From: arno@natisbad.org (Arnaud Ebalard)
To: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Boris Brezillon <boris.brezillon@free-electrons.com>,
Thomas Petazzoni <thomas.petazzoni@free-electrons.com>,
Jason Cooper <jason@lakedaemon.net>,
Herbert Xu <herbert@gondor.apana.org.au>,
"David S. Miller" <davem@davemloft.net>,
linux-crypto@vger.kernel.org
Subject: Re: [PATCH 4/6] crypto: marvell: use memcpy_fromio()/memcpy_toio()
Date: Tue, 20 Oct 2015 01:26:55 +0200 [thread overview]
Message-ID: <87oafucudc.fsf@natisbad.org> (raw)
In-Reply-To: <E1Znrnf-00064N-OS@rmk-PC.arm.linux.org.uk> (Russell King's message of "Sun, 18 Oct 2015 18:31:15 +0100")
Hi Russell,
Russell King <rmk+kernel@arm.linux.org.uk> writes:
> Use the IO memcpy() functions when copying from/to MMIO memory.
> These locations were found via sparse.
On recent MVEBU hardware, *_std_* function are not expected to be used
because we will instead use the TDMA-based versions. So the only
possible penalty we could get (if any) from this change on recent
devices is for the copy of the IV in mv_cesa_ablkcipher_process(). Out
of curiosity, I took a quick look at what is generated and it seems this
results in a call to mmiocpy():
00000340 <mv_cesa_ablkcipher_process>:
340: e1a0c00d mov ip, sp
344: e92dd830 push {r4, r5, fp, ip, lr, pc}
348: e24cb004 sub fp, ip, #4
34c: e24dd008 sub sp, sp, #8
....
3a4: e5943010 ldr r3, [r4, #16]
3a8: e5951008 ldr r1, [r5, #8]
3ac: e594001c ldr r0, [r4, #28]
3b0: e2811040 add r1, r1, #64 ; 0x40
3b4: e593201c ldr r2, [r3, #28]
3b8: ebfffffe bl 0 <mmiocpy>
which if I am not mistaken is provided by arch/arm/lib/memcpy.S via:
ENTRY(mmiocpy)
ENTRY(memcpy)
#include "copy_template.S"
ENDPROC(memcpy)
ENDPROC(mmiocpy)
Am I right in concluding this will end up being the code of a usual
memcpy(), i.e. the change is a noop in the final code compared to
previous version?
Cheers,
a+
>
> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
> ---
> drivers/crypto/marvell/cipher.c | 11 ++++++-----
> drivers/crypto/marvell/hash.c | 16 ++++++++--------
> 2 files changed, 14 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/crypto/marvell/cipher.c b/drivers/crypto/marvell/cipher.c
> index 4db2d632204f..6edae64bb387 100644
> --- a/drivers/crypto/marvell/cipher.c
> +++ b/drivers/crypto/marvell/cipher.c
> @@ -98,10 +98,10 @@ static void mv_cesa_ablkcipher_std_step(struct ablkcipher_request *req)
>
> /* FIXME: only update enc_len field */
> if (!sreq->skip_ctx) {
> - memcpy(engine->sram, &sreq->op, sizeof(sreq->op));
> + memcpy_toio(engine->sram, &sreq->op, sizeof(sreq->op));
> sreq->skip_ctx = true;
> } else {
> - memcpy(engine->sram, &sreq->op, sizeof(sreq->op.desc));
> + memcpy_toio(engine->sram, &sreq->op, sizeof(sreq->op.desc));
> }
>
> mv_cesa_set_int_mask(engine, CESA_SA_INT_ACCEL0_DONE);
> @@ -145,8 +145,9 @@ static int mv_cesa_ablkcipher_process(struct crypto_async_request *req,
> if (ret)
> return ret;
>
> - memcpy(ablkreq->info, engine->sram + CESA_SA_CRYPT_IV_SRAM_OFFSET,
> - crypto_ablkcipher_ivsize(crypto_ablkcipher_reqtfm(ablkreq)));
> + memcpy_fromio(ablkreq->info,
> + engine->sram + CESA_SA_CRYPT_IV_SRAM_OFFSET,
> + crypto_ablkcipher_ivsize(crypto_ablkcipher_reqtfm(ablkreq)));
>
> return 0;
> }
> @@ -181,7 +182,7 @@ mv_cesa_ablkcipher_std_prepare(struct ablkcipher_request *req)
> sreq->size = 0;
> sreq->offset = 0;
> mv_cesa_adjust_op(engine, &sreq->op);
> - memcpy(engine->sram, &sreq->op, sizeof(sreq->op));
> + memcpy_toio(engine->sram, &sreq->op, sizeof(sreq->op));
> }
>
> static inline void mv_cesa_ablkcipher_prepare(struct crypto_async_request *req,
> diff --git a/drivers/crypto/marvell/hash.c b/drivers/crypto/marvell/hash.c
> index 84ddc4cbfa9d..8330815d72fc 100644
> --- a/drivers/crypto/marvell/hash.c
> +++ b/drivers/crypto/marvell/hash.c
> @@ -209,8 +209,8 @@ static void mv_cesa_ahash_std_step(struct ahash_request *req)
> size_t len;
>
> if (creq->cache_ptr)
> - memcpy(engine->sram + CESA_SA_DATA_SRAM_OFFSET, creq->cache,
> - creq->cache_ptr);
> + memcpy_toio(engine->sram + CESA_SA_DATA_SRAM_OFFSET,
> + creq->cache, creq->cache_ptr);
>
> len = min_t(size_t, req->nbytes + creq->cache_ptr - sreq->offset,
> CESA_SA_SRAM_PAYLOAD_SIZE);
> @@ -251,10 +251,10 @@ static void mv_cesa_ahash_std_step(struct ahash_request *req)
> if (len + trailerlen > CESA_SA_SRAM_PAYLOAD_SIZE) {
> len &= CESA_HASH_BLOCK_SIZE_MSK;
> new_cache_ptr = 64 - trailerlen;
> - memcpy(creq->cache,
> - engine->sram +
> - CESA_SA_DATA_SRAM_OFFSET + len,
> - new_cache_ptr);
> + memcpy_fromio(creq->cache,
> + engine->sram +
> + CESA_SA_DATA_SRAM_OFFSET + len,
> + new_cache_ptr);
> } else {
> len += mv_cesa_ahash_pad_req(creq,
> engine->sram + len +
> @@ -272,7 +272,7 @@ static void mv_cesa_ahash_std_step(struct ahash_request *req)
> mv_cesa_update_op_cfg(op, frag_mode, CESA_SA_DESC_CFG_FRAG_MSK);
>
> /* FIXME: only update enc_len field */
> - memcpy(engine->sram, op, sizeof(*op));
> + memcpy_toio(engine->sram, op, sizeof(*op));
>
> if (frag_mode == CESA_SA_DESC_CFG_FIRST_FRAG)
> mv_cesa_update_op_cfg(op, CESA_SA_DESC_CFG_MID_FRAG,
> @@ -312,7 +312,7 @@ static void mv_cesa_ahash_std_prepare(struct ahash_request *req)
>
> sreq->offset = 0;
> mv_cesa_adjust_op(engine, &creq->op_tmpl);
> - memcpy(engine->sram, &creq->op_tmpl, sizeof(creq->op_tmpl));
> + memcpy_toio(engine->sram, &creq->op_tmpl, sizeof(creq->op_tmpl));
> }
>
> static void mv_cesa_ahash_step(struct crypto_async_request *req)
next prev parent reply other threads:[~2015-10-19 23:27 UTC|newest]
Thread overview: 96+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-09 10:29 [PATCH 0/3] crypto: fixes for Marvell hash Russell King - ARM Linux
2015-10-09 10:29 ` [PATCH 1/3] crypto: ensure algif_hash does not pass a zero-sized state Russell King
2015-10-09 10:34 ` Herbert Xu
2015-10-09 10:41 ` Russell King - ARM Linux
2015-10-09 10:42 ` Herbert Xu
2015-10-09 10:29 ` [PATCH 2/3] crypto: marvell: fix stack smashing in marvell/hash.c Russell King
2015-10-09 10:29 ` [PATCH 3/3] crypto: marvell: initialise struct mv_cesa_ahash_req Russell King
2015-10-09 10:46 ` [PATCH v2 0/3] crypto: fixes for Marvell hash Russell King - ARM Linux
