[SRU][Xenial][PATCH] Revert "mm: check that mm is still valid in madvise()"

[Ian May]

BugLink: https://bugs.launchpad.net/bugs/1913853

This reverts commit d8f6fb767b2fed2e0ca85d85acfa9a0669e34f48.

Several users have reported system hangs.

A deadlock was also confirmed with autotest ubuntu_stress_smoke_test.

Call Trace:
  schedule
  rwsem_down_read_failed
  call_rwsem_down_read_failed
  down_read
  SyS_madvise
  __do_page_fault
  trace_do_page_fault
  entry_SYSCALL_64_fastpath

Removing this patch resolves autotest deadlock

Signed-off-by: Ian May <ian.may at canonical.com>
---
 mm/madvise.c | 18 ------------------
 1 file changed, 18 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index cf2b217a647a..f548c66154ee 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -20,7 +20,6 @@
 #include <linux/backing-dev.h>
 #include <linux/swap.h>
 #include <linux/swapops.h>
-#include <linux/sched.h>
 
 /*
  * Any behaviour which results in changes to the vma->vm_flags needs to
@@ -492,23 +491,6 @@ SYSCALL_DEFINE3(madvise, unsigned long, start, size_t, len_in, int, behavior)
 	write = madvise_need_mmap_write(behavior);
 	if (write)
 		down_write(&current->mm->mmap_sem);
-
-		/*
-		 * We may have stolen the mm from another process
-		 * that is undergoing core dumping.
-		 *
-		 * Right now that's io_ring, in the future it may
-		 * be remote process management and not "current"
-		 * at all.
-		 *
-		 * We need to fix core dumping to not do this,
-		 * but for now we have the mmget_still_valid()
-		 * model.
-		 */
-		if (!mmget_still_valid(current->mm)) {
-			up_write(&current->mm->mmap_sem);
-			return -EINTR;
-		}
 	else
 		down_read(&current->mm->mmap_sem);
 
-- 
2.25.1