A Git based version of Magic Lantern, for those unwilling or unable to work using Mercurial. The vast majority of branches have been removed, with those thought to be important brought in individually and merged.
GNU General Public License v2.0
142
stars
47
forks
source link
EDmac module can leak significant amounts of mem #124
CreateResLockEntry() returns LockEntry *. modules/edmac/edmac.c can do this in a loop due to find_free_edmac_channels(), and never frees it.
LockEntry has pResources that is also allocated inside CreateResLockEntry(), of variable size (although fixed and small in the case of find_free_edmac_channels()). Max is 32 loops, on 200D 0x28 bytes per loop, 1280 bytes max per call.
Presumably we should free pResources and the LockEntry, although I have not tested this. I think you want to Unlock before the free, to remove avoid dangling pointers in ResourceInfo_global.
Things to check:
is it really a leak? I believe this only due to static analysis on 200D, wants to be tested
is the free safe on all gens? Only checked 200D code so far
audit all other uses of CreateResLockEntry() (quick check suggests we never free, but maybe we only assign once so it doesn't matter, we never lose the ref)
CreateResLockEntry() returns LockEntry *. modules/edmac/edmac.c can do this in a loop due to find_free_edmac_channels(), and never frees it.
LockEntry has pResources that is also allocated inside CreateResLockEntry(), of variable size (although fixed and small in the case of find_free_edmac_channels()). Max is 32 loops, on 200D 0x28 bytes per loop, 1280 bytes max per call.
Presumably we should free pResources and the LockEntry, although I have not tested this. I think you want to Unlock before the free, to remove avoid dangling pointers in ResourceInfo_global.
Things to check: