Files
@ 155d653e278f
Branch filter:
Location: Copyleft/guide/presentations/ccs-report-examples/ccs-examples.md
155d653e278f
17.9 KiB
text/markdown
Continued editing for minor errors, clarity, until I.2.
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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 | % Examples of CCS Reports
% Bradley M. Kuhn & Karen M. Sandler
% Tuesday 9 May 2017
# CCS
Complete, Corresponding Source
# How GPLv3 says CCS.
<hr/>
> The "Corresponding Source" for a work in object code form means all the source code needed to generate, install, and (for an executable work) run the object code and to modify the work, including scripts to control those activities.
<span class="fitonslide">
<p align=right>
— GPLv3§1
</p>
</span>
# How GPLv2 says CCS.
<hr/>
> You may copy and distribute the Program (or a work based on it, under
> § 2) in object code or executable form under the terms of § 1
> & 2 above provided that you … [a]ccompany it with the complete
> corresponding machine-readable source code … The source code for a
> work means the preferred form of the work for making modifications to it.
> For an executable work, complete source code means all the source code for
> all modules it contains, plus any associated interface definition files,
> plus the scripts used to control compilation and installation of the
> executable.
<p align=right>
— GPLv2§3
</p>
</span>
# The 11 Words That Consumed My Life
+ GPLv2 enforcement, for embedded products, is all about the these eleven
words.
+ I could give an entire talk on any one of these 11 words.
+ Yes, I can even give 20-30 minute treatises on each use of “the”.
+ Yet, when enforcement processes are at their best, they're about the spirit
behind these words, not the words themselves.
<hr>
> the scripts used to control compilation and installation of the executable.
<p align=right>
— GPLv2§3
</p>
</span>
# The 11 Words That Consumed My Life
+ Basic reference rule:
+ Can a developer reasonably skilled in the art of embedded software
build your sources, take the (copylefted) executables and install
them?
+ Enforcement spends its most attention on testing CSS
“candidates” to verify that.
<hr>
> the scripts used to control compilation and installation of the executable.
<p align=right>
— GPLv2§3
</p>
</span>
# CCS "Round" Reports
+ Evaluate each CCS candidate.
+ Someone "skilled in the art" attempts to build.
+ Detailed notes are sent, asking for new CCS candidate "round".
+ Some anonyomized real world examples.
# No Build Instructions
<hr/>
The primary issues we found were a dearth of build instructions as well
as a complete lack of installation instructions. There was no
information that mentioned how one might build each package so we had to
guess about which Makefile and/or build script to run for each package.
And in many cases it was not possible or straight-forward to build - this
must be resolved in the next source candidate.
# Making General Recommendations
<hr/>
We generally recommend that the source release be a single file (ie. one
tarball containing all packages required for the build) that includes a
README or similar in the main directory explaining exactly how to build
and install all of the packages. See section 21.2 of
http://compliance.guide/pristine-example for an excellent example.
# Suspicious, But Not Captious.
<hr/>
Also, we noticed that some packages mentioned in the "OPEN SOURCE
SOFTWARE NOTICE" included with the device (and also downloaded as part of
the source release; see
Open_Source_Software_Notice_and_Privacy_Policy.pdf ) could not be found
in the source release. For example, we found "Software: Samba 3.0.XX" in
the notice, but could not find any trace of Samba in the source release.
Please ensure that all the software included in the notice is included in
the source release as well.
# Nesting Doll Packages
<hr/>
Once extracted, the 3 .rar files above produce the following output
files:
* busybox-1.21.1.rar
* AB_A0101.123.tar.gz
* source.rar
* a small text file that gives two-word descriptions of the above files
# Actual(ly Trying to) Build
<hr/>
This file included no instructions for how one might build it so we
tried to run "make" but received the following error:
$ make
.../busybox-1.21.1/scripts/gcc-version.sh: line 11:
arm-none-linux-gnueabi-gcc: command not found
# Toolchain?
+ The toolchain is rarely considered mandatory as part of “the
scripts”.
+ Admittedly, it doesn't *control* compilation, it *is* compilation.
+ The script here is explaining precisely what type of toolchain is needed.
