On Mon, Nov 23, 2020 at 02:09:05PM -0800, syzbot wrote: > Hello, > > syzbot has tested the proposed patch but the reproducer is still triggering an issue: > memory leak in rxrpc_lookup_local > > BUG: memory leak > unreferenced object 0xffff888117ab9900 (size 256): > comm "syz-executor.0", pid 8883, jiffies 4294943811 (age 433.620s) > hex dump (first 32 bytes): > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ > 00 00 00 00 0a 00 00 00 00 80 cb 17 81 88 ff ff ................ > backtrace: > [<000000009003383a>] kmalloc include/linux/slab.h:552 [inline] > [<000000009003383a>] kzalloc include/linux/slab.h:664 [inline] > [<000000009003383a>] rxrpc_alloc_local net/rxrpc/local_object.c:79 [inline] > [<000000009003383a>] rxrpc_lookup_local+0x1c1/0x760 net/rxrpc/local_object.c:244 > [<00000000609410d3>] rxrpc_bind+0x174/0x240 net/rxrpc/af_rxrpc.c:149 > [<00000000661f73ad>] afs_open_socket+0xdb/0x200 fs/afs/rxrpc.c:64 > [<00000000e3eb5768>] afs_net_init+0x2b4/0x340 fs/afs/main.c:126 > [<000000002c6bf109>] ops_init+0x4e/0x190 net/core/net_namespace.c:152 > [<000000009ce0aa62>] setup_net+0xdb/0x2d0 net/core/net_namespace.c:342 > [<00000000db8c8dc2>] copy_net_ns+0x14b/0x320 net/core/net_namespace.c:483 > [<00000000b04b70a8>] create_new_namespaces+0x199/0x4e0 kernel/nsproxy.c:110 > [<000000005dc01eb8>] unshare_nsproxy_namespaces+0x9b/0x120 kernel/nsproxy.c:231 > [<00000000422ec6bd>] ksys_unshare+0x2fe/0x5c0 kernel/fork.c:2949 > [<0000000042f77bee>] __do_sys_unshare kernel/fork.c:3017 [inline] > [<0000000042f77bee>] __se_sys_unshare kernel/fork.c:3015 [inline] > [<0000000042f77bee>] __x64_sys_unshare+0x12/0x20 kernel/fork.c:3015 > [<00000000e58e69f9>] do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46 > [<000000000a67195e>] entry_SYSCALL_64_after_hwframe+0x44/0xa9 Okay, that confirms it. This is a completely different memory leak, as can be seen by comparing the stack trace with the previous one. The problem with the gspca driver is gone. Mauro/Hans, what should I do with the patch? Alan Stern
