2022-09-22 03:14:51

by Kees Cook

Subject: [PATCH 03/12] net: ipa: Proactively round up to kmalloc bucket size

Instead of discovering the kmalloc bucket size _after_ allocation, round
up proactively so the allocation is explicitly made for the full size,
allowing the compiler to correctly reason about the resulting size of
the buffer through the existing __alloc_size() hint.

Cc: Alex Elder <[email protected]>
Cc: "David S. Miller" <[email protected]>
Cc: Eric Dumazet <[email protected]>
Cc: Jakub Kicinski <[email protected]>
Cc: Paolo Abeni <[email protected]>
Cc: [email protected]
Signed-off-by: Kees Cook <[email protected]>
---
drivers/net/ipa/gsi_trans.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ipa/gsi_trans.c b/drivers/net/ipa/gsi_trans.c
index 18e7e8c405be..cec968854dcf 100644
--- a/drivers/net/ipa/gsi_trans.c
+++ b/drivers/net/ipa/gsi_trans.c
@@ -89,6 +89,7 @@ int gsi_trans_pool_init(struct gsi_trans_pool *pool, size_t size, u32 count,
u32 max_alloc)
{
void *virt;
+ size_t allocate;

if (!size)
return -EINVAL;
@@ -104,13 +105,15 @@ int gsi_trans_pool_init(struct gsi_trans_pool *pool, size_t size, u32 count,
* If there aren't enough entries starting at the free index,
* we just allocate free entries from the beginning of the pool.
*/
- virt = kcalloc(count + max_alloc - 1, size, GFP_KERNEL);
+ allocate = size_mul(count + max_alloc - 1, size);
+ allocate = kmalloc_size_roundup(allocate);
+ virt = kzalloc(allocate, GFP_KERNEL);
if (!virt)
return -ENOMEM;

pool->base = virt;
/* If the allocator gave us any extra memory, use it */
- pool->count = ksize(pool->base) / size;
+ pool->count = allocate / size;
pool->free = 0;
pool->max_alloc = max_alloc;
pool->size = size;
--
2.34.1

2022-09-22 13:46:43

by Alex Elder

Subject: Re: [PATCH 03/12] net: ipa: Proactively round up to kmalloc bucket size

On 9/21/22 10:10 PM, Kees Cook wrote:
> Instead of discovering the kmalloc bucket size _after_ allocation, round
> up proactively so the allocation is explicitly made for the full size,
> allowing the compiler to correctly reason about the resulting size of
> the buffer through the existing __alloc_size() hint.
>
> Cc: Alex Elder <[email protected]>
> Cc: "David S. Miller" <[email protected]>
> Cc: Eric Dumazet <[email protected]>
> Cc: Jakub Kicinski <[email protected]>
> Cc: Paolo Abeni <[email protected]>
> Cc: [email protected]
> Signed-off-by: Kees Cook <[email protected]>
> ---
> drivers/net/ipa/gsi_trans.c | 7 +++++--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/ipa/gsi_trans.c b/drivers/net/ipa/gsi_trans.c
> index 18e7e8c405be..cec968854dcf 100644
> --- a/drivers/net/ipa/gsi_trans.c
> +++ b/drivers/net/ipa/gsi_trans.c
> @@ -89,6 +89,7 @@ int gsi_trans_pool_init(struct gsi_trans_pool *pool, size_t size, u32 count,
> u32 max_alloc)
> {
> void *virt;
> + size_t allocate;

I don't care about this but the reverse Christmas tree
convention would put the "allocate" variable definition
above "virt".

Whether you fix that or not, this patch looks good to me.

Reviewed-by: Alex Elder <[email protected]>

> if (!size)
> return -EINVAL;
> @@ -104,13 +105,15 @@ int gsi_trans_pool_init(struct gsi_trans_pool *pool, size_t size, u32 count,
> * If there aren't enough entries starting at the free index,
> * we just allocate free entries from the beginning of the pool.
> */
> - virt = kcalloc(count + max_alloc - 1, size, GFP_KERNEL);
> + allocate = size_mul(count + max_alloc - 1, size);
> + allocate = kmalloc_size_roundup(allocate);
> + virt = kzalloc(allocate, GFP_KERNEL);
> if (!virt)
> return -ENOMEM;
>
> pool->base = virt;
> /* If the allocator gave us any extra memory, use it */
> - pool->count = ksize(pool->base) / size;
> + pool->count = allocate / size;
> pool->free = 0;
> pool->max_alloc = max_alloc;
> pool->size = size;

2022-09-22 16:12:55

by Kees Cook

Subject: Re: [PATCH 03/12] net: ipa: Proactively round up to kmalloc bucket size

On Thu, Sep 22, 2022 at 08:45:19AM -0500, Alex Elder wrote:
> On 9/21/22 10:10 PM, Kees Cook wrote:
> > Instead of discovering the kmalloc bucket size _after_ allocation, round
> > up proactively so the allocation is explicitly made for the full size,
> > allowing the compiler to correctly reason about the resulting size of
> > the buffer through the existing __alloc_size() hint.
> >
> > Cc: Alex Elder <[email protected]>
> > Cc: "David S. Miller" <[email protected]>
> > Cc: Eric Dumazet <[email protected]>
> > Cc: Jakub Kicinski <[email protected]>
> > Cc: Paolo Abeni <[email protected]>
> > Cc: [email protected]
> > Signed-off-by: Kees Cook <[email protected]>
> > ---
> > drivers/net/ipa/gsi_trans.c | 7 +++++--
> > 1 file changed, 5 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/ipa/gsi_trans.c b/drivers/net/ipa/gsi_trans.c
> > index 18e7e8c405be..cec968854dcf 100644
> > --- a/drivers/net/ipa/gsi_trans.c
> > +++ b/drivers/net/ipa/gsi_trans.c
> > @@ -89,6 +89,7 @@ int gsi_trans_pool_init(struct gsi_trans_pool *pool, size_t size, u32 count,
> > u32 max_alloc)
> > {
> > void *virt;
> > + size_t allocate;
>
> I don't care about this but the reverse Christmas tree
> convention would put the "allocate" variable definition
> above "virt".

Oops, yes; thank you!

-Kees

--
Kees Cook