+ Something like: “GCC vX built with the following ./configure
line” is usually adequate.
+ But including the toolchain is a nice step to make it easy for your users.
<hr>
> the scripts used to **control compilation** and installation of the executable.
<p align=right>
— GPLv2§3
</p>
</span>
# We Guess at Compiler Anyway
<hr/>
So we searched for an arm-none-linux-gnueabi- cross-compiler in the
other files but could not find one. We then tried to use our own (be
editing the PATH appropriately), which did get us past this error. Note
that this is not acceptable in a source release - the cross-compiler
that a user must use needs to be clearly indicated (name, version, etc.)
and/or included with the source release.
# Feedback on Small Problems
<hr/>
Once we had the custom cross-compiler configured, we then ran into these
errors:
$ make
.../busybox-1.21.1/scripts/gen_build_files.sh: Permission denied
make: *** [gen_build_files] Error 127
$ make
.../busybox-1.21.1/scripts/mkconfigs: Permission denied
make: *** [include/config/MARKER] Error 126
$ make
/bin/sh: applets/usage_compressed: Permission denied
make[1]: *** [include/usage_compressed.h] Error 126
make: *** [applets_dir] Error 2
$ make
.../busybox-1.21.1/scripts/trylink: Permission denied
make: *** [busybox_unstripped] Error 126
In each case, we found the mentioned file and then added executable
permissions to it (ie. "chmod u+x scripts/gen_build_files.sh"). This
must be fixed in the next source release - please set the executable
bits on the above files appropriately in the archive file you
distribute.
# Install Instructions missing
<hr/>
After fixing the above, a "busybox" binary was generated. However,
there were no instructions to indicate how one might install this binary
on the device. Such instructions are required by GPLv2, under which
BusyBox is licensed. Please include the instructions in your next
source release.
# Build "Only Seems" To build
<hr/>
For the AAB_A0101.123.tar.gz package, we ran "./build.sh", the build
took about 140 seconds, which is less than one would expect for building
all of the programs listed in the "OPEN SOURCE SOFTWARE NOTICE". The
only files we could immediately find that were clearly the result of
this "./build.sh" invocation were some kernel image binaries, found in
path/path/path/path/path/KERNEL_OBJ . This path was not mentioned at
all and we had to guess at where they might be.
# Maybe Proprietary Kernel Modules?
<hr/>
Furthermore, there were no .ko files generated, which is abnormal for a
build of the kernel, Linux. Please ensure that all .ko files which are
used on the system are generated with "./build.sh" or a similar script.
# Weird versioning
<hr/>
* The following libraries have different versions in the firmware than
is built from the candidate CCS. Specifically, your candidate CCS
contains version "1800", and the firmware has version "2400". Since
most of these libraries are licensed under the LGPL, you are required
to have the complete, corresponding source present for the correct
version as distributed in the firmware. You also must include the
"scripts to control compilation and installation of the executable".
* lib/libgio-2.0.so.0.2400.2
* lib/libglib-2.0.so.0.2400.2
* lib/libgmodule-2.0.so.0.2400.2
* lib/libgobject-2.0.so.0.2400.2
* lib/libgthread-2.0.so.0.2400.2
* lib/libz.so.1.2.5 (version 1.2.2 is provided in the sources)
# Weird Build Issues Over Many Candidates
<hr/>
You mentioned in your Round 6 commentary that you have corrected the
thatlib issues. However, we are unable to see what you mean. There are
now two copies of thatlib, one in 2624.7_524/uclinux-rootfs/lib/thatlib/,
as well as the one in yourlibs. We aren't sure which one you intend to
be built to generate the binaries on the firmware. When we try to build
the yourlibs one from scratch, by cleaning the whole area, we get the
following build issues. Here's what we did:
# Getting Really build-technical
<hr/>
We ran:
make -C libsrc/thatlib install
which did not work because of a missing Makefile error. We read the
build source and discovered that the Makefile, etc, for that directory
is generated by running:
cd libsrc/thatlib/thatlib-0.9.22_mipsel-uclibc; sh configure_thatlib_mipsel-uclibc
Once we did that
make -C libsrc/thatlib install
worked correctly. The only remaining binaries were in build source and
discovered that the Makefile, etc, for that directory is generated by
running:
cd libsrc/thatlib/thatlib-0.9.22_mipsel-uclibc; sh configure_thatlib_mipsel-uclibc
# Getting Really build-technical
<hr/>
Once we did that
make -C libsrc/thatlib install
worked correctly. The only remaining binaries were in
./libsrc/thatlib/\{YOURLIB_ROOT_DIR\}/ which looks like a build with a
misconfigured environment somehow, so we simply removed that
directory.
Then, after running make clean, thatlib failed with the following
errors. Random .o/.so files laying around in the thatlib source
directory, and then it failing to build correctly after they are
removed. If there some set of .so files you claim are not required
as part of the C&CS since thatlib is LGPL'd, we understand that, but
the rest of the sources must build and install those other .so's.
Here's the build error we get in the bdvdlibs version:
# Getting Really build-technical
<hr/>
mkdir .libs/libthatlibwm_default.a.tmp
(cd .libs/libthatlibwm_default.a.tmp && ar x ../../.libs/libthatlibwm_default.a)
mkdir .libs/libthatlibwm_default.a.tmp
(cd .libs/libthatlibwm_default.a.tmp && ar x ../../.libs/libthatlibwm_default.a)
/opt/toolchains/crosstools_sf-linux-2.6.18.0_gcc-4.2-9ts_uclibc-nptl-0.9.29-20070423_20080702/bin//mipsel-uclibc-
ld -o libthatlibwm_default.o -r .libs/libthatlibwm_default.a.tmp/*.o
/opt/toolchains/crosstools_sf-linux-2.6.18.0_gcc-4.2-9ts_uclibc-nptl-0.9.29-20070423_20080702/bin//mipsel-uclibc-
ld: .libs/libthatlibwm_default.a.tmp/default.o: Relocations in generic ELF (EM: 3)
/opt/toolchains/crosstools_sf-linux-2.6.18.0_gcc-4.2-9ts_uclibc-nptl-0.9.29-20070423_20080702/bin//mipsel-uclibc-
ld: .libs/libthatlibwm_default.a.tmp/default.o: Relocations in generic ELF (EM: 3)
/opt/toolchains/crosstools_sf-linux-2.6.18.0_gcc-4.2-9ts_uclibc-nptl-0.9.29-20070423_20080702/bin//mipsel-uclibc-
ld: .libs/libthatlibwm_default.a.tmp/default.o: Relocations in generic ELF (EM: 3)
/opt/toolchains/crosstools_sf-linux-2.6.18.0_gcc-4.2-9ts_uclibc-nptl-0.9.29-20070423_20080702/bin//mipsel-uclibc-
ld: .libs/libthatlibwm_default.a.tmp/default.o: Relocations in generic ELF (EM: 3)
.libs/libthatlibwm_default.a.tmp/default.o: could not read symbols: File in wrong format
make[4]: *** [libthatlibwm_default.o] Error 1
# Proprietary Linux Modules Are Everywhere
<hr/>
We did find one .ko file that was already included in the package, but
wasn't built when we ran "./build.sh". This is
path/path/android_X.X/device/device-type/mydevice.ko , which notes that
its license is "GPL v2" in the modinfo, but for which we could find no
source code in the source release. Please ensure that the source code
for mydevice.ko is included in the next source candidate.
# Proprietary Linux Modules Are Everywhere
<hr/>
* The following files are derivative of the kernel named Linux and
therefore covered by the GPL. However, no source code, scripts to
control compilation nor installation are included in your CCS
candidate:
lib/modules/myfilesystem.ko
lib/modules/mydevicecontroller.ko
lib/modules/myblockdevice.ko
lib/modules/mypcicard.ko
# Non-Technical GPL Compliance Issues
<hr/>
Regarding over the air updates: we'd like to see a screenshot or other
details documenting what has now been implemented by BestBuy to make
sure the offer for source appears to users appropriately after
upgrade. There was a consensus reached on the last conference call
how this would be done, so we only need follow up and implementation
on that.
# Binary Comparison.
<hr/>
Note that we did not receive a firmware image to compare this with
(though we do have the device). Company's website did not appear to
have any firmware images available for download. It would be helpful to
have such an image for the next CCS check.
The above source candidate was downloaded from
http:///sourcez.company.com/en/search/index.htm?keywords=X1234Y, which
was alluded to in Company's 2017-01-18 email to us that said:
"You can check this website
http://sourcez.company.com/en/search/index.htm "
The email did not mention how to use that website, but we found that by
entering "X1234Y" into the top right search box that we could find the
source file list.
Note that the offer for source included in the web UI of the device said
to email NAME@COMPANY.com , which is how the above instructions for
downloading the source were received.
# A Pristine Example
+ Enforcement must often use a “know it when I see it” standard.
+ i.e., can we take your CCS build it, and install it?
+ We've reached compliant CCS with hundreds of companies:
+ but that didn't mean the CCS was pretty.
+ Thanks to ThinkPenguin, we finally have an example of beautiful embedded
product compliance.
# Lessons Learned from Pristine Example
+ The full paper for this talk is available online:
+ [compliance.guide/pristine-example](http://compliance.guide/pristine-example)
+ It's part of the larger tutorial called [*Copyleft and the GNU General
Public License: A Comprehensive Tutorial and Guide*](https://copyleft.org/guide/)
at copyleft.org.
# Give a roadmap in a README
+ Scripts doesn't only mean shell scripts and Makefiles.
+ Think of the script of a play or movie.
+ If your build process includes human intervention …
+ … then the script are a written explanation of what the human must
do.
<hr>
> **the scripts** used to control compilation and installation of the executable.
<p align=right>
— GPLv2§3
</p>
</span>
# ThinkPengiun's README
A file called “README” at the top-level directory said:
In order to build firmware images for your router, the following needs to be installed:
gcc, binutils, bzip2, flex, python, perl, make, find, grep, diff, unzip,
gawk, getopt, libz-dev and libc headers.
Please use “make menuconfig” to configure your appreciated configuration
for the toolchain and firmware. Please note that the default configuration
is what was used to build the firmware image for your router. It is advised
that you use this configuration.
Simply running “make” will build your firmware. The build system will
download all sources, build the cross-compile toolchain, the kernel and all
chosen applications.
To build your own firmware you need to have access to a GNU/Linux system
(case-sensitive filesystem required).
# Make Sure It Builds
+ Can your CCS pass this test?
+ Give you source release to another developer from another department.
+ Ask them to follow the instructions you wrote.
+ They should get the equivalent binaries you get in building.
+ Very few organizations bother to do this.
+ It's probably the most useful step to verify compliance, yet *no*
compliance process recommendations I've ever seen include this.
<hr>
> the scripts used to **control compilation** and installation of the executable.
<p align=right>
— GPLv2§3
</p>
</span>
# It's not “make install”
+ Server system software can offer a “make install” that
reasonable works to meet installation requirements.
+ Embedded products are admittedly difficult to install.
+ To comply here, you'll usually just have write out the instructions.
+ It is required; don't skip this part.
<hr>
> the scripts used to **control** compilation and **installation** of the executable.
<p align=right>
— GPLv2§3
</p>
</span>
# Missing hardware components
+ Inclusion of specialized installation hardware is not a
“script”.
+ In our ThinkPenguin example, we had to go buy a USB serial adapter to
install the modified firmware.
+ Just tell the user what they have to go buy for the install to work.
# More Info / Talk License
<img align="right" src="cc-by-sa-4-0_88x31.png" />
+ Specific Sections of Copyleft Guide relating to these topics:
- [The Pristine Example](https://copyleft.org/guide/comprehensive-gpl-guidech22.html#x29-15900021)
- [Details of a Compliant Distribution](https://copyleft.org/guide/comprehensive-gpl-guidech16.html#x21-12700015)
<span class="fitonslide">
<p>Presentation and slides are: Copyright © Bradley M. Kuhn (2008–2011, 2015, 2017), Karen M. Sandler (2017), and are licensed under the <a rel="license" href="https://creativecommons.org/licenses/by-sa/4.0/legalcode">Creative Commons Attribution-Share Alike 4.0 International License</a>. </p>
</span>
|