* [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() @ 2022-08-03 5:02 Peilin Ye 2022-08-03 13:16 ` Paul Moore 2022-08-03 22:23 ` [PATCH v2] " Peilin Ye 0 siblings, 2 replies; 10+ messages in thread From: Peilin Ye @ 2022-08-03 5:02 UTC (permalink / raw) To: Jens Axboe, Pavel Begunkov, Paul Moore, Eric Paris Cc: Peilin Ye, io-uring, linux-kernel, linux-audit, Peilin Ye From: Peilin Ye <[email protected]> Currently @audit_context is allocated twice for io_uring workers: 1. copy_process() calls audit_alloc(); 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which is effectively audit_alloc()) and overwrites @audit_context, causing: BUG: memory leak unreferenced object 0xffff888144547400 (size 1024): <...> hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 [<ffffffff81239e63>] copy_process+0xcd3/0x2340 [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 [<ffffffff81686604>] create_io_worker+0xb4/0x230 [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 [<ffffffff816768b3>] io_queue_async+0x113/0x180 [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 [<ffffffff8125a688>] get_signal+0xc18/0xf10 [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 Then, 3. io_sq_thread() or io_wqe_worker() frees @audit_context using audit_free(); 4. do_exit() eventually calls audit_free() again, which is okay because audit_free() does a NULL check. Free the old @audit_context first in audit_alloc_kernel(), and delete the redundant calls to audit_free() for less confusion. Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") Cc: [email protected] Signed-off-by: Peilin Ye <[email protected]> --- Hi all, A better way to fix this memleak would probably be checking @args->io_thread in copy_process()? Something like: if (args->io_thread) retval = audit_alloc_kernel(); else retval = audit_alloc(); But I didn't want to add another if to copy_process() for this bugfix. Please suggest, thanks! Peilin Ye fs/io-wq.c | 1 - fs/io_uring.c | 2 -- kernel/auditsc.c | 1 + 3 files changed, 1 insertion(+), 3 deletions(-) diff --git a/fs/io-wq.c b/fs/io-wq.c index 824623bcf1a5..0f4804a5e873 100644 --- a/fs/io-wq.c +++ b/fs/io-wq.c @@ -670,7 +670,6 @@ static int io_wqe_worker(void *data) if (test_bit(IO_WQ_BIT_EXIT, &wq->state)) io_worker_handle_work(worker); - audit_free(current); io_worker_exit(worker); return 0; } diff --git a/fs/io_uring.c b/fs/io_uring.c index e8e769be9ed0..0f27914f37f2 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -9283,8 +9283,6 @@ static int io_sq_thread(void *data) io_run_task_work(); mutex_unlock(&sqd->lock); - audit_free(current); - complete(&sqd->exited); do_exit(0); } diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 3a8c9d744800..7948090fd12f 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -1095,6 +1095,7 @@ int audit_alloc_kernel(struct task_struct *tsk) * 2. The {set,clear}_task_syscall_work() ops likely have zero effect * on these internal kernel tasks, but they probably don't hurt either. */ + __audit_free(current); return audit_alloc(tsk); } -- 2.20.1 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-03 5:02 [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() Peilin Ye @ 2022-08-03 13:16 ` Paul Moore 2022-08-03 19:28 ` Paul Moore 2022-08-03 22:23 ` [PATCH v2] " Peilin Ye 1 sibling, 1 reply; 10+ messages in thread From: Paul Moore @ 2022-08-03 13:16 UTC (permalink / raw) To: Peilin Ye Cc: Jens Axboe, Pavel Begunkov, Eric Paris, Peilin Ye, io-uring, linux-kernel, linux-audit On Wed, Aug 3, 2022 at 1:03 AM Peilin Ye <[email protected]> wrote: > > From: Peilin Ye <[email protected]> > > Currently @audit_context is allocated twice for io_uring workers: > > 1. copy_process() calls audit_alloc(); > 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which > is effectively audit_alloc()) and overwrites @audit_context, > causing: > > BUG: memory leak > unreferenced object 0xffff888144547400 (size 1024): > <...> > hex dump (first 32 bytes): > 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ > backtrace: > [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 > [<ffffffff81239e63>] copy_process+0xcd3/0x2340 > [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 > [<ffffffff81686604>] create_io_worker+0xb4/0x230 > [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 > [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 > [<ffffffff816768b3>] io_queue_async+0x113/0x180 > [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 > [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 > [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 > [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 > [<ffffffff8125a688>] get_signal+0xc18/0xf10 > [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 > [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 > [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 > [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 > > Then, > > 3. io_sq_thread() or io_wqe_worker() frees @audit_context using > audit_free(); > 4. do_exit() eventually calls audit_free() again, which is okay > because audit_free() does a NULL check. > > Free the old @audit_context first in audit_alloc_kernel(), and delete > the redundant calls to audit_free() for less confusion. > > Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") > Cc: [email protected] > Signed-off-by: Peilin Ye <[email protected]> > --- > Hi all, > > A better way to fix this memleak would probably be checking > @args->io_thread in copy_process()? Something like: > > if (args->io_thread) > retval = audit_alloc_kernel(); > else > retval = audit_alloc(); > > But I didn't want to add another if to copy_process() for this bugfix. > Please suggest, thanks! Thanks for the report and patch! I'll take a closer look at this today and get back to you. > fs/io-wq.c | 1 - > fs/io_uring.c | 2 -- > kernel/auditsc.c | 1 + > 3 files changed, 1 insertion(+), 3 deletions(-) -- paul-moore.com ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-03 13:16 ` Paul Moore @ 2022-08-03 19:28 ` Paul Moore 2022-08-03 19:39 ` Jens Axboe 2022-08-03 20:24 ` Peilin Ye 0 siblings, 2 replies; 10+ messages in thread From: Paul Moore @ 2022-08-03 19:28 UTC (permalink / raw) To: Peilin Ye Cc: Jens Axboe, Pavel Begunkov, Eric Paris, Peilin Ye, io-uring, linux-kernel, linux-audit On Wed, Aug 3, 2022 at 9:16 AM Paul Moore <[email protected]> wrote: > On Wed, Aug 3, 2022 at 1:03 AM Peilin Ye <[email protected]> wrote: > > > > Currently @audit_context is allocated twice for io_uring workers: > > > > 1. copy_process() calls audit_alloc(); > > 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which > > is effectively audit_alloc()) and overwrites @audit_context, > > causing: > > > > BUG: memory leak > > unreferenced object 0xffff888144547400 (size 1024): > > <...> > > hex dump (first 32 bytes): > > 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ > > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ > > backtrace: > > [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 > > [<ffffffff81239e63>] copy_process+0xcd3/0x2340 > > [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 > > [<ffffffff81686604>] create_io_worker+0xb4/0x230 > > [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 > > [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 > > [<ffffffff816768b3>] io_queue_async+0x113/0x180 > > [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 > > [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 > > [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 > > [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 > > [<ffffffff8125a688>] get_signal+0xc18/0xf10 > > [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 > > [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 > > [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 > > [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 > > > > Then, > > > > 3. io_sq_thread() or io_wqe_worker() frees @audit_context using > > audit_free(); > > 4. do_exit() eventually calls audit_free() again, which is okay > > because audit_free() does a NULL check. > > > > Free the old @audit_context first in audit_alloc_kernel(), and delete > > the redundant calls to audit_free() for less confusion. > > > > Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") > > Cc: [email protected] > > Signed-off-by: Peilin Ye <[email protected]> > > --- > > Hi all, > > > > A better way to fix this memleak would probably be checking > > @args->io_thread in copy_process()? Something like: > > > > if (args->io_thread) > > retval = audit_alloc_kernel(); > > else > > retval = audit_alloc(); > > > > But I didn't want to add another if to copy_process() for this bugfix. > > Please suggest, thanks! > > Thanks for the report and patch! I'll take a closer look at this > today and get back to you. I think the best solution to this is simply to remove the calls to audit_alloc_kernel() in the io_uring and io-wq code, as well as the audit_alloc_kernel() function itself. As long as create_io_thread() ends up calling copy_process to create the new kernel thread the audit_context should be allocated correctly. Peilin Ye, are you able to draft a patch to do that and give it a test? For those that may be wondering how this happened (I definitely was!), it looks like when I first started working on the LSM/audit support for io_uring it was before the v5.12-rc1 release when create_io_thread() was introduced. Prior to create_io_thread() it appears that io_uring/io-wq wasn't calling into copy_process() and thus was not getting an audit_context allocated in the kernel thread's task_struct; the solution for those original development drafts was to add a call to a new audit_alloc_kernel() which would handle the audit_context allocation. Unfortunately, I didn't notice the move to create_io_thread() during development and the redundant audit_alloc_kernel() calls remained :/ -- paul-moore.com ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-03 19:28 ` Paul Moore @ 2022-08-03 19:39 ` Jens Axboe 2022-08-03 20:24 ` Peilin Ye 1 sibling, 0 replies; 10+ messages in thread From: Jens Axboe @ 2022-08-03 19:39 UTC (permalink / raw) To: Paul Moore, Peilin Ye Cc: Pavel Begunkov, Eric Paris, Peilin Ye, io-uring, linux-kernel, linux-audit On 8/3/22 1:28 PM, Paul Moore wrote: > On Wed, Aug 3, 2022 at 9:16 AM Paul Moore <[email protected]> wrote: >> On Wed, Aug 3, 2022 at 1:03 AM Peilin Ye <[email protected]> wrote: >>> >>> Currently @audit_context is allocated twice for io_uring workers: >>> >>> 1. copy_process() calls audit_alloc(); >>> 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which >>> is effectively audit_alloc()) and overwrites @audit_context, >>> causing: >>> >>> BUG: memory leak >>> unreferenced object 0xffff888144547400 (size 1024): >>> <...> >>> hex dump (first 32 bytes): >>> 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ >>> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ >>> backtrace: >>> [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 >>> [<ffffffff81239e63>] copy_process+0xcd3/0x2340 >>> [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 >>> [<ffffffff81686604>] create_io_worker+0xb4/0x230 >>> [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 >>> [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 >>> [<ffffffff816768b3>] io_queue_async+0x113/0x180 >>> [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 >>> [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 >>> [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 >>> [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 >>> [<ffffffff8125a688>] get_signal+0xc18/0xf10 >>> [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 >>> [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 >>> [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 >>> [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 >>> >>> Then, >>> >>> 3. io_sq_thread() or io_wqe_worker() frees @audit_context using >>> audit_free(); >>> 4. do_exit() eventually calls audit_free() again, which is okay >>> because audit_free() does a NULL check. >>> >>> Free the old @audit_context first in audit_alloc_kernel(), and delete >>> the redundant calls to audit_free() for less confusion. >>> >>> Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") >>> Cc: [email protected] >>> Signed-off-by: Peilin Ye <[email protected]> >>> --- >>> Hi all, >>> >>> A