| Message ID | 20220726072451.142427-1-wangkefeng.wang@huawei.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | mm: add warning if __vm_enough_memory fails | expand |
On 26.07.22 09:24, Kefeng Wang wrote: > If a process has no enough memory to allocate a new virtual mapping, we > may meet kinds of error, eg, fork cannot allocate memory, SIGBUS error > in shmem, but it is difficult to confirm them, let's add some debug > information to easy to check this scenario if __vm_enough_memory fails. > > Reported-by: Yongqiang Liu <liuyongqiang13@huawei.com> > Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com> > --- > mm/util.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/mm/util.c b/mm/util.c > index 1266a33a49ea..19bfff8a0ad6 100644 > --- a/mm/util.c > +++ b/mm/util.c > @@ -1020,6 +1020,8 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin) > if (percpu_counter_read_positive(&vm_committed_as) < allowed) > return 0; > error: > + pr_warn("%s: pid: %d, comm: %s, no enough memory for the allocation\n", > + __func__, current->pid, current->comm); > vm_unacct_memory(pages); > > return -ENOMEM; Users can easily spam the kernel log, no? Maybe at least ratelimit.
On 2022/7/26 21:45, David Hildenbrand wrote: > On 26.07.22 09:24, Kefeng Wang wrote: >> If a process has no enough memory to allocate a new virtual mapping, we >> may meet kinds of error, eg, fork cannot allocate memory, SIGBUS error >> in shmem, but it is difficult to confirm them, let's add some debug >> information to easy to check this scenario if __vm_enough_memory fails. >> >> Reported-by: Yongqiang Liu <liuyongqiang13@huawei.com> >> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com> >> --- >> mm/util.c | 2 ++ >> 1 file changed, 2 insertions(+) >> >> diff --git a/mm/util.c b/mm/util.c >> index 1266a33a49ea..19bfff8a0ad6 100644 >> --- a/mm/util.c >> +++ b/mm/util.c >> @@ -1020,6 +1020,8 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin) >> if (percpu_counter_read_positive(&vm_committed_as) < allowed) >> return 0; >> error: >> + pr_warn("%s: pid: %d, comm: %s, no enough memory for the allocation\n", >> + __func__, current->pid, current->comm); >> vm_unacct_memory(pages); >> >> return -ENOMEM; > Users can easily spam the kernel log, no? Maybe at least ratelimit. pr_warn_ratelimited is better, will update, thanks. >
diff --git a/mm/util.c b/mm/util.c index 1266a33a49ea..19bfff8a0ad6 100644 --- a/mm/util.c +++ b/mm/util.c @@ -1020,6 +1020,8 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin) if (percpu_counter_read_positive(&vm_committed_as) < allowed) return 0; error: + pr_warn("%s: pid: %d, comm: %s, no enough memory for the allocation\n", + __func__, current->pid, current->comm); vm_unacct_memory(pages); return -ENOMEM;
If a process has no enough memory to allocate a new virtual mapping, we may meet kinds of error, eg, fork cannot allocate memory, SIGBUS error in shmem, but it is difficult to confirm them, let's add some debug information to easy to check this scenario if __vm_enough_memory fails. Reported-by: Yongqiang Liu <liuyongqiang13@huawei.com> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com> --- mm/util.c | 2 ++ 1 file changed, 2 insertions(+)