diff mbox series

[v5,RESEND,06/17] mm/ioremap: add slab availability checking in ioremap_prot

Message ID 20230515090848.833045-7-bhe@redhat.com
State New
Headers
Series mm: ioremap: Convert architectures to take GENERIC_IOREMAP way |

Commit Message

Baoquan He May 15, 2023, 9:08 a.m. UTC 
  Several architectures has done checking if slab if available in
ioremap_prot(). In fact it should be done in generic ioremap_prot()
since on any architecutre, slab allocator must be available before
get_vm_area_caller() and vunmap() are used.

Add the checking into generic_ioremap_prot().

Suggested-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Signed-off-by: Baoquan He <bhe@redhat.com>
---
 mm/ioremap.c | 4 ++++
 1 file changed, 4 insertions(+)

Comments

Mike Rapoport May 16, 2023, 6:48 a.m. UTC | #1 
On Mon, May 15, 2023 at 05:08:37PM +0800, Baoquan He wrote:
> Several architectures has done checking if slab if available in
> ioremap_prot(). In fact it should be done in generic ioremap_prot()
> since on any architecutre, slab allocator must be available before
> get_vm_area_caller() and vunmap() are used.
> 
> Add the checking into generic_ioremap_prot().
> 
> Suggested-by: Christophe Leroy <christophe.leroy@csgroup.eu>
> Signed-off-by: Baoquan He <bhe@redhat.com>

Reviewed-by: Mike Rapoport (IBM) <rppt@kernel.org>

> ---
>  mm/ioremap.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/mm/ioremap.c b/mm/ioremap.c
> index 9f34a8f90b58..2fbe6b9bc50e 100644
> --- a/mm/ioremap.c
> +++ b/mm/ioremap.c
> @@ -18,6 +18,10 @@ void __iomem *generic_ioremap_prot(phys_addr_t phys_addr, size_t size,
>  	phys_addr_t last_addr;
>  	struct vm_struct *area;
>  
> +	/* An early platform driver might end up here */
> +	if (!slab_is_available())
> +		return NULL;
> +
>  	/* Disallow wrap-around or zero size */
>  	last_addr = phys_addr + size - 1;
>  	if (!size || last_addr < phys_addr)
> -- 
> 2.34.1
> 
>
Christoph Hellwig May 17, 2023, 6:30 a.m. UTC | #2 
On Mon, May 15, 2023 at 05:08:37PM +0800, Baoquan He wrote:
> Several architectures has done checking if slab if available in
> ioremap_prot(). In fact it should be done in generic ioremap_prot()
> since on any architecutre, slab allocator must be available before
> get_vm_area_caller() and vunmap() are used.
> 
> Add the checking into generic_ioremap_prot().

Should we add a WARN_ON/WARN_ON_ONCE to aid debugging?

Otherwise looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>
Baoquan He May 18, 2023, 3:43 a.m. UTC | #3 
On 05/16/23 at 11:30pm, Christoph Hellwig wrote:
> On Mon, May 15, 2023 at 05:08:37PM +0800, Baoquan He wrote:
> > Several architectures has done checking if slab if available in
> > ioremap_prot(). In fact it should be done in generic ioremap_prot()
> > since on any architecutre, slab allocator must be available before
> > get_vm_area_caller() and vunmap() are used.
> > 
> > Add the checking into generic_ioremap_prot().
> 
> Should we add a WARN_ON/WARN_ON_ONCE to aid debugging?

Sounds like a great idea, will add WARN_ON_ONCE as below. Thanks.

	if (WARN_ON_ONCE(!slab_is_available()))
		return NULL;
diff mbox series

Patch

diff --git a/mm/ioremap.c b/mm/ioremap.c
index 9f34a8f90b58..2fbe6b9bc50e 100644
--- a/mm/ioremap.c
+++ b/mm/ioremap.c
@@ -18,6 +18,10 @@  void __iomem *generic_ioremap_prot(phys_addr_t phys_addr, size_t size,
 	phys_addr_t last_addr;
 	struct vm_struct *area;
 
+	/* An early platform driver might end up here */
+	if (!slab_is_available())
+		return NULL;
+
 	/* Disallow wrap-around or zero size */
 	last_addr = phys_addr + size - 1;
 	if (!size || last_addr < phys_addr)