2015-10-09 10:48 ` [PATCH v2 1/3] crypto: ensure algif_hash does not pass a zero-sized state Russell King
2015-10-09 10:48 ` [PATCH v2 2/3] crypto: marvell: fix stack smashing in marvell/hash.c Russell King
2015-10-09 16:13 ` Boris Brezillon
2015-10-09 10:48 ` [PATCH v2 3/3] crypto: marvell: initialise struct mv_cesa_ahash_req Russell King
2015-10-09 16:15 ` Boris Brezillon
2015-10-09 12:42 ` [PATCH v2 0/3] crypto: fixes for Marvell hash Russell King - ARM Linux
2015-10-09 16:12 ` Boris Brezillon
2015-10-09 19:43 ` [PATCH v3 0/5] " Russell King - ARM Linux
2015-10-09 19:43 ` [PATCH v3 1/5] crypto: ensure algif_hash does not pass a zero-sized state Russell King
2015-10-10 16:46 ` Boris Brezillon
2015-10-10 16:52 ` Russell King - ARM Linux
2015-10-11 6:59 ` Herbert Xu
2015-10-11 6:57 ` Herbert Xu
2015-10-13 14:33 ` Herbert Xu
2015-10-15 9:39 ` Russell King - ARM Linux
2015-10-15 9:41 ` Herbert Xu
2015-10-15 12:59 ` Russell King - ARM Linux
2015-10-15 13:13 ` Herbert Xu
2015-10-16 23:24 ` Victoria Milhoan
2015-10-17 7:56 ` Russell King - ARM Linux
2015-10-09 19:43 ` [PATCH v3 2/5] crypto: marvell: fix stack smashing in marvell/hash.c Russell King
2015-10-09 19:43 ` [PATCH v3 3/5] crypto: marvell: initialise struct mv_cesa_ahash_req Russell King
2015-10-09 19:50 ` Boris Brezillon
2015-10-09 19:52 ` Russell King - ARM Linux
2015-10-09 19:43 ` [PATCH v3 4/5] crypto: marvell: fix wrong hash results Russell King
2015-10-09 19:51 ` Boris Brezillon
2015-10-09 19:43 ` [PATCH v3 5/5] crypto: marvell: factor out common import functions Russell King
2015-10-09 19:55 ` Boris Brezillon
2015-10-09 20:14 ` [PATCH v3b 5/5] crypto: marvell: factor out common import/export functions Russell King
2015-10-09 20:19 ` Boris Brezillon
2015-10-09 22:37 ` Arnaud Ebalard
2015-10-09 23:51 ` Russell King - ARM Linux
2015-10-10 10:31 ` Arnaud Ebalard
2015-10-10 11:29 ` Russell King - ARM Linux
2015-10-10 16:17 ` Russell King - ARM Linux
2015-10-11 6:55 ` Herbert Xu
2015-10-13 13:00 ` Herbert Xu
2015-10-13 13:55 ` Russell King - ARM Linux
2015-10-13 13:57 ` Herbert Xu
2015-10-13 13:59 ` Russell King - ARM Linux
2015-10-13 14:01 ` Herbert Xu
2015-10-10 18:07 ` Marek Vasut
2015-10-09 19:57 ` [PATCH v3 0/5] crypto: fixes for Marvell hash Boris Brezillon
2015-10-18 16:16 ` [PATCH 00/18] crypto: further fixes for Marvell CESA hash Russell King - ARM Linux
2015-10-18 16:23 ` [PATCH 01/18] crypto: marvell: easier way to get the transform Russell King
2015-10-19 1:37 ` crypto: ahash - Add crypto_ahash_blocksize Herbert Xu
2015-10-18 16:23 ` [PATCH 02/18] crypto: marvell: keep creq->state in CPU endian format at all times Russell King
2015-10-18 16:23 ` [PATCH 03/18] crypto: marvell: add flag to determine algorithm endianness Russell King
2015-10-19 15:04 ` Jason Cooper
2015-10-19 15:25 ` Russell King - ARM Linux
2015-10-19 16:15 ` Jason Cooper
2015-10-19 16:18 ` Herbert Xu
2015-10-18 16:23 ` [PATCH 04/18] crypto: marvell: fix the bit length endianness Russell King
2015-10-18 16:23 ` [PATCH 05/18] crypto: marvell: ensure template operation is initialised Russell King
2015-10-18 16:23 ` [PATCH 06/18] crypto: marvell: const-ify argument to mv_cesa_get_op_cfg() Russell King
2015-10-18 16:24 ` [PATCH 07/18] crypto: marvell: factor out first fragment decisions to helper Russell King
2015-10-18 16:24 ` [PATCH 08/18] crypto: marvell: factor out adding an operation and launching it Russell King
2015-10-18 16:24 ` [PATCH 09/18] crypto: marvell: always ensure mid-fragments after first-fragment Russell King
2015-10-18 16:24 ` [PATCH 10/18] crypto: marvell: move mv_cesa_dma_add_frag() calls Russell King
2015-10-18 16:24 ` [PATCH 11/18] crypto: marvell: use presence of scatterlist to determine data load Russell King
2015-10-18 16:24 ` [PATCH 12/18] crypto: marvell: ensure iter.base.op_len is the full op length Russell King
2015-10-18 16:24 ` [PATCH 13/18] crypto: marvell: avoid adding final operation within loop Russell King
2015-10-18 16:24 ` [PATCH 14/18] crypto: marvell: rearrange last request handling Russell King
2015-10-18 16:24 ` [PATCH 15/18] crypto: marvell: rearrange handling for hw finished hashes Russell King
2015-10-18 16:24 ` [PATCH 16/18] crypto: marvell: rearrange handling for sw padded hashes Russell King
2015-10-18 16:24 ` [PATCH 17/18] crypto: marvell: fix first-fragment handling in mv_cesa_ahash_dma_last_req() Russell King
2015-10-19 22:53 ` Arnaud Ebalard
2015-10-18 16:24 ` [PATCH 18/18] crypto: marvell/cesa: fix memory leak Russell King
2015-10-18 17:18 ` [PATCH 00/18] crypto: further fixes for Marvell CESA hash Boris Brezillon
2015-10-18 23:57 ` Arnaud Ebalard
2015-10-19 22:57 ` Arnaud Ebalard
2015-10-18 17:30 ` [PATCH 0/6] Sparse related fixes Russell King - ARM Linux
2015-10-18 17:31 ` [PATCH 1/6] crypto: marvell: use readl_relaxed()/writel_relaxed() Russell King
2015-10-18 17:31 ` [PATCH 2/6] crypto: marvell: use dma_addr_t for cur_dma Russell King
2015-10-18 17:31 ` [PATCH 3/6] crypto: marvell: use gfp_t for gfp flags Russell King
2015-10-18 17:31 ` [PATCH 4/6] crypto: marvell: use memcpy_fromio()/memcpy_toio() Russell King
2015-10-19 23:26 ` Arnaud Ebalard [this message]
2015-10-20 7:58 ` Russell King - ARM Linux
2015-10-18 17:31 ` [PATCH 5/6] crypto: marvell: fix missing cpu_to_le32() in mv_cesa_dma_add_op() Russell King
2015-10-18 17:31 ` [PATCH 6/6] crypto: marvell: use __le32 for hardware descriptors Russell King
2015-10-18 17:49 ` [PATCH 0/6] Sparse related fixes Boris Brezillon
2015-10-19 23:29 ` Arnaud Ebalard
2015-10-20 14:21 ` Herbert Xu
2015-10-20 14:20 ` [PATCH 00/18] crypto: further fixes for Marvell CESA hash Herbert Xu
2015-10-09 12:12 ` [PATCH 0/3] crypto: fixes for Marvell hash Thomas Petazzoni
2015-10-09 12:31 ` Russell King - ARM Linux
2015-10-09 12:40 ` Thomas Petazzoni
2015-10-09 14:35 ` Herbert Xu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87oafucudc.fsf@natisbad.org \
--to=arno@natisbad.org \
--cc=boris.brezillon@free-electrons.com \
--cc=davem@davemloft.net \
--cc=herbert@gondor.apana.org.au \
--cc=jason@lakedaemon.net \
--cc=linux-crypto@vger.kernel.org \
--cc=rmk+kernel@arm.linux.org.uk \
--cc=thomas.petazzoni@free-electrons.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).