Commit graph

8 commits

Author SHA1 Message Date
Christopher Ferris
f4ada9c9ce Return total footprint, not high water mark.
The mallinfo usmblks value returned by dlmalloc is a little misleading.
It's not the current max, it's the historical high water mark. This
leads to dumpsys meminfo producing native memory numbers that don't add up.
Change this to the real total footprint, not this high water mark.

Bug: 17265653
Change-Id: Id0293a1b50c9b0be8795405049f537a51ab0e8b7
2014-10-13 12:29:30 -07:00
Christopher Ferris
a403780538 Put all allocation functions into dispatch table.
Implement these new functions for all of the debug malloc types.

Fix a number of bugs in the debug malloc functions related to overflow
conditions.
Fix a bug in dlpvalloc due to an overflow condition.

Fix various other bugs in the debug malloc functions.

Add new tests for malloc functions.

Bug: 11225066

Change-Id: Idf50f389603e2157645565bc15cd9365eec2e9dd
2014-06-12 15:08:18 -07:00
Ian Rogers
09f76cdd9b Fix build warning of initialization but no use.
first_run is only used with LOCK_AT_FORK, however, care needs to be
taken with the #if to avoid compiler warnings when LOCK_AT_FORK isn't
being used.

Change-Id: I6bcfac2ab8732c91607f4a1bcd7c0019f29c2eec
2012-09-06 09:59:13 -07:00
Ian Rogers
82978cae4e Merge "Avoid malloc lock while calling pthread_atfork." 2012-09-05 07:41:32 -07:00
Elliott Hughes
63deae5e4f Add missing MALLOC_FAILURE_ACTION calls to dlmalloc.
Without these, sometimes malloc(3) returns NULL without setting errno.

Change-Id: I4708c3f675bf2c878ddcaf012fde7848b255826b
2012-09-04 15:42:23 -07:00
Kirill Artamonov
3e2d2936b0 Avoid malloc lock while calling pthread_atfork.
Expecting the memory in a forked child process to be sane wrt threading
is a bad idea. An example of a problem is when the parent process has
the malloc lock and a child process is forked. The malloc lock in the
child will appear locked by a thread that doesn't exist.

This change aims to make bionic more compatible with glibc by reseting
the malloc lock in the child forked process, as well as holding it
during the fork. This is a feature in dlmalloc 2.8.6 called
LOCK_AT_FORK. In general this feature isn't necessary as a forked
process will then exec. Some bad applications rely on being able to use
features like malloc before the exec and having multiple threads running
in the parent program. This isn't a problem with glibc and this patch
makes it not a problem for bionic.

Unfortunately for use in bionic, LOCK_AT_FORK has an issue as internally
it uses pthread_atfork that in bionic uses malloc. This leads to the
LOCK_AT_FORK initialization deadlocking with pthread_atfork's call to
malloc due to the malloc lock. This change moves the pthread_atfork logic
in LOCK_AT_FORK to be called without the malloc lock held.

Change-Id: Id68175a564a6abb936ee4488b44d9479f7311f69
2012-08-31 11:11:40 -07:00
Ian Rogers
c6d95add30 Upgrade to dlmalloc 2.8.6.
This fixes a bug and enables the use of MORECORE_CONTIGUOUS.

Change-Id: Ia7c5d32bdc46e99b3ecb92ee94d1f702c4385d5d
2012-08-29 18:18:29 -07:00
Ian Rogers
999089181e Upgrade to dlmalloc 2.8.5.
Move dlmalloc code to upstream-dlmalloc to make pulling upstream changes
easier.
Declare pvalloc and malloc_usable_size routines present in malloc.h but with
missing implementations. Remove other functions from malloc.h that have
no implementation nor use in Android.

Change-Id: Ia6472ec6cbebc9ad1ef99f4669de9d33fcc2efb4
2012-08-20 14:12:20 -07:00