# Re: \[PATCH\] crypto: allow to assign gfp\_t for \_\_crypto\_alloc\_tfm

- From: Jaegeuk Kim \<jaegeuk@kernel.org\>
- Date: 2015-05-19 06:24:38 
- List: linux-ext4
- Page: https://ratatoskr.run/linux-ext4/2015/05/4496596
- Thread: https://ratatoskr.run/linux-crypto/2015/05/4496563/t.md

```text
On Tue, May 19, 2015 at 01:49:45PM +0800, Herbert Xu wrote:
> On Mon, May 18, 2015 at 10:46:56PM -0700, Jaegeuk Kim wrote:
> > This patch adds a parameter, gfp_t, for __crypto_alloc_tfm.
> > Now EXT4 and F2FS use the crypto engine to encrypt a page when writing it to the
> > disk.
> > That happens during ->writepage and it needs to allocate memory with
> > GFP_NOFS.
> > 
> > Otherwise, in the f2fs case, kernel reports such the following warning.
> 
> Normally crypto structures should only be allocated on control
> paths where sleeping or swapping is not an issue.  Why is ext4
> doing crypto allocations on the data path?
> 

Recently, Ted introduced per-file encryption feature as follows.

https://lwn.net/Articles/639427/

The call path in fs/ext4/crypto.c is:
 - writepage
  - ext4_encrypt
   - ext4_get_crypto_ctx
    - crypto_alloc_ablkcipher

AFAIK, this way can achieve to reduce memory footprint gracefully.
Just before submitting bios, fs allocates the required memory, and then end_io
will free them in pair.

Thanks,

> Cheers,
> -- 
> Email: Herbert Xu [ref]
> Home Page: http://gondor.apana.org.au/~herbert/
> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
```