better way to fix this memleak would probably be checking >>> @args->io_thread in copy_process()? Something like: >>> >>> if (args->io_thread) >>> retval = audit_alloc_kernel(); >>> else >>> retval = audit_alloc(); >>> >>> But I didn't want to add another if to copy_process() for this bugfix. >>> Please suggest, thanks! >> >> Thanks for the report and patch! I'll take a closer look at this >> today and get back to you. > > I think the best solution to this is simply to remove the calls to > audit_alloc_kernel() in the io_uring and io-wq code, as well as the > audit_alloc_kernel() function itself. As long as create_io_thread() > ends up calling copy_process to create the new kernel thread the > audit_context should be allocated correctly. Peilin Ye, are you able > to draft a patch to do that and give it a test? > > For those that may be wondering how this happened (I definitely was!), > it looks like when I first started working on the LSM/audit support > for io_uring it was before the v5.12-rc1 release when > create_io_thread() was introduced. Prior to create_io_thread() it > appears that io_uring/io-wq wasn't calling into copy_process() and > thus was not getting an audit_context allocated in the kernel thread's > task_struct; the solution for those original development drafts was to > add a call to a new audit_alloc_kernel() which would handle the > audit_context allocation. Unfortunately, I didn't notice the move to > create_io_thread() during development and the redundant > audit_alloc_kernel() calls remained :/ I agree with your analysis and suggested solution. Post the native io-wq workers create_io_thread() -> copy_process() is always used for io-wq (and sqpoll, for that matter). -- Jens Axboe ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-03 19:28 ` Paul Moore 2022-08-03 19:39 ` Jens Axboe @ 2022-08-03 20:24 ` Peilin Ye 1 sibling, 0 replies; 10+ messages in thread From: Peilin Ye @ 2022-08-03 20:24 UTC (permalink / raw) To: Paul Moore Cc: Jens Axboe, Pavel Begunkov, Eric Paris, Peilin Ye, io-uring, linux-kernel, linux-audit Hi, On Wed, Aug 03, 2022 at 03:28:19PM -0400, Paul Moore wrote: > On Wed, Aug 3, 2022 at 9:16 AM Paul Moore <[email protected]> wrote: > > On Wed, Aug 3, 2022 at 1:03 AM Peilin Ye <[email protected]> wrote: > > > Hi all, > > > > > > A better way to fix this memleak would probably be checking > > > @args->io_thread in copy_process()? Something like: > > > > > > if (args->io_thread) > > > retval = audit_alloc_kernel(); > > > else > > > retval = audit_alloc(); > > > > > > But I didn't want to add another if to copy_process() for this bugfix. > > > Please suggest, thanks! > > > > Thanks for the report and patch! I'll take a closer look at this > > today and get back to you. > > I think the best solution to this is simply to remove the calls to > audit_alloc_kernel() in the io_uring and io-wq code, as well as the > audit_alloc_kernel() function itself. As long as create_io_thread() > ends up calling copy_process to create the new kernel thread the > audit_context should be allocated correctly. Peilin Ye, are you able > to draft a patch to do that and give it a test? Sure, I will write a v2 today. Thanks for the suggestion! > For those that may be wondering how this happened (I definitely was!), > it looks like when I first started working on the LSM/audit support > for io_uring it was before the v5.12-rc1 release when > create_io_thread() was introduced. Prior to create_io_thread() it > appears that io_uring/io-wq wasn't calling into copy_process() and > thus was not getting an audit_context allocated in the kernel thread's > task_struct; the solution for those original development drafts was to > add a call to a new audit_alloc_kernel() which would handle the > audit_context allocation. Unfortunately, I didn't notice the move to > create_io_thread() during development and the redundant > audit_alloc_kernel() calls remained :/ Thanks, Peilin Ye ^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH v2] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-03 5:02 [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() Peilin