96 [PU]
171 92 48 [04X]
172 92 95 [IT]
original size=128280, new size=59849
reply other threads:[~2003-04-15 22:47 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3E9C922E.9030108@techsource.com \
--to=miller@techsource.com \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox