[v3] kasan: vmalloc: Fixes for the percpu allocator and vrealloc

[PATCH v3 3/3] kasan: Unpoison vms[area] addresses with a common tag

Posted by Maciej Wieczor-Retman 2 weeks ago

From: Maciej Wieczor-Retman <maciej.wieczor-retman@intel.com>

A KASAN tag mismatch, possibly causing a kernel panic, can be observed
on systems with a tag-based KASAN enabled and with multiple NUMA nodes.
It was reported on arm64 and reproduced on x86. It can be explained in
the following points:

	1. There can be more than one virtual memory chunk.
	2. Chunk's base address has a tag.
	3. The base address points at the first chunk and thus inherits
	   the tag of the first chunk.
	4. The subsequent chunks will be accessed with the tag from the
	   first chunk.
	5. Thus, the subsequent chunks need to have their tag set to
	   match that of the first chunk.

Use the new vmalloc flag that disables random tag assignment in
__kasan_unpoison_vmalloc() - pass the same random tag to all the
vm_structs by tagging the pointers before they go inside
__kasan_unpoison_vmalloc(). Assigning a common tag resolves the pcpu
chunk address mismatch.

Fixes: 1d96320f8d53 ("kasan, vmalloc: add vmalloc tagging for SW_TAGS")
Cc: <stable@vger.kernel.org> # 6.1+
Signed-off-by: Maciej Wieczor-Retman <maciej.wieczor-retman@intel.com>
---
Changelog v3:
- Redo the patch by using a flag instead of a new argument in
  __kasan_unpoison_vmalloc() (Andrey Konovalov)

Changelog v2:
- Revise the whole patch to match the fixed refactorization from the
  first patch.

Changelog v1:
- Rewrite the patch message to point at the user impact of the issue.
- Move helper to common.c so it can be compiled in all KASAN modes.

 mm/kasan/common.c | 23 ++++++++++++++++++++---
 1 file changed, 20 insertions(+), 3 deletions(-)

diff --git a/mm/kasan/common.c b/mm/kasan/common.c
index 1ed6289d471a..496bb2c56911 100644
--- a/mm/kasan/common.c
+++ b/mm/kasan/common.c
@@ -591,11 +591,28 @@ void __kasan_unpoison_vmap_areas(struct vm_struct **vms, int nr_vms,
 	unsigned long size;
 	void *addr;
 	int area;
+	u8 tag;
+
+	/*
+	 * If KASAN_VMALLOC_KEEP_TAG was set at this point, all vms[] pointers
+	 * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
+	 * KASAN checks down the line.
+	 */
+	if (flags & KASAN_VMALLOC_KEEP_TAG) {
+		pr_warn("KASAN_VMALLOC_KEEP_TAG flag shouldn't be already set!\n");
+		return;
+	}
+
+	size = vms[0]->size;
+	addr = vms[0]->addr;
+	vms[0]->addr = __kasan_unpoison_vmalloc(addr, size, flags);
+	tag = get_tag(vms[0]->addr);
 
-	for (area = 0 ; area < nr_vms ; area++) {
+	for (area = 1 ; area < nr_vms ; area++) {
 		size = vms[area]->size;
-		addr = vms[area]->addr;
-		vms[area]->addr = __kasan_unpoison_vmalloc(addr, size, flags);
+		addr = set_tag(vms[area]->addr, tag);
+		vms[area]->addr =
+			__kasan_unpoison_vmalloc(addr, size, flags | KASAN_VMALLOC_KEEP_TAG);
 	}
 }
 #endif
-- 
2.52.0