Ye 2022-08-03 13:16 ` Paul Moore @ 2022-08-03 22:23 ` Peilin Ye 2022-08-04 13:51 ` Paul Moore 2022-08-04 14:36 ` Jens Axboe 1 sibling, 2 replies; 10+ messages in thread From: Peilin Ye @ 2022-08-03 22:23 UTC (permalink / raw) To: Jens Axboe, Pavel Begunkov, Paul Moore, Eric Paris Cc: Peilin Ye, io-uring, linux-kernel, linux-audit, Peilin Ye From: Peilin Ye <[email protected]> Currently @audit_context is allocated twice for io_uring workers: 1. copy_process() calls audit_alloc(); 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which is effectively audit_alloc()) and overwrites @audit_context, causing: BUG: memory leak unreferenced object 0xffff888144547400 (size 1024): <...> hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 [<ffffffff81239e63>] copy_process+0xcd3/0x2340 [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 [<ffffffff81686604>] create_io_worker+0xb4/0x230 [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 [<ffffffff816768b3>] io_queue_async+0x113/0x180 [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 [<ffffffff8125a688>] get_signal+0xc18/0xf10 [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 Then, 3. io_sq_thread() or io_wqe_worker() frees @audit_context using audit_free(); 4. do_exit() eventually calls audit_free() again, which is okay because audit_free() does a NULL check. As suggested by Paul Moore, fix it by deleting audit_alloc_kernel() and redundant audit_free() calls. Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") Suggested-by: Paul Moore <[email protected]> Cc: [email protected] Signed-off-by: Peilin Ye <[email protected]> --- Change since v1: - Delete audit_alloc_kernel() (Paul Moore) fs/io-wq.c | 3 --- fs/io_uring.c | 4 ---- include/linux/audit.h | 5 ----- kernel/auditsc.c | 25 ------------------------- 4 files changed, 37 deletions(-) diff --git a/fs/io-wq.c b/fs/io-wq.c index 824623bcf1a5..4d8a77a2a150 100644 --- a/fs/io-wq.c +++ b/fs/io-wq.c @@ -634,8 +634,6 @@ static int io_wqe_worker(void *data) snprintf(buf, sizeof(buf), "iou-wrk-%d", wq->task->pid); set_task_comm(current, buf); - audit_alloc_kernel(current); - while (!test_bit(IO_WQ_BIT_EXIT, &wq->state)) { long ret; @@ -670,7 +668,6 @@ static int io_wqe_worker(void *data) if (test_bit(IO_WQ_BIT_EXIT, &wq->state)) io_worker_handle_work(worker); - audit_free(current); io_worker_exit(worker); return 0; } diff --git a/fs/io_uring.c b/fs/io_uring.c index e8e769be9ed0..e0da0d2f71f2 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -9208,8 +9208,6 @@ static int io_sq_thread(void *data) set_cpus_allowed_ptr(current, cpu_online_mask); current->flags |= PF_NO_SETAFFINITY; - audit_alloc_kernel(current); - mutex_lock(&sqd->lock); while (1) { bool cap_entries, sqt_spin = false; @@ -9283,8 +9281,6 @@ static int io_sq_thread(void *data) io_run_task_work(); mutex_unlock(&sqd->lock); - audit_free(current); - complete(&sqd->exited); do_exit(0); } diff --git a/include/linux/audit.h b/include/linux/audit.h index cece70231138..c313df466e6e 100644 --- a/include/linux/audit.h +++ b/include/linux/audit.h @@ -287,7 +287,6 @@ static inline int audit_signal_info(int sig, struct task_struct *t) /* These are defined in auditsc.c */ /* Public API */ extern int audit_alloc(struct task_struct *task); -extern int audit_alloc_kernel(struct task_struct *task); extern void __audit_free(struct task_struct *task); extern void __audit_uring_entry(u8 op); extern void __audit_uring_exit(int success, long code); @@ -580,10 +579,6 @@ static inline int audit_alloc(struct task_struct *task) { return 0; } -static inline int audit_alloc_kernel(struct task_struct *task) -{ - return 0; -} static inline void audit_free(struct task_struct *task) { } static inline void audit_uring_entry(u8 op) diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 3a8c9d744800..dd8d9ab747c3 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -1073,31 +1073,6 @@ int audit_alloc(struct task_struct *tsk) return 0; } -/** - * audit_alloc_kernel - allocate an audit_context for a kernel task - * @tsk: the kernel task - * - * Similar to the audit_alloc() function, but intended for kernel