PaX exception for PagerDuty #23

Closed
ghost opened this Issue Aug 27, 2015 · 7 comments

Comments

Projects
None yet
1 participant
@ghost

ghost commented Aug 27, 2015

This seems to still occur on the 8/26 build. Turning on PaX Soft Mode prevents the crash from happening.

Not urgent, just wanted to pass this on in case there's some other corner case not covered by the fix from #18...

Sorry about referencing tons of other issues with the stack trace in #22. Will avoid doing that in the future.

--------- beginning of crash
F/libc (11173): Fatal signal 11 (SIGSEGV), code 1, fault addr 0x15f in tid 11246 (Thread-39)
I/DEBUG ( 202): property debug.db.uid not set; NOT waiting for gdb.
I/DEBUG ( 202): HINT: adb shell setprop debug.db.uid 100000
I/DEBUG ( 202): HINT: adb forward tcp:5039 tcp:5039
I/ActivityManager( 796): Displayed com.pagerduty.android/.LoginActivity: +498ms (total +1s409ms)
I/Timeline( 796): Timeline: Activity_windows_visible id: ActivityRecord{2e8398e1 u0 com.pagerduty.android/.LoginActivity t169} time:1031753
I/DEBUG ( 202): *** *** *** *** *** *** *** *** *** *** *** *** *** *** ** ***
I/DEBUG ( 202): Build fingerprint: 'google/hammerhead/hammerhead:5.1.1/LMY48B/1863243:user/release-keys'
I/DEBUG ( 202): Revision: '11'
I/DEBUG ( 202): ABI: 'arm'
I/DEBUG ( 202): pid: 11173, tid: 11246, name: Thread-39 >>> com.pagerduty.android <<<
I/DEBUG ( 202): signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x15f
I/DEBUG ( 202): r0 00000000 r1 8cb655c0 r2 8f1a8cc0 r3 8cb62000
I/DEBUG ( 202): r4 00000000 r5 8cb62000 r6 00000001 r7 009aec8e
I/DEBUG ( 202): r8 8cb62878 r9 918a1234 sl 918a0cec fp 00000001
I/DEBUG ( 202): ip 00000000 sp 8e7f4758 lr 91324973 pc 913246c8 cpsr 80010030
I/DEBUG ( 202):
I/DEBUG ( 202): backtrace:
I/DEBUG ( 202): 00 pc 011396c8 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 01 pc 0113996f /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 02 pc 01141937 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 03 pc 01141b11 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 04 pc 0124da15 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 05 pc 011b813f /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 06 pc 0130180f /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 07 pc 010d1bb1 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 08 pc 00c5eeab /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 09 pc 00b38305 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 10 pc 00b383fb /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 11 pc 00b3524f /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 12 pc 007ebffd /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 13 pc 00d51a79 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 14 pc 00d4e751 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 15 pc 00d5104b /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 16 pc 01389fff /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 17 pc 00fe07ed /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 18 pc 00fe065b /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 19 pc 0023f8a1 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 20 pc 0021f73b /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 21 pc 00220c77 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 22 pc 00221093 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 23 pc 0022127f /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 24 pc 0021fa01 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 25 pc 00226d89 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 26 pc 0021f2e9 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 27 pc 0023464d /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 28 pc 00231243 /system/lib/libwebviewchromium.so
I/DEBUG ( 202): 29 pc 00013acd /system/lib/libc.so (__pthread_start(void
)+30)
I/DEBUG ( 202): 30 pc 000119ff /system/lib/libc.so (__start_thread+6)

@thestinger

This comment has been minimized.

Show comment Hide comment
@thestinger

thestinger Aug 27, 2015

Contributor

Sorry about referencing tons of other issues with the stack trace in #22.

No worries. FWIW, you can wrap code blocks with a ~~~ line above and below instead of editing the content.

Contributor

thestinger commented Aug 27, 2015

Sorry about referencing tons of other issues with the stack trace in #22.

No worries. FWIW, you can wrap code blocks with a ~~~ line above and below instead of editing the content.

@thestinger

This comment has been minimized.

Show comment Hide comment
@thestinger

thestinger Aug 27, 2015

Contributor

Strange, this seems to be automatically handled for me. The PackageParser code I wrote logs the expected message, and it gets the permission for dynamic code generation added:

D/PackageParser(  608): adding MPROTECT exception due to setJavaScriptEnabled call in com.pagerduty.android

Not sure what's going wrong.

Contributor

thestinger commented Aug 27, 2015

Strange, this seems to be automatically handled for me. The PackageParser code I wrote logs the expected message, and it gets the permission for dynamic code generation added:

D/PackageParser(  608): adding MPROTECT exception due to setJavaScriptEnabled call in com.pagerduty.android

Not sure what's going wrong.

@thestinger

This comment has been minimized.

Show comment Hide comment
@thestinger

thestinger Aug 27, 2015

Contributor

@andy11: I think I see the problem. I don't handle this in the code path deciding if new permissions can be granted for apps that are already installed. I'm guessing that you already had this app installed before the update. Can you try uninstalling and reinstalling it?

I'll have to figure out a way to work around this issue.

Contributor

thestinger commented Aug 27, 2015

@andy11: I think I see the problem. I don't handle this in the code path deciding if new permissions can be granted for apps that are already installed. I'm guessing that you already had this app installed before the update. Can you try uninstalling and reinstalling it?

I'll have to figure out a way to work around this issue.

@thestinger

This comment has been minimized.

Show comment Hide comment
@thestinger

thestinger Aug 27, 2015

Contributor

On the other hand now that the code to deal with this is in place, the problem won't occur anymore (i.e. it will be detected at first install). So, assuming that's the issue, nothing really has to be fixed.

Contributor

thestinger commented Aug 27, 2015

On the other hand now that the code to deal with this is in place, the problem won't occur anymore (i.e. it will be detected at first install). So, assuming that's the issue, nothing really has to be fixed.

@ghost

This comment has been minimized.

Show comment Hide comment
@ghost

ghost Aug 27, 2015

Ah, that did it! May not even be worth handling if 99.99% of your users will be starting fresh from builds where the update is already in place. I'll try re-installing any apps that I have issues with. Thanks!

ghost commented Aug 27, 2015

Ah, that did it! May not even be worth handling if 99.99% of your users will be starting fresh from builds where the update is already in place. I'll try re-installing any apps that I have issues with. Thanks!

@thestinger

This comment has been minimized.

Show comment Hide comment
@thestinger

thestinger Aug 27, 2015

Contributor

It might be worth working around because I'll add new heuristics like this in the future (for example, scanning for libraries with text relocations).

Contributor

thestinger commented Aug 27, 2015

It might be worth working around because I'll add new heuristics like this in the future (for example, scanning for libraries with text relocations).

@thestinger

This comment has been minimized.

Show comment Hide comment
@thestinger

thestinger Sep 1, 2015

Contributor

Closing in favour of #24.

Contributor

thestinger commented Sep 1, 2015

Closing in favour of #24.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment