> Le Mon, Jun 06, 2022 at 02:49:31PM +0800, Neal Liu a écrit :
> > Hash and Crypto Engine (HACE) is designed to accelerate the throughput
> > of hash data digest, encryption, and decryption.
> >
> > Basically, HACE can be divided into two independently engines
> > - Hash Engine and Crypto Engine. This patch aims to add HACE hash
> > engine driver for hash accelerator.
> >
> > Signed-off-by: Neal Liu <neal_liu@xxxxxxxxxxxxxx>
> > Signed-off-by: Johnny Huang <johnny_huang@xxxxxxxxxxxxxx>
> > ---
> > MAINTAINERS | 7 +
> > drivers/crypto/Kconfig | 1 +
> > drivers/crypto/Makefile | 1 +
> > drivers/crypto/aspeed/Kconfig | 22 +
> > drivers/crypto/aspeed/Makefile | 6 +
> > drivers/crypto/aspeed/aspeed-hace-hash.c | 1409
> ++++++++++++++++++++++
> > drivers/crypto/aspeed/aspeed-hace.c | 221 ++++
> > drivers/crypto/aspeed/aspeed-hace.h | 182 +++
> > 8 files changed, 1849 insertions(+)
> > create mode 100644 drivers/crypto/aspeed/Kconfig create mode 100644
> > drivers/crypto/aspeed/Makefile create mode 100644
> > drivers/crypto/aspeed/aspeed-hace-hash.c
> > create mode 100644 drivers/crypto/aspeed/aspeed-hace.c
> > create mode 100644 drivers/crypto/aspeed/aspeed-hace.h
>
> Hello please see my comments below.
>
> [...]
>
> > diff --git a/drivers/crypto/aspeed/aspeed-hace-hash.c
> > b/drivers/crypto/aspeed/aspeed-hace-hash.c
> > new file mode 100644
> > index 000000000000..9b003e12b2c8
> > --- /dev/null
> > +++ b/drivers/crypto/aspeed/aspeed-hace-hash.c
> > @@ -0,0 +1,1409 @@
> > +// SPDX-License-Identifier: GPL-2.0+
> > +/*
> > + * Copyright (c) 2021 Aspeed Technology Inc.
> > + */
> > +
> > +#include "aspeed-hace.h"
> > +
> > +//#define ASPEED_AHASH_DEBUG
>
> Please remove this
Sure !
>
> [...]
>
> > +
> > +#ifdef ASPEED_AHASH_DEBUG
> > +#define AHASH_DBG(h, fmt, ...) \
> > + dev_info((h)->dev, "%s() " fmt, __func__, ##__VA_ARGS__) #else
> > +#define AHASH_DBG(h, fmt, ...) \
> > + ((void)(h))
> > +#endif
>
> Why not using dev_dbg() ?
dev_dbg() would be nice. I'll revise it.
>
> [...]
> > + sg_len = dma_map_sg(hace_dev->dev, rctx->src_sg, rctx->src_nents,
> > + DMA_TO_DEVICE);
> > + if (!sg_len) {
> > + dev_warn(hace_dev->dev, "dma_map_sg() src error\n");
> > + rc = -ENOMEM;
> > + goto end;
> > + }
>
> This test does handle the possible negative return of dma_map_sg()
Zero is returned for any error of dma_map_sg().
>
> [...]
> > + memset(bctx->ipad + keylen, 0, bs - keylen);
> > + memcpy(bctx->opad, bctx->ipad, bs);
> > +
> > + for (i = 0; i < bs; i++) {
> > + bctx->ipad[i] ^= 0x36;
> > + bctx->opad[i] ^= 0x5c;
>
> Please use HMAC_OPAD_VALUE and HMAC_IPAD_VALUE from
> include/crypto/hmac.h
>
Thanks for the notice. I'll revise it in next patch, thanks.
> [...]
> > +int aspeed_register_hace_hash_algs(struct aspeed_hace_dev *hace_dev)
> > +{
> > + int rc, i;
> > +
> > + AHASH_DBG(hace_dev, "\n");
> > +
> > + for (i = 0; i < ARRAY_SIZE(aspeed_ahash_algs); i++) {
> > + aspeed_ahash_algs[i].hace_dev = hace_dev;
> > + rc = crypto_register_ahash(&aspeed_ahash_algs[i].alg.ahash);
> > + if (rc)
> > + return rc;
> > + }
>
> If any hash fail to register, the function exits but you will still unregister all
> hashes (even ones not registered) on aspeed_unregister_hace_hash_algs().
Yes. crypto/algapi.c would handle this part even if algo is not registered.
But of course we could use a variable to store which algo is registered, which is not.
>
> [...]
> > +static int aspeed_hace_probe(struct platform_device *pdev) {
> > + const struct of_device_id *hace_dev_id;
> > + struct aspeed_engine_hash *hash_engine;
> > + struct aspeed_hace_dev *hace_dev;
> > + struct resource *res;
> > + int rc;
> > +
> > + hace_dev = devm_kzalloc(&pdev->dev, sizeof(struct aspeed_hace_dev),
> > + GFP_KERNEL);
> > + if (!hace_dev)
> > + return -ENOMEM;
> > +
> > + hace_dev_id = of_match_device(aspeed_hace_of_matches, &pdev->dev);
> > + if (!hace_dev_id) {
> > + dev_err(&pdev->dev, "Failed to match hace dev id\n");
> > + return -EINVAL;
> > + }
> > +
> > + hace_dev->dev = &pdev->dev;
> > + hace_dev->version = (unsigned long)hace_dev_id->data;
> > + hash_engine = &hace_dev->hash_engine;
> > +
> > + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > +
> > + platform_set_drvdata(pdev, hace_dev);
> > +
> > + spin_lock_init(&hash_engine->lock);
> > + tasklet_init(&hash_engine->done_task, aspeed_hace_hash_done_task,
> > + (unsigned long)hace_dev);
> > + tasklet_init(&hash_engine->queue_task, aspeed_hace_hash_queue_task,
> > + (unsigned long)hace_dev);
> > + crypto_init_queue(&hash_engine->queue,
> ASPEED_HASH_QUEUE_LENGTH);
> > +
> > + hace_dev->regs = devm_ioremap_resource(&pdev->dev, res);
> > + if (!hace_dev->regs) {
> > + dev_err(&pdev->dev, "Failed to map resources\n");
> > + return -ENOMEM;
> > + }
> > +
> > + hace_dev->irq = platform_get_irq(pdev, 0);
> > + if (!hace_dev->irq) {
> > + dev_err(&pdev->dev, "Failed to get interrupt\n");
> > + return -ENXIO;
> > + }
> > +
> > + rc = devm_request_irq(&pdev->dev, hace_dev->irq, aspeed_hace_irq, 0,
> > + dev_name(&pdev->dev), hace_dev);
> > + if (rc) {
> > + dev_err(&pdev->dev, "Failed to request interrupt\n");
> > + return rc;
> > + }
> > +
> > + hace_dev->clk = devm_clk_get(&pdev->dev, NULL);
> > + if (IS_ERR(hace_dev->clk)) {
> > + dev_err(&pdev->dev, "Failed to get clk\n");
> > + return -ENODEV;
> > + }
> > +
> > + rc = clk_prepare_enable(hace_dev->clk);
> > + if (rc) {
> > + dev_err(&pdev->dev, "Failed to enable clock 0x%x\n", rc);
> > + return rc;
> > + }
> > +
> > + hash_engine->ahash_src_addr =
> > + dma_alloc_coherent(&pdev->dev,
> > + ASPEED_HASH_SRC_DMA_BUF_LEN,
> > + &hash_engine->ahash_src_dma_addr,
> > + GFP_KERNEL);
> > + if (!hash_engine->ahash_src_addr) {
> > + dev_err(&pdev->dev, "Failed to allocate dma buffer\n");
> > + rc = -ENOMEM;
> > + goto end;
> > + }
> > +
> > + rc = aspeed_hace_register(hace_dev);
> > + if (rc) {
> > + dev_err(&pdev->dev, "Failed to register algs, rc:0x%x\n", rc);
> > + rc = 0;
> > + }
>
> So you print an error but you keep the driver loaded even if no hashes (or a
> limited number of them) are registered, for which interest ?
>
Take PATCH 5/5 as consideration, aspeed_hace_register() would register both hashes & skciphers.
So even one of them failed, the driver still can be loaded.
> [...]
> > +struct aspeed_sham_ctx {
> > + struct aspeed_hace_dev *hace_dev;
> > + unsigned long flags; /* hmac flag */
> > +
> > + /* fallback stuff */
> > + struct aspeed_sha_hmac_ctx base[0];
>
> I am not sure to understand the [0] purpose, and the comment said something
> about fallback, but your driver dont do any real fallback.
[0] is a convenience to reference it as a pointer.
The comment should be removed, thanks.
>
> [...]
> > +struct aspeed_hace_dev {
> > + void __iomem *regs;
> > + void __iomem *sec_regs;
>
> sec_regs is unused.
I miss this part, I'll revise it in next patch, thanks.
>
> > + struct device *dev;
> > + int irq;
> > + struct clk *clk;
> > + unsigned long version;
> > + struct aspeed_engine_hash hash_engine;
> > +};
> > +
> > +struct aspeed_hace_alg {
> > + struct aspeed_hace_dev *hace_dev;
> > + union {
> > + struct skcipher_alg skcipher;
> > + struct aead_alg aead;
> > + struct ahash_alg ahash;
> > + struct kpp_alg kpp;
> > + struct akcipher_alg akcipher;
>
> Your patch dont do any kpp or akcipher (and aead/skcipher also).
> Please drop this.
I miss this part, I'll revise it in next patch, thanks.
>
> Regards
