From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jaegeuk Kim Subject: Re: [PATCH] crypto: allow to assign gfp_t for __crypto_alloc_tfm Date: Tue, 19 May 2015 00:27:40 -0700 Message-ID: <20150519072740.GA40281@jaegeuk-mac02.hsd1.ca.comcast.net> References: <1432014416-39326-1-git-send-email-jaegeuk@kernel.org> <20150519054945.GA28060@gondor.apana.org.au> <20150519062430.GA39588@jaegeuk-mac02.hsd1.ca.comcast.net> <20150519063211.GA28347@gondor.apana.org.au> <20150519065812.GA40012@jaegeuk-mac02.hsd1.ca.comcast.net> <20150519065929.GA28610@gondor.apana.org.au> <20150519071317.GB40012@jaegeuk-mac02.hsd1.ca.comcast.net> <20150519071521.GA28862@gondor.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: davem@davemloft.net, linux-crypto@vger.kernel.org, linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net, ecryptfs@vger.kernel.org, linux-arm-kernel@lists.infradead.org To: Herbert Xu Return-path: Content-Disposition: inline In-Reply-To: <20150519071521.GA28862@gondor.apana.org.au> Sender: ecryptfs-owner@vger.kernel.org List-Id: linux-crypto.vger.kernel.org On Tue, May 19, 2015 at 03:15:21PM +0800, Herbert Xu wrote: > On Tue, May 19, 2015 at 12:13:17AM -0700, Jaegeuk Kim wrote: > > > > The key generation is done by ext4_generate_encryption_key in > > fs/ext4/crypto_key.c. > > And, ext4_file_mmap and ext4_file_open trigger it. > > Well that's where you should be doing crypto_alloc_ablkcipher > and crypto_ablkcipher_setkey. > > The whole point of a crypto tfm is to represent a key so any time > you get one you should create a crypto tfm. Carrying around a raw > key is just wrong. So, IMHO, it can consume memory too much, since tfm should be allocated for every inodes and be alive until inode eviction. Apart from giving GFP_NOFS, do you mean that it is a wrong approach? Thanks, > > Cheers, > -- > Email: Herbert Xu > Home Page: http://gondor.apana.org.au/~herbert/ > PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt From mboxrd@z Thu Jan 1 00:00:00 1970 From: jaegeuk@kernel.org (Jaegeuk Kim) Date: Tue, 19 May 2015 00:27:40 -0700 Subject: [PATCH] crypto: allow to assign gfp_t for __crypto_alloc_tfm In-Reply-To: <20150519071521.GA28862@gondor.apana.org.au> References: <1432014416-39326-1-git-send-email-jaegeuk@kernel.org> <20150519054945.GA28060@gondor.apana.org.au> <20150519062430.GA39588@jaegeuk-mac02.hsd1.ca.comcast.net> <20150519063211.GA28347@gondor.apana.org.au> <20150519065812.GA40012@jaegeuk-mac02.hsd1.ca.comcast.net> <20150519065929.GA28610@gondor.apana.org.au> <20150519071317.GB40012@jaegeuk-mac02.hsd1.ca.comcast.net> <20150519071521.GA28862@gondor.apana.org.au> Message-ID: <20150519072740.GA40281@jaegeuk-mac02.hsd1.ca.comcast.net> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Tue, May 19, 2015 at 03:15:21PM +0800, Herbert Xu wrote: > On Tue, May 19, 2015 at 12:13:17AM -0700, Jaegeuk Kim wrote: > > > > The key generation is done by ext4_generate_encryption_key in > > fs/ext4/crypto_key.c. > > And, ext4_file_mmap and ext4_file_open trigger it. > > Well that's where you should be doing crypto_alloc_ablkcipher > and crypto_ablkcipher_setkey. > > The whole point of a crypto tfm is to represent a key so any time > you get one you should create a crypto tfm. Carrying around a raw > key is just wrong. So, IMHO, it can consume memory too much, since tfm should be allocated for every inodes and be alive until inode eviction. Apart from giving GFP_NOFS, do you mean that it is a wrong approach? Thanks, > > Cheers, > -- > Email: Herbert Xu > Home Page: http://gondor.apana.org.au/~herbert/ > PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt