i#147: Alternate-bitwidth client support #4324

derekbruening · 2020-06-14T15:25:30Z

Adds new options and interfaces to specify alternate-bitwidth client
libraries for use when the application creates a child process of the
other bitwidth.

For DR, adds -client_lib32 and -client_lib64 options. Switches main
usage to use the appropriate option, with its contents then copied
into -client_lib (to avoid the pain of removing that option).

For drconfiglib, adds dr_register_client_ex() with
dr_client_iterator_next_ex() to support querying other-bitwidth client
registrations.

Adds a new libutil.drconfig_test for drconfiglib. Fixes a bug found
by the test: existing client queries were cutting off the last
character of the path and options.

For drrun and drconfig, adds "-c32" and "-c64" options, with an
additional "--" separating the client options between them.

For tool files, adds CLIENT{32,64}_{REL,ABS}. Updates drcov,
drcpusim, and drcachesim to use the new syntax and drcachesim's
launcher to process it. Tested these manually:
===========================================================================
$ ninja install
$ rm *.log
$ ../exports/bin64/drrun -t drcov -- ~/dr/test/execve64 ~/dr/test/hello32
$ l -t .log
20K drcov.execve64.109583.0000.proc.log 20K drcov.execve64.109585.0000.proc.log
20K drcov.hello32.109585.0000.proc.log
$ rm -rf drm.dir
$ ../exports/bin64/drrun -verbose -t drcachesim -verbose 1 -offline -- ~/dr/test/execve64 ~/dr/test/hello32
$ l -td *.dir
4.0K drmemtrace.hello32.117714.7095.dir/ 4.0K drmemtrace.execve64.117714.6260.dir/
4.0K drmemtrace.execve64.117713.9314.dir/
===========================================================================

Adds tests of -c32/-c64 to the existing cross-arch linux.execve{32,64}
tests (Windows won't work until #803 is addressed).
The tests look like this:
===========================================================================
$ cmake . && ctest -V -R 'linux.execve(32|64)'

8: Running test command: "/home/bruening/dr/git/build_x64_dbg_tests/bin64/drrun" "-32" "-dr_home" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit" "-stderr_mask" "0xC" "-dumpcore_mask" "0" "-c32" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "-c64" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/linux.execve32" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/linux.execve-sub64"
8: large_options passed: -paramA foo -paramB bar
8: parent is running under DynamoRIO
8: parent waiting for child
8: child is running under DynamoRIO
8: large_options passed: -paramA foo -paramB bar
8: it_worked
8: running under DynamoRIO
8: large_options exiting
8: child has exited
8: large_options exiting
8:
1/2 Test #8: linux.execve32 ...................   Passed    3.93 sec

9: Test command: /home/bruening/dr/git/build_x64_dbg_tests/bin64/drrun "-stderr_mask" "0xC" "-dumpcore_mask" "0" "-c32" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "-c64" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/linux.execve64" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/linux.execve-sub32"
9: Test timeout computed to be: 1500
9: large_options passed: -paramA foo -paramB bar
9: parent is running under DynamoRIO
9: parent waiting for child
9: child is running under DynamoRIO
9: large_options passed: -paramA foo -paramB bar
9: it_worked
9: running under DynamoRIO
9: large_options exiting
9: child has exited
9: large_options exiting
2/2 Test #9: linux.execve64 ...................   Passed    0.86 sec
===========================================================================

Issue: #147, #803
Fixes: #147

Adds new options and interfaces to specify alternate-bitwidth client libraries for use when the application creates a child process of the other bitwidth. For DR, adds -client_lib32 and -client_lib64 options. Switches main usage to use the appropriate option, with its contents then copied into -client_lib (to avoid the pain of removing that options). For drconfiglib, adds dr_register_client_ex() with dr_client_iterator_next_ex() to support querying other-bitwidth client registrations. Adds a new libutil.drconfig_test for drconfiglib. Fixes a bug found by the test: existing client queries were cutting off the last character of the path and options. For drrun and drconfig, adds "-c32" and "-c64" options, with an additional "--" separating the client options between them. For tool files, adds CLIENT{32,64}_{REL,ABS}. Updates drcov, drcpusim, and drcachesim to use the new syntax and drcachesim's launcher to process it. Tested these manually: =========================================================================== $ ninja install $ rm *.log $ ../exports/bin64/drrun -t drcov -- ~/dr/test/execve64 ~/dr/test/hello32 $ l -t *.log 20K drcov.execve64.109583.0000.proc.log 20K drcov.execve64.109585.0000.proc.log 20K drcov.hello32.109585.0000.proc.log $ rm -rf drm*.dir $ ../exports/bin64/drrun -verbose -t drcachesim -verbose 1 -offline -- ~/dr/test/execve64 ~/dr/test/hello32 $ l -td *.dir 4.0K drmemtrace.hello32.117714.7095.dir/ 4.0K drmemtrace.execve64.117714.6260.dir/ 4.0K drmemtrace.execve64.117713.9314.dir/ =========================================================================== Adds tests of -c32/-c64 to the existing cross-arch linux.execve{32,64} tests (Windows won't work until #803 is addressed). The tests look like this: =========================================================================== $ cmake . && ctest -V -R 'linux.execve(32|64)' 8: Running test command: "/home/bruening/dr/git/build_x64_dbg_tests/bin64/drrun" "-32" "-dr_home" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit" "-stderr_mask" "0xC" "-dumpcore_mask" "0" "-c32" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "-c64" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/linux.execve32" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/linux.execve-sub64" 8: large_options passed: -paramA foo -paramB bar 8: parent is running under DynamoRIO 8: parent waiting for child 8: child is running under DynamoRIO 8: large_options passed: -paramA foo -paramB bar 8: it_worked 8: running under DynamoRIO 8: large_options exiting 8: child has exited 8: large_options exiting 8: 1/2 Test #8: linux.execve32 ................... Passed 3.93 sec 9: Test command: /home/bruening/dr/git/build_x64_dbg_tests/bin64/drrun "-stderr_mask" "0xC" "-dumpcore_mask" "0" "-c32" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "-c64" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/libclient.large_options.dll.so" "-paramA" "foo" "-paramB" "bar" "--" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/bin/linux.execve64" "/home/bruening/dr/git/build_x64_dbg_tests/suite/tests/32bit/suite/tests/bin/linux.execve-sub32" 9: Test timeout computed to be: 1500 9: large_options passed: -paramA foo -paramB bar 9: parent is running under DynamoRIO 9: parent waiting for child 9: child is running under DynamoRIO 9: large_options passed: -paramA foo -paramB bar 9: it_worked 9: running under DynamoRIO 9: large_options exiting 9: child has exited 9: large_options exiting 2/2 Test #9: linux.execve64 ................... Passed 0.86 sec =========================================================================== Issue: #147, #803 Fixes: #147

derekbruening · 2020-06-14T15:27:58Z

@hunterbr72 here is Step 1 of solving the alternate-bitwidth. Step 2 is the Windows injection #803.

… drconfiglib.dll into bin dir on Windows; fix unfilled-in field on query; fix clang build warning

derekbruening · 2020-06-14T20:52:42Z

Appveyor's test is not finding the drconfiglib.dll copy:

229/251 Test #230: code_api|libutil.drconfig_test ...............................Exit code 0xc0000135
21465***Exception:   0.02 sec

It passes locally on my machine. Grrr.

johnfxgalea

I'll look at the rest later on today.

api/docs/release.dox

clients/drcachesim/common/options.cpp

core/lib/dr_config.h

johnfxgalea · 2020-06-15T02:30:53Z

core/lib/dr_config.h

@@ -685,15 +800,17 @@ DR_EXPORT
 *                              configurations to unset.
 *
 * \return      The number of clients registered for the given process and platform.
+ *              An alternative-bitwidth client is counted as a separate client.
 */
 size_t
 dr_num_registered_clients(const char *process_name, process_id_t pid, bool global,


Not sure, but has this function ever been used to determine the next available client ID? Now client IDs might not be sequential and may skip valid values due to the registration of alternative bitwidth clients. I don't think this is a problem, just an observation. No biggie.

The next available priority I think you mean. Yes, it is in fact used for that very thing in drdeploy.c. These alt clients end up being treated as regular clients in most respects so I guess it's not really a gap. The interface is a little weird that way but I think it's better than my first attempt which tacked on an alt path as an attribute to a single client, not allowing different options (which was voted for with the -c32/-c64).

Actually, I'm also referring to Client IDs, assuming alternative bitwidth clients are assigned with the same ID as their regular clients. But like I said, this is just an observation, not an issue.

core/lib/dr_config.h

libutil/dr_config.c

…nt tweaks

johnfxgalea · 2020-06-15T17:01:39Z

Can I have a look now?

derekbruening · 2020-06-15T17:04:33Z

Yes, PTAL

…imar then alt

johnfxgalea · 2020-06-15T17:26:04Z

core/lib/dr_config.h

@@ -685,15 +800,17 @@ DR_EXPORT
 *                              configurations to unset.
 *
 * \return      The number of clients registered for the given process and platform.
+ *              An alternative-bitwidth client is counted as a separate client.
 */
 size_t
 dr_num_registered_clients(const char *process_name, process_id_t pid, bool global,


Actually, I'm also referring to Client IDs, assuming alternative bitwidth clients are assigned with the same ID as their regular clients. But like I said, this is just an observation, not an issue.

suite/tests/libutil/drconfig_test.c

api/docs/intro.dox

derekbruening · 2020-06-15T18:30:48Z

Sorry, had to put in another commit b/c drdeploy was registering alt clients first from a tool file (just easier in that order w/ the code how it was), so shifted it down to solve. That should fix the CI failures.

derekbruening · 2020-06-15T18:47:42Z

Actually, I'm also referring to Client IDs, assuming alternative bitwidth clients are assigned with the same ID as their regular clients. But like I said, this is just an observation, not an issue.

Github won't let me reply in the thread so replying here: but a client ID can be anything and doesn't have to be sequential (the priority is what's sequential -- though looking at it now if we did it again I'd probably make them the same thing).

derekbruening · 2020-06-15T19:49:38Z

Grr this ordering enforcement is causing more problems: drdeploy was just appending, which for 64-bit would mean you'd have to pass "-c64 -c32" order. I guess it can insert.

derekbruening · 2020-06-15T19:53:49Z

No order enforcement, or "-c3264" combined arg, was certainly simpler.

…der; add another test case; fix Windows TCHAR vs char buf

derekbruening requested a review from johnfxgalea June 14, 2020 15:25

derekbruening mentioned this pull request Jun 14, 2020

client interactions with cross-arch execve #147

Closed

Fix pre-existing Windows drconfig handle leak found by new test; copy…

c5125f7

… drconfiglib.dll into bin dir on Windows; fix unfilled-in field on query; fix clang build warning

Fix drconfiglib copy rule

2362e8d

johnfxgalea requested changes Jun 15, 2020

View reviewed changes

Enforce primary then alt; use DR's snprintf; add a copy helper; comme…

02e1952

…nt tweaks

Reverse order of client reg for tool files to comply with required pr…

a31bb86

…imar then alt

johnfxgalea approved these changes Jun 15, 2020

View reviewed changes

derekbruening added 2 commits June 15, 2020 17:31

Reverse order of -c32 -c64 for x64 builds to get alt 2nd; document or…

f4720ef

…der; add another test case; fix Windows TCHAR vs char buf

Fix clang warning

eac4c1d

derekbruening merged commit 4471c9b into master Jun 15, 2020

derekbruening deleted the i147-other-arch-client branch June 15, 2020 23:47

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

i#147: Alternate-bitwidth client support #4324

i#147: Alternate-bitwidth client support #4324

derekbruening commented Jun 14, 2020

derekbruening commented Jun 14, 2020

derekbruening commented Jun 14, 2020

johnfxgalea left a comment

johnfxgalea Jun 15, 2020 •

edited

Loading

derekbruening Jun 15, 2020

johnfxgalea Jun 15, 2020

johnfxgalea commented Jun 15, 2020

derekbruening commented Jun 15, 2020

johnfxgalea Jun 15, 2020

derekbruening commented Jun 15, 2020

derekbruening commented Jun 15, 2020

derekbruening commented Jun 15, 2020

derekbruening commented Jun 15, 2020

i#147: Alternate-bitwidth client support #4324

i#147: Alternate-bitwidth client support #4324

Conversation

derekbruening commented Jun 14, 2020

derekbruening commented Jun 14, 2020

derekbruening commented Jun 14, 2020

johnfxgalea left a comment

Choose a reason for hiding this comment

johnfxgalea Jun 15, 2020 • edited Loading

Choose a reason for hiding this comment

derekbruening Jun 15, 2020

Choose a reason for hiding this comment

johnfxgalea Jun 15, 2020

Choose a reason for hiding this comment

johnfxgalea commented Jun 15, 2020

derekbruening commented Jun 15, 2020

johnfxgalea Jun 15, 2020

Choose a reason for hiding this comment

derekbruening commented Jun 15, 2020

derekbruening commented Jun 15, 2020

derekbruening commented Jun 15, 2020

derekbruening commented Jun 15, 2020

johnfxgalea Jun 15, 2020 •

edited

Loading