Re: [PATCH v3 3/3] kasan: Unpoison vms[area] addresses with a common tag

Posted by Andrey Konovalov 2 weeks ago

On Thu, Dec 4, 2025 at 8:00 PM Maciej Wieczor-Retman
<m.wieczorretman@pm.me> wrote:
>
> From: Maciej Wieczor-Retman <maciej.wieczor-retman@intel.com>
>
> A KASAN tag mismatch, possibly causing a kernel panic, can be observed
> on systems with a tag-based KASAN enabled and with multiple NUMA nodes.
> It was reported on arm64 and reproduced on x86. It can be explained in
> the following points:
>
>         1. There can be more than one virtual memory chunk.
>         2. Chunk's base address has a tag.
>         3. The base address points at the first chunk and thus inherits
>            the tag of the first chunk.
>         4. The subsequent chunks will be accessed with the tag from the
>            first chunk.
>         5. Thus, the subsequent chunks need to have their tag set to
>            match that of the first chunk.
>
> Use the new vmalloc flag that disables random tag assignment in
> __kasan_unpoison_vmalloc() - pass the same random tag to all the
> vm_structs by tagging the pointers before they go inside
> __kasan_unpoison_vmalloc(). Assigning a common tag resolves the pcpu
> chunk address mismatch.
>
> Fixes: 1d96320f8d53 ("kasan, vmalloc: add vmalloc tagging for SW_TAGS")
> Cc: <stable@vger.kernel.org> # 6.1+
> Signed-off-by: Maciej Wieczor-Retman <maciej.wieczor-retman@intel.com>
> ---
> Changelog v3:
> - Redo the patch by using a flag instead of a new argument in
>   __kasan_unpoison_vmalloc() (Andrey Konovalov)
>
> Changelog v2:
> - Revise the whole patch to match the fixed refactorization from the
>   first patch.
>
> Changelog v1:
> - Rewrite the patch message to point at the user impact of the issue.
> - Move helper to common.c so it can be compiled in all KASAN modes.
>
>  mm/kasan/common.c | 23 ++++++++++++++++++++---
>  1 file changed, 20 insertions(+), 3 deletions(-)
>
> diff --git a/mm/kasan/common.c b/mm/kasan/common.c
> index 1ed6289d471a..496bb2c56911 100644
> --- a/mm/kasan/common.c
> +++ b/mm/kasan/common.c
> @@ -591,11 +591,28 @@ void __kasan_unpoison_vmap_areas(struct vm_struct **vms, int nr_vms,
>         unsigned long size;
>         void *addr;
>         int area;
> +       u8 tag;
> +
> +       /*
> +        * If KASAN_VMALLOC_KEEP_TAG was set at this point, all vms[] pointers
> +        * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
> +        * KASAN checks down the line.
> +        */
> +       if (flags & KASAN_VMALLOC_KEEP_TAG) {

I think we can do a WARN_ON() here: passing KASAN_VMALLOC_KEEP_TAG to
this function would be a bug in KASAN annotations and thus a kernel
bug. Therefore, printing a WARNING seems justified.

> +               pr_warn("KASAN_VMALLOC_KEEP_TAG flag shouldn't be already set!\n");
> +               return;
> +       }
> +
> +       size = vms[0]->size;
> +       addr = vms[0]->addr;
> +       vms[0]->addr = __kasan_unpoison_vmalloc(addr, size, flags);
> +       tag = get_tag(vms[0]->addr);
>
> -       for (area = 0 ; area < nr_vms ; area++) {
> +       for (area = 1 ; area < nr_vms ; area++) {
>                 size = vms[area]->size;
> -               addr = vms[area]->addr;
> -               vms[area]->addr = __kasan_unpoison_vmalloc(addr, size, flags);
> +               addr = set_tag(vms[area]->addr, tag);
> +               vms[area]->addr =
> +                       __kasan_unpoison_vmalloc(addr, size, flags | KASAN_VMALLOC_KEEP_TAG);
>         }
>  }
>  #endif
> --
> 2.52.0
>

With WARN_ON():

Reviewed-by: Andrey Konovalov <andreyknvl@gmail.com>

Thank you!

Re: [PATCH v3 3/3] kasan: Unpoison vms[area] addresses with a common tag

Posted by Andrew Morton 2 weeks ago

On Fri, 5 Dec 2025 02:09:06 +0100 Andrey Konovalov <andreyknvl@gmail.com> wrote:

> > --- a/mm/kasan/common.c
> > +++ b/mm/kasan/common.c
> > @@ -591,11 +591,28 @@ void __kasan_unpoison_vmap_areas(struct vm_struct **vms, int nr_vms,
> >         unsigned long size;
> >         void *addr;
> >         int area;
> > +       u8 tag;
> > +
> > +       /*
> > +        * If KASAN_VMALLOC_KEEP_TAG was set at this point, all vms[] pointers
> > +        * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
> > +        * KASAN checks down the line.
> > +        */
> > +       if (flags & KASAN_VMALLOC_KEEP_TAG) {
> 
> I think we can do a WARN_ON() here: passing KASAN_VMALLOC_KEEP_TAG to
> this function would be a bug in KASAN annotations and thus a kernel
> bug. Therefore, printing a WARNING seems justified.

This?

--- a/mm/kasan/common.c~kasan-unpoison-vms-addresses-with-a-common-tag-fix
+++ a/mm/kasan/common.c
@@ -598,7 +598,7 @@ void __kasan_unpoison_vmap_areas(struct
 	 * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
 	 * KASAN checks down the line.
 	 */
-	if (flags & KASAN_VMALLOC_KEEP_TAG) {
+	if (WARN_ON_ONCE(flags & KASAN_VMALLOC_KEEP_TAG)) {
 		pr_warn("KASAN_VMALLOC_KEEP_TAG flag shouldn't be already set!\n");
 		return;
 	}
_

Re: [PATCH v3 3/3] kasan: Unpoison vms[area] addresses with a common tag

Posted by Andrey Konovalov 2 weeks ago

On Fri, Dec 5, 2025 at 4:22 AM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Fri, 5 Dec 2025 02:09:06 +0100 Andrey Konovalov <andreyknvl@gmail.com> wrote:
>
> > > --- a/mm/kasan/common.c
> > > +++ b/mm/kasan/common.c
> > > @@ -591,11 +591,28 @@ void __kasan_unpoison_vmap_areas(struct vm_struct **vms, int nr_vms,
> > >         unsigned long size;
> > >         void *addr;
> > >         int area;
> > > +       u8 tag;
> > > +
> > > +       /*
> > > +        * If KASAN_VMALLOC_KEEP_TAG was set at this point, all vms[] pointers
> > > +        * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
> > > +        * KASAN checks down the line.
> > > +        */
> > > +       if (flags & KASAN_VMALLOC_KEEP_TAG) {
> >
> > I think we can do a WARN_ON() here: passing KASAN_VMALLOC_KEEP_TAG to
> > this function would be a bug in KASAN annotations and thus a kernel
> > bug. Therefore, printing a WARNING seems justified.
>
> This?
>
> --- a/mm/kasan/common.c~kasan-unpoison-vms-addresses-with-a-common-tag-fix
> +++ a/mm/kasan/common.c
> @@ -598,7 +598,7 @@ void __kasan_unpoison_vmap_areas(struct
>          * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
>          * KASAN checks down the line.
>          */
> -       if (flags & KASAN_VMALLOC_KEEP_TAG) {
> +       if (WARN_ON_ONCE(flags & KASAN_VMALLOC_KEEP_TAG)) {
>                 pr_warn("KASAN_VMALLOC_KEEP_TAG flag shouldn't be already set!\n");
>                 return;
>         }
> _
>

Can also drop pr_warn(), but this is fine too. Thanks!

Re: [PATCH v3 3/3] kasan: Unpoison vms[area] addresses with a common tag

Posted by Maciej Wieczór-Retman 2 weeks ago

Thanks for checking the patches out, do you want me to send v4 with this
correction or is it redundant now that Andrew already wrote it?

Kind regards
Maciej Wieczór-Retman

On 2025-12-05 at 04:38:27 +0100, Andrey Konovalov wrote:
>On Fri, Dec 5, 2025 at 4:22 AM Andrew Morton <akpm@linux-foundation.org> wrote:
>>
>> On Fri, 5 Dec 2025 02:09:06 +0100 Andrey Konovalov <andreyknvl@gmail.com> wrote:
>>
>> > > --- a/mm/kasan/common.c
>> > > +++ b/mm/kasan/common.c
>> > > @@ -591,11 +591,28 @@ void __kasan_unpoison_vmap_areas(struct vm_struct **vms, int nr_vms,
>> > >         unsigned long size;
>> > >         void *addr;
>> > >         int area;
>> > > +       u8 tag;
>> > > +
>> > > +       /*
>> > > +        * If KASAN_VMALLOC_KEEP_TAG was set at this point, all vms[] pointers
>> > > +        * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
>> > > +        * KASAN checks down the line.
>> > > +        */
>> > > +       if (flags & KASAN_VMALLOC_KEEP_TAG) {
>> >
>> > I think we can do a WARN_ON() here: passing KASAN_VMALLOC_KEEP_TAG to
>> > this function would be a bug in KASAN annotations and thus a kernel
>> > bug. Therefore, printing a WARNING seems justified.
>>
>> This?
>>
>> --- a/mm/kasan/common.c~kasan-unpoison-vms-addresses-with-a-common-tag-fix
>> +++ a/mm/kasan/common.c
>> @@ -598,7 +598,7 @@ void __kasan_unpoison_vmap_areas(struct
>>          * would be unpoisoned with the KASAN_TAG_KERNEL which would disable
>>          * KASAN checks down the line.
>>          */
>> -       if (flags & KASAN_VMALLOC_KEEP_TAG) {
>> +       if (WARN_ON_ONCE(flags & KASAN_VMALLOC_KEEP_TAG)) {
>>                 pr_warn("KASAN_VMALLOC_KEEP_TAG flag shouldn't be already set!\n");
>>                 return;
>>         }
>> _
>>
>
>Can also drop pr_warn(), but this is fine too. Thanks!

Re: [PATCH v3 3/3] kasan: Unpoison vms[area] addresses with a common tag

Posted by Andrey Konovalov 2 weeks ago

On Fri, Dec 5, 2025 at 8:55 AM Maciej Wieczór-Retman
<m.wieczorretman@pm.me> wrote:
>
> Thanks for checking the patches out, do you want me to send v4 with this
> correction or is it redundant now that Andrew already wrote it?

Either way is fine with me, thanks!