private - * threads. Returns zero on success, negative values on failure. - */ -int audit_alloc_kernel(struct task_struct *tsk) -{ - /* - * At the moment we are just going to call into audit_alloc() to - * simplify the code, but there two things to keep in mind with this - * approach: - * - * 1. Filtering internal kernel tasks is a bit laughable in almost all - * cases, but there is at least one case where there is a benefit: - * the '-a task,never' case allows the admin to effectively disable - * task auditing at runtime. - * - * 2. The {set,clear}_task_syscall_work() ops likely have zero effect - * on these internal kernel tasks, but they probably don't hurt either. - */ - return audit_alloc(tsk); -} - static inline void audit_free_context(struct audit_context *context) { /* resetting is extra work, but it is likely just noise */ -- 2.20.1 ^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH v2] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-03 22:23 ` [PATCH v2] " Peilin Ye @ 2022-08-04 13:51 ` Paul Moore 2022-08-04 14:32 ` Jens Axboe 2022-08-04 14:36 ` Jens Axboe 1 sibling, 1 reply; 10+ messages in thread From: Paul Moore @ 2022-08-04 13:51 UTC (permalink / raw) To: Peilin Ye Cc: Jens Axboe, Pavel Begunkov, Eric Paris, Peilin Ye, io-uring, linux-kernel, linux-audit On Wed, Aug 3, 2022 at 6:24 PM Peilin Ye <[email protected]> wrote: > > From: Peilin Ye <[email protected]> > > Currently @audit_context is allocated twice for io_uring workers: > > 1. copy_process() calls audit_alloc(); > 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which > is effectively audit_alloc()) and overwrites @audit_context, > causing: > > BUG: memory leak > unreferenced object 0xffff888144547400 (size 1024): > <...> > hex dump (first 32 bytes): > 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ > backtrace: > [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 > [<ffffffff81239e63>] copy_process+0xcd3/0x2340 > [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 > [<ffffffff81686604>] create_io_worker+0xb4/0x230 > [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 > [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 > [<ffffffff816768b3>] io_queue_async+0x113/0x180 > [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 > [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 > [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 > [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 > [<ffffffff8125a688>] get_signal+0xc18/0xf10 > [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 > [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 > [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 > [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 > > Then, > > 3. io_sq_thread() or io_wqe_worker() frees @audit_context using > audit_free(); > 4. do_exit() eventually calls audit_free() again, which is okay > because audit_free() does a NULL check. > > As suggested by Paul Moore, fix it by deleting audit_alloc_kernel() and > redundant audit_free() calls. > > Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") > Suggested-by: Paul Moore <[email protected]> > Cc: [email protected] > Signed-off-by: Peilin Ye <[email protected]> > --- > Change since v1: > - Delete audit_alloc_kernel() (Paul Moore) > > fs/io-wq.c | 3 --- > fs/io_uring.c | 4 ---- > include/linux/audit.h | 5 ----- > kernel/auditsc.c | 25 ------------------------- > 4 files changed, 37 deletions(-) This looks good to me, thanks! Although it looks like the io_uring related changes will need to be applied by hand as they are pointing to the old layout under fs/ as opposed to the newer layout in io_uring/ introduced during this merge window. Jens, did you want to take this via the io_uring tree or should I take it via the audit tree? If the latter, an ACK would be appreciated, if the former my ACK is below. Acked-by: Paul Moore <[email protected]> -- paul-moore.com ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v2] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-04 13:51 ` Paul Moore @ 2022-08-04 14:32 ` Jens Axboe 2022-08-04 14:44 ` Paul Moore 0 siblings, 1 reply; 10+ messages in thread From: Jens Axboe @ 2022-08-04 14:32 UTC (permalink / raw) To: Paul Moore, Peilin Ye Cc: Pavel Begunkov, Eric Paris, Peilin Ye, io-uring, linux-kernel, linux-audit On 8/4/22 7:51 AM, Paul Moore wrote: > On Wed, Aug 3, 2022 at 6:24 PM Peilin Ye <[email protected]> wrote: >> >> From: Peilin Ye <[email protected]> >> >> Currently @audit_context is allocated twice for io_uring workers: >> >> 1. copy_process() calls audit_alloc(); >> 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which >> is effectively audit_alloc()) and overwrites @audit_context, >> causing: >> >> BUG: memory leak >> unreferenced object 0xffff888144547400 (size 1024): >> <...> >> hex dump (first 32 bytes): >> 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ >> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ >> backtrace: >> [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 >> [<ffffffff81239e63>] copy_process+0xcd3/0x2340 >> [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 >> [<ffffffff81686604>] create_io_worker+0xb4/0x230 >> [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 >> [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 >> [<ffffffff816768b3>] io_queue_async+0x113/0x180 >> [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 >> [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 >> [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 >> [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 >> [<ffffffff8125a688>] get_signal+0xc18/0xf10 >> [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 >> [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 >> [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 >> [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 >> >> Then, >> >> 3. io_sq_thread() or io_wqe_worker() frees @audit_context using >> audit_free(); >> 4. do_exit() eventually calls audit_free() again, which is okay >> because audit_free() does a NULL check. >> >> As suggested by Paul Moore, fix it by deleting audit_alloc_kernel() and >> redundant audit_free() calls. >> >> Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") >> Suggested-by: Paul Moore <[email protected]> >> Cc: [email protected] >> Signed-off-by: Peilin Ye <[email protected]> >> --- >> Change since v1: >> - Delete audit_alloc_kernel() (Paul Moore) >> >> fs/io-wq.c | 3 --- >> fs/io_uring.c | 4 ---- >> include/linux/audit.h | 5 ----- >> kernel/auditsc.c | 25 ------------------------- >> 4 files changed, 37 deletions(-) > > This looks good to me, thanks! Although it looks like the io_uring > related changes will need to be applied by hand as they are pointing > to the old layout under fs/ as opposed to the newer layout in > io_uring/ introduced during this merge window. > > Jens, did you want to take this via the io_uring tree or should I take > it via the audit tree? If the latter, an ACK would be appreciated, if > the former my ACK is below. > > Acked-by: Paul Moore <[email protected]> Probably better if I take it, since I need to massage it into the current tree anyway. We can then use this one as the base for the stable backports that are going to be required. -- Jens Axboe ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v2] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-04 14:32 ` Jens Axboe @ 2022-08-04 14:44 ` Paul Moore 0 siblings, 0 replies; 10+ messages in thread From: Paul Moore @ 2022-08-04 14:44 UTC (permalink / raw) To: Jens Axboe Cc: Peilin Ye, Pavel Begunkov, Eric Paris, Peilin Ye, io-uring, linux-kernel, linux-audit On Thu, Aug 4, 2022 at 10:32 AM Jens Axboe <[email protected]> wrote: > On 8/4/22 7:51 AM, Paul Moore wrote: > > On Wed, Aug 3, 2022 at 6:24 PM Peilin Ye <[email protected]> wrote: > >> > >> From: Peilin Ye <[email protected]> > >> > >> Currently @audit_context is allocated twice for io_uring workers: > >> > >> 1. copy_process() calls audit_alloc(); > >> 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which > >> is effectively audit_alloc()) and overwrites @audit_context, > >> causing: > >> > >> BUG: memory leak > >> unreferenced object 0xffff888144547400 (size 1024): > >> <...> > >> hex dump (first 32 bytes): > >> 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................ > >> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ > >> backtrace: > >> [<ffffffff8135cfc3>] audit_alloc+0x133/0x210 > >> [<ffffffff81239e63>] copy_process+0xcd3/0x2340 > >> [<ffffffff8123b5f3>] create_io_thread+0x63/0x90 > >> [<ffffffff81686604>] create_io_worker+0xb4/0x230 > >> [<ffffffff81686f68>] io_wqe_enqueue+0x248/0x3b0 > >> [<ffffffff8167663a>] io_queue_iowq+0xba/0x200 > >> [<ffffffff816768b3>] io_queue_async+0x113/0x180 > >> [<ffffffff816840df>] io_req_task_submit+0x18f/0x1a0 > >> [<ffffffff816841cd>] io_apoll_task_func+0xdd/0x120 > >> [<ffffffff8167d49f>] tctx_task_work+0x11f/0x570 > >> [<ffffffff81272c4e>] task_work_run+0x7e/0xc0 > >> [<ffffffff8125a688>] get_signal+0xc18/0xf10 > >> [<ffffffff8111645b>] arch_do_signal_or_restart+0x2b/0x730 > >> [<ffffffff812ea44e>] exit_to_user_mode_prepare+0x5e/0x180 > >> [<ffffffff844ae1b2>] syscall_exit_to_user_mode+0x12/0x20 > >> [<ffffffff844a7e80>] do_syscall_64+0x40/0x80 > >> > >> Then, > >> > >> 3. io_sq_thread() or io_wqe_worker() frees @audit_context using > >> audit_free(); > >> 4. do_exit() eventually calls audit_free() again, which is okay > >> because audit_free() does a NULL check. > >> > >> As suggested by Paul Moore, fix it by deleting audit_alloc_kernel() and > >> redundant audit_free() calls. > >> > >> Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") > >> Suggested-by: Paul Moore <[email protected]> > >> Cc: [email protected] > >> Signed-off-by: Peilin Ye <[email protected]> > >> --- > >> Change since v1: > >> - Delete audit_alloc_kernel() (Paul Moore) > >> > >> fs/io-wq.c | 3 --- > >> fs/io_uring.c | 4 ---- > >> include/linux/audit.h | 5 ----- > >> kernel/auditsc.c | 25 ------------------------- > >> 4 files changed, 37 deletions(-) > > > > This looks good to me, thanks! Although it looks like the io_uring > > related changes will need to be applied by hand as they are pointing > > to the old layout under fs/ as opposed to the newer layout in > > io_uring/ introduced during this merge window. > > > > Jens, did you want to take this via the io_uring tree or should I take > > it via the audit tree? If the latter, an ACK would be appreciated, if > > the former my ACK is below. > > > > Acked-by: Paul Moore <[email protected]> > > Probably better if I take it, since I need to massage it into the > current tree anyway. We can then use this one as the base for the stable > backports that are going to be required. Sounds good to me, thanks everyone. -- paul-moore.com ^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v2] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() 2022-08-03 22:23 ` [PATCH v2] " Peilin Ye 2022-08-04 13:51 ` Paul Moore @ 2022-08-04 14:36 ` Jens Axboe 1 sibling, 0 replies; 10+ messages in thread From: Jens Axboe @ 2022-08-04 14:36 UTC (permalink / raw) To: paul, yepeilin.cs, eparis, asml.silence Cc: linux-audit, peilin.ye, io-uring, linux-kernel On Wed, 3 Aug 2022 15:23:43 -0700, Peilin Ye wrote: > From: Peilin Ye <[email protected]> > > Currently @audit_context is allocated twice for io_uring workers: > > 1. copy_process() calls audit_alloc(); > 2. io_sq_thread() or io_wqe_worker() calls audit_alloc_kernel() (which > is effectively audit_alloc()) and overwrites @audit_context, > causing: > > [...] Applied, thanks! [1/1] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() commit: f482aa98652795846cc55da98ebe331eb74f3d0b Best regards, -- Jens Axboe ^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2022-08-04 14:45 UTC | newest] Thread overview: 10+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2022-08-03 5:02 [PATCH] audit, io_uring, io-wq: Fix memory leak in io_sq_thread() and io_wqe_worker() Peilin Ye 2022-08-03 13:16 ` Paul Moore 2022-08-03 19:28 ` Paul Moore 2022-08-03 19:39 ` Jens Axboe 2022-08-03 20:24 ` Peilin Ye 2022-08-03 22:23 ` [PATCH v2] " Peilin Ye 2022-08-04 13:51 ` Paul Moore 2022-08-04 14:32 ` Jens Axboe 2022-08-04 14:44 ` Paul Moore 2022-08-04 14:36 ` Jens Axboe
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox