[PATCH] mm: use VM_BUG_ON*() helpers to dump more debugging info

[PATCH] mm: use VM_BUG_ON*() helpers to dump more debugging info

[qiwuchen55]

From: chenqiwu <chenqiwu@xiaomi.com>

This patch use VM_BUG_ON*() helpers instead of simple BUG_ON()
in some of the main mm codes. If CONFIG_DEBUG_VM is set, we can
get more debugging information when the bug is hit.

Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
---
 mm/memory.c   | 25 +++++++++++++------------
 mm/mmap.c     |  8 ++++----
 mm/rmap.c     | 10 +++++-----
 mm/swapfile.c | 16 ++++++++--------
 mm/vmscan.c   |  6 +++---
 5 files changed, 33 insertions(+), 32 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index 586271f..082472f 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -912,7 +912,7 @@ static inline int copy_pud_range(struct mm_struct *dst_mm, struct mm_struct *src
 		if (pud_trans_huge(*src_pud) || pud_devmap(*src_pud)) {
 			int err;
 
-			VM_BUG_ON_VMA(next-addr != HPAGE_PUD_SIZE, vma);
+			VM_BUG_ON_VMA(next - addr != HPAGE_PUD_SIZE, vma);
 			err = copy_huge_pud(dst_mm, src_mm,
 					    dst_pud, src_pud, addr, vma);
 			if (err == -ENOMEM)
@@ -1245,7 +1245,7 @@ void unmap_page_range(struct mmu_gather *tlb,
 	pgd_t *pgd;
 	unsigned long next;
 
-	BUG_ON(addr >= end);
+	VM_BUG_ON(addr >= end);
 	tlb_start_vma(tlb, vma);
 	pgd = pgd_offset(vma->vm_mm, addr);
 	do {
@@ -1507,8 +1507,8 @@ int vm_insert_page(struct vm_area_struct *vma, unsigned long addr,
 	if (!page_count(page))
 		return -EINVAL;
 	if (!(vma->vm_flags & VM_MIXEDMAP)) {
-		BUG_ON(down_read_trylock(&vma->vm_mm->mmap_sem));
-		BUG_ON(vma->vm_flags & VM_PFNMAP);
+		VM_BUG_ON_VMA(down_read_trylock(&vma->vm_mm->mmap_sem), vma);
+		VM_BUG_ON_VMA(vma->vm_flags & VM_PFNMAP, vma);
 		vma->vm_flags |= VM_MIXEDMAP;
 	}
 	return insert_page(vma, addr, page, vma->vm_page_prot);
@@ -1679,11 +1679,12 @@ vm_fault_t vmf_insert_pfn_prot(struct vm_area_struct *vma, unsigned long addr,
 	 * consistency in testing and feature parity among all, so we should
 	 * try to keep these invariants in place for everybody.
 	 */
-	BUG_ON(!(vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP)));
-	BUG_ON((vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP)) ==
-						(VM_PFNMAP|VM_MIXEDMAP));
-	BUG_ON((vma->vm_flags & VM_PFNMAP) && is_cow_mapping(vma->vm_flags));
-	BUG_ON((vma->vm_flags & VM_MIXEDMAP) && pfn_valid(pfn));
+	VM_BUG_ON_VMA(!(vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP)), vma);
+	VM_BUG_ON_VMA((vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP)) ==
+					(VM_PFNMAP|VM_MIXEDMAP), vma);
+	VM_BUG_ON_VMA((vma->vm_flags & VM_PFNMAP) &&
+					is_cow_mapping(vma->vm_flags), vma);
+	VM_BUG_ON_VMA((vma->vm_flags & VM_MIXEDMAP) && pfn_valid(pfn), vma);
 
 	if (addr < vma->vm_start || addr >= vma->vm_end)
 		return VM_FAULT_SIGBUS;
@@ -1987,7 +1988,7 @@ int remap_pfn_range(struct vm_area_struct *vma, unsigned long addr,
 
 	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
 
-	BUG_ON(addr >= end);
+	VM_BUG_ON(addr >= end);
 	pfn -= addr >> PAGE_SHIFT;
 	pgd = pgd_offset(mm, addr);
 	flush_cache_range(vma, addr, end);
@@ -2075,7 +2076,7 @@ static int apply_to_pte_range(struct mm_struct *mm, pmd_t *pmd,
 			pte_offset_map_lock(mm, pmd, addr, &ptl);
 	}
 
-	BUG_ON(pmd_huge(*pmd));
+	VM_BUG_ON(pmd_huge(*pmd));
 
 	arch_enter_lazy_mmu_mode();
 
@@ -2102,7 +2103,7 @@ static int apply_to_pmd_range(struct mm_struct *mm, pud_t *pud,
 	unsigned long next;
 	int err = 0;
 
-	BUG_ON(pud_huge(*pud));
+	VM_BUG_ON(pud_huge(*pud));
 
 	if (create) {
 		pmd = pmd_alloc(mm, pud, addr);
diff --git a/mm/mmap.c b/mm/mmap.c
index 94ae183..6a0d8ad 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -3192,7 +3192,7 @@ int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma)
 	 * Similarly in do_mmap_pgoff and in do_brk.
 	 */
 	if (vma_is_anonymous(vma)) {
-		BUG_ON(vma->anon_vma);
+		VM_BUG_ON_VMA(vma->anon_vma, vma);
 		vma->vm_pgoff = vma->vm_start >> PAGE_SHIFT;
 	}
 
@@ -3550,7 +3550,7 @@ int mm_take_all_locks(struct mm_struct *mm)
 	struct vm_area_struct *vma;
 	struct anon_vma_chain *avc;
 
-	BUG_ON(down_read_trylock(&mm->mmap_sem));
+	VM_BUG_ON_MM(down_read_trylock(&mm->mmap_sem), mm);
 
 	mutex_lock(&mm_all_locks_mutex);
 
@@ -3630,8 +3630,8 @@ void mm_drop_all_locks(struct mm_struct *mm)
 	struct vm_area_struct *vma;
 	struct anon_vma_chain *avc;
 
-	BUG_ON(down_read_trylock(&mm->mmap_sem));
-	BUG_ON(!mutex_is_locked(&mm_all_locks_mutex));
+	VM_BUG_ON_MM(down_read_trylock(&mm->mmap_sem), mm);
+	VM_BUG_ON(!mutex_is_locked(&mm_all_locks_mutex));
 
 	for (vma = mm->mmap; vma; vma = vma->vm_next) {
 		if (vma->anon_vma)
diff --git a/mm/rmap.c b/mm/rmap.c
index 2df75a1..13ed1ac 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -999,7 +999,7 @@ int page_mkclean(struct page *page)
 		.invalid_vma = invalid_mkclean_vma,
 	};
 
-	BUG_ON(!PageLocked(page));
+	VM_BUG_ON_PAGE(!PageLocked(page), page);
 
 	if (!page_mapped(page))
 		return 0;
@@ -1054,7 +1054,7 @@ static void __page_set_anon_rmap(struct page *page,
 {
 	struct anon_vma *anon_vma = vma->anon_vma;
 
-	BUG_ON(!anon_vma);
+	VM_BUG_ON_VMA(!anon_vma, vma);
 
 	if (PageAnon(page))
 		return;
@@ -1965,8 +1965,8 @@ void hugepage_add_anon_rmap(struct page *page,
 	struct anon_vma *anon_vma = vma->anon_vma;
 	int first;
 
-	BUG_ON(!PageLocked(page));
-	BUG_ON(!anon_vma);
+	VM_BUG_ON_PAGE(!PageLocked(page), page);
+	VM_BUG_ON_VMA(!anon_vma, vma);
 	/* address might be in next vma when migration races vma_adjust */
 	first = atomic_inc_and_test(compound_mapcount_ptr(page));
 	if (first)
@@ -1976,7 +1976,7 @@ void hugepage_add_anon_rmap(struct page *page,
 void hugepage_add_new_anon_rmap(struct page *page,
 			struct vm_area_struct *vma, unsigned long address)
 {
-	BUG_ON(address < vma->vm_start || address >= vma->vm_end);
+	VM_BUG_ON_VMA(address < vma->vm_start || address >= vma->vm_end, vma);
 	atomic_set(compound_mapcount_ptr(page), 0);
 	if (hpage_pincount_available(page))
 		atomic_set(compound_pincount_ptr(page), 0);
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 273a923..986ae8d 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -2326,7 +2326,7 @@ static void destroy_swap_extents(struct swap_info_struct *sis)
 
 	if (parent) {
 		se = rb_entry(parent, struct swap_extent, rb_node);
-		BUG_ON(se->start_page + se->nr_pages != start_page);
+		VM_BUG_ON(se->start_page + se->nr_pages != start_page);
 		if (se->start_block + se->nr_pages == start_block) {
 			/* Merge it */
 			se->nr_pages += nr_pages;
@@ -2528,7 +2528,7 @@ bool has_usable_swap(void)
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	BUG_ON(!current->mm);
+	VM_BUG_ON(!current->mm);
 
 	pathname = getname(specialfile);
 	if (IS_ERR(pathname))
@@ -3586,7 +3586,7 @@ int add_swap_count_continuation(swp_entry_t entry, gfp_t gfp_mask)
 	 * but it does always reset its private field.
 	 */
 	if (!page_private(head)) {
-		BUG_ON(count & COUNT_CONTINUED);
+		VM_BUG_ON(count & COUNT_CONTINUED);
 		INIT_LIST_HEAD(&head->lru);
 		set_page_private(head, SWP_CONTINUED);
 		si->flags |= SWP_CONTINUED;
@@ -3647,7 +3647,7 @@ static bool swap_count_continued(struct swap_info_struct *si,
 
 	head = vmalloc_to_page(si->swap_map + offset);
 	if (page_private(head) != SWP_CONTINUED) {
-		BUG_ON(count & COUNT_CONTINUED);
+		VM_BUG_ON_PAGE(count & COUNT_CONTINUED, head);
 		return false;		/* need to add count continuation */
 	}
 
@@ -3666,7 +3666,7 @@ static bool swap_count_continued(struct swap_info_struct *si,
 		while (*map == (SWAP_CONT_MAX | COUNT_CONTINUED)) {
 			kunmap_atomic(map);
 			page = list_entry(page->lru.next, struct page, lru);
-			BUG_ON(page == head);
+			VM_BUG_ON_PAGE(page == head, page);
 			map = kmap_atomic(page) + offset;
 		}
 		if (*map == SWAP_CONT_MAX) {
@@ -3694,14 +3694,14 @@ static bool swap_count_continued(struct swap_info_struct *si,
 		/*
 		 * Think of how you subtract 1 from 1000
 		 */
-		BUG_ON(count != COUNT_CONTINUED);
+		VM_BUG_ON(count != COUNT_CONTINUED);
 		while (*map == COUNT_CONTINUED) {
 			kunmap_atomic(map);
 			page = list_entry(page->lru.next, struct page, lru);
-			BUG_ON(page == head);
+			VM_BUG_ON_PAGE(page == head, page);
 			map = kmap_atomic(page) + offset;
 		}
-		BUG_ON(*map == 0);
+		VM_BUG_ON(*map == 0);
 		*map -= 1;
 		if (*map == 0)
 			count = 0;
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 2e8e690..1b4bc87 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -231,7 +231,7 @@ static void unregister_memcg_shrinker(struct shrinker *shrinker)
 {
 	int id = shrinker->id;
 
-	BUG_ON(id < 0);
+	VM_BUG_ON(id < 0);
 
 	down_write(&shrinker_rwsem);
 	idr_remove(&shrinker_idr, id);
@@ -854,8 +854,8 @@ static int __remove_mapping(struct address_space *mapping, struct page *page,
 	unsigned long flags;
 	int refcount;
 
-	BUG_ON(!PageLocked(page));
-	BUG_ON(mapping != page_mapping(page));
+	VM_BUG_ON_PAGE(!PageLocked(page), page);
+	VM_BUG_ON_PAGE(mapping != page_mapping(page), page);
 
 	xa_lock_irqsave(&mapping->i_pages, flags);
 	/*
-- 
1.9.1

Re: [PATCH] mm: use VM_BUG_ON*() helpers to dump more debugging info

[David Hildenbrand]

On 06.04.20 05:55, qiwuchen55@gmail.com wrote:
> From: chenqiwu <chenqiwu@xiaomi.com>
> 
> This patch use VM_BUG_ON*() helpers instead of simple BUG_ON()
> in some of the main mm codes. If CONFIG_DEBUG_VM is set, we can
> get more debugging information when the bug is hit.
> 

The "issue" in this context of VM_BUG_ON*() is that, without
CONFIG_DEBUG_VM,  there won't really be any runtime checks anymore,
meaning a production system would not stop and BUG_ON() (which would be
disruptive, but there is a chance to debug this), instead it would
happily continue to run, eventually messing up something else.

This is a clear change introduced in this series.

My gut feeling is that we want to convert this on a per-case basis instead.

> Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
> ---
>  mm/memory.c   | 25 +++++++++++++------------
>  mm/mmap.c     |  8 ++++----
>  mm/rmap.c     | 10 +++++-----
>  mm/swapfile.c | 16 ++++++++--------
>  mm/vmscan.c   |  6 +++---
>  5 files changed, 33 insertions(+), 32 deletions(-)
> 
> diff --git a/mm/memory.c b/mm/memory.c
> index 586271f..082472f 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -912,7 +912,7 @@ static inline int copy_pud_range(struct mm_struct *dst_mm, struct mm_struct *src
>  		if (pud_trans_huge(*src_pud) || pud_devmap(*src_pud)) {
>  			int err;
>  
> -			VM_BUG_ON_VMA(next-addr != HPAGE_PUD_SIZE, vma);
> +			VM_BUG_ON_VMA(next - addr != HPAGE_PUD_SIZE, vma);

unrelated change.


-- 
Thanks,

David / dhildenb

Re: [PATCH] mm: use VM_BUG_ON*() helpers to dump more debugging info

[Michal Hocko]

On Mon 06-04-20 09:40:05, David Hildenbrand wrote:
> On 06.04.20 05:55, qiwuchen55@gmail.com wrote:
> > From: chenqiwu <chenqiwu@xiaomi.com>
> > 
> > This patch use VM_BUG_ON*() helpers instead of simple BUG_ON()
> > in some of the main mm codes. If CONFIG_DEBUG_VM is set, we can
> > get more debugging information when the bug is hit.
> > 
> 
> The "issue" in this context of VM_BUG_ON*() is that, without
> CONFIG_DEBUG_VM,  there won't really be any runtime checks anymore,
> meaning a production system would not stop and BUG_ON() (which would be
> disruptive, but there is a chance to debug this), instead it would
> happily continue to run, eventually messing up something else.
> 
> This is a clear change introduced in this series.
> 
> My gut feeling is that we want to convert this on a per-case basis instead.

This definitely should be done on per-case basis. Many of those bug ons
are historical and they wouldn't be allowed these days. So I would
definitely recommend going through each of them and re-evaluate whether
they are really needed and whether they serve any purpose. A rule of
thumb is that if we can handle the situation more gracefully then the
BUG_ON should be simply dropped. If a verbatim output in a debugging
mode would serve a good purpose then WARN{_ONECE} would serve a good
purpose and if a crash is would help debugging in a DEBUG_VM mode then
use the VM_BUG_ON instead.
-- 
Michal Hocko
SUSE Labs