Re: [PATCH v2 8/8] scsi/pmcraid: Use sgl_alloc_order() and sgl_free_order()
On 10/17/2017 12:49 AM, Bart Van Assche wrote:
> Use the sgl_alloc_order() and sgl_free_order() functions instead
> of open coding these functions.
>
> Signed-off-by: Bart Van Assche
> Reviewed-by: Johannes Thumshirn
> Cc: linux-s...@vger.kernel.org
> Cc: Martin K. Petersen
> Cc: Anil Ravindranath
> ---
> drivers/scsi/Kconfig | 1 +
> drivers/scsi/pmcraid.c | 43 ---
> drivers/scsi/pmcraid.h | 2 +-
> 3 files changed, 6 insertions(+), 40 deletions(-)
>
> diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
> index d11e75e76195..632200ec36a6 100644
> --- a/drivers/scsi/Kconfig
> +++ b/drivers/scsi/Kconfig
> @@ -1576,6 +1576,7 @@ config ZFCP
> config SCSI_PMCRAID
> tristate "PMC SIERRA Linux MaxRAID adapter support"
> depends on PCI && SCSI && NET
> + select SGL_ALLOC
> ---help---
> This driver supports the PMC SIERRA MaxRAID adapters.
>
> diff --git a/drivers/scsi/pmcraid.c b/drivers/scsi/pmcraid.c
> index b4d6cd8cd1ad..95fc0352f9bb 100644
> --- a/drivers/scsi/pmcraid.c
> +++ b/drivers/scsi/pmcraid.c
> @@ -3232,12 +3232,7 @@ static int pmcraid_build_ioadl(
> */
> static void pmcraid_free_sglist(struct pmcraid_sglist *sglist)
> {
> - int i;
> -
> - for (i = 0; i < sglist->num_sg; i++)
> - __free_pages(sg_page(&(sglist->scatterlist[i])),
> - sglist->order);
> -
> + sgl_free_order(sglist->scatterlist, sglist->order);
> kfree(sglist);
> }
>
> @@ -3254,50 +3249,20 @@ static void pmcraid_free_sglist(struct pmcraid_sglist
> *sglist)
> static struct pmcraid_sglist *pmcraid_alloc_sglist(int buflen)
> {
> struct pmcraid_sglist *sglist;
> - struct scatterlist *scatterlist;
> - struct page *page;
> - int num_elem, i, j;
> int sg_size;
> int order;
> - int bsize_elem;
>
> sg_size = buflen / (PMCRAID_MAX_IOADLS - 1);
> order = (sg_size > 0) ? get_order(sg_size) : 0;
> - bsize_elem = PAGE_SIZE * (1 << order);
> -
> - /* Determine the actual number of sg entries needed */
> - if (buflen % bsize_elem)
> - num_elem = (buflen / bsize_elem) + 1;
> - else
> - num_elem = buflen / bsize_elem;
>
> /* Allocate a scatter/gather list for the DMA */
> - sglist = kzalloc(sizeof(struct pmcraid_sglist) +
> - (sizeof(struct scatterlist) * (num_elem - 1)),
> - GFP_KERNEL);
> -
> + sglist = kzalloc(sizeof(struct pmcraid_sglist), GFP_KERNEL);
> if (sglist == NULL)
> return NULL;
>
> - scatterlist = sglist->scatterlist;
> - sg_init_table(scatterlist, num_elem);
> sglist->order = order;
> - sglist->num_sg = num_elem;
> - sg_size = buflen;
> -
> - for (i = 0; i < num_elem; i++) {
> - page = alloc_pages(GFP_KERNEL|GFP_DMA|__GFP_ZERO, order);
> - if (!page) {
> - for (j = i - 1; j >= 0; j--)
> - __free_pages(sg_page(&scatterlist[j]), order);
> - kfree(sglist);
> - return NULL;
> - }
> -
> - sg_set_page(&scatterlist[i], page,
> - sg_size < bsize_elem ? sg_size : bsize_elem, 0);
> - sg_size -= bsize_elem;
> - }
> + sgl_alloc_order(buflen, order, false,
> + GFP_KERNEL | GFP_DMA | __GFP_ZERO, &sglist->num_sg);
>
> return sglist;
> }
> diff --git a/drivers/scsi/pmcraid.h b/drivers/scsi/pmcraid.h
> index 44da91712115..754ef30927e2 100644
> --- a/drivers/scsi/pmcraid.h
> +++ b/drivers/scsi/pmcraid.h
> @@ -542,7 +542,7 @@ struct pmcraid_sglist {
> u32 order;
> u32 num_sg;
> u32 num_dma_sg;
> - struct scatterlist scatterlist[1];
> + struct scatterlist *scatterlist;
> };
>
> /* page D0 inquiry data of focal point resource */
>
The comment to ipr applied here, too.
We need input from the pmcraid folks if this is a valid conversion.
Cheers,
Hannes
--
Dr. Hannes ReineckeTeamlead Storage & Networking
h...@suse.de +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
Re: [PATCH v2 7/8] scsi/pmcraid: Remove an unused structure member
On 10/17/2017 12:49 AM, Bart Van Assche wrote:
> Signed-off-by: Bart Van Assche
> Reviewed-by: Johannes Thumshirn
> Cc: linux-s...@vger.kernel.org
> Cc: Martin K. Petersen
> Cc: Anil Ravindranath
> ---
> drivers/scsi/pmcraid.h | 1 -
> 1 file changed, 1 deletion(-)
>
> diff --git a/drivers/scsi/pmcraid.h b/drivers/scsi/pmcraid.h
> index 8bfac72a242b..44da91712115 100644
> --- a/drivers/scsi/pmcraid.h
> +++ b/drivers/scsi/pmcraid.h
> @@ -542,7 +542,6 @@ struct pmcraid_sglist {
> u32 order;
> u32 num_sg;
> u32 num_dma_sg;
> - u32 buffer_len;
> struct scatterlist scatterlist[1];
> };
>
>
This actually is the same story that we've had with ipr (and, looking at
the code, those two drivers look awfully similar ...).
pmcraid_sglist looks as if it's a hardware-dependent structure, so just
removing one entry from the middle of a structure might not be a good idea.
But this is something for the pmcraid folks to clarify.
Cheers,
Hannes
--
Dr. Hannes ReineckeTeamlead Storage & Networking
h...@suse.de +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
Re: [PATCH v2 6/8] scsi/ipr: Use sgl_alloc_order() and sgl_free_order()
On 10/17/2017 12:49 AM, Bart Van Assche wrote:
> Use the sgl_alloc_order() and sgl_free_order() functions instead
> of open coding these functions.
>
> Signed-off-by: Bart Van Assche
> Reviewed-by: Johannes Thumshirn
> Cc: linux-s...@vger.kernel.org
> Cc: Martin K. Petersen
> Cc: Brian King
> ---
> drivers/scsi/Kconfig | 1 +
> drivers/scsi/ipr.c | 49 -
> drivers/scsi/ipr.h | 2 +-
> 3 files changed, 10 insertions(+), 42 deletions(-)
>
> diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
> index 41366339b950..d11e75e76195 100644
> --- a/drivers/scsi/Kconfig
> +++ b/drivers/scsi/Kconfig
> @@ -1058,6 +1058,7 @@ config SCSI_IPR
> depends on PCI && SCSI && ATA
> select FW_LOADER
> select IRQ_POLL
> + select SGL_ALLOC
> ---help---
> This driver supports the IBM Power Linux family RAID adapters.
> This includes IBM pSeries 5712, 5703, 5709, and 570A, as well
> diff --git a/drivers/scsi/ipr.c b/drivers/scsi/ipr.c
> index f838bd73befa..6fed5db6255e 100644
> --- a/drivers/scsi/ipr.c
> +++ b/drivers/scsi/ipr.c
> @@ -3815,10 +3815,8 @@ static struct device_attribute ipr_iopoll_weight_attr
> = {
> **/
> static struct ipr_sglist *ipr_alloc_ucode_buffer(int buf_len)
> {
> - int sg_size, order, bsize_elem, num_elem, i, j;
> + int sg_size, order;
> struct ipr_sglist *sglist;
> - struct scatterlist *scatterlist;
> - struct page *page;
>
> /* Get the minimum size per scatter/gather element */
> sg_size = buf_len / (IPR_MAX_SGLIST - 1);
> @@ -3826,45 +3824,18 @@ static struct ipr_sglist *ipr_alloc_ucode_buffer(int
> buf_len)
> /* Get the actual size per element */
> order = get_order(sg_size);
>
> - /* Determine the actual number of bytes per element */
> - bsize_elem = PAGE_SIZE * (1 << order);
> -
> - /* Determine the actual number of sg entries needed */
> - if (buf_len % bsize_elem)
> - num_elem = (buf_len / bsize_elem) + 1;
> - else
> - num_elem = buf_len / bsize_elem;
> -
> /* Allocate a scatter/gather list for the DMA */
> - sglist = kzalloc(sizeof(struct ipr_sglist) +
> - (sizeof(struct scatterlist) * (num_elem - 1)),
> - GFP_KERNEL);
> -
> + sglist = kzalloc(sizeof(struct ipr_sglist), GFP_KERNEL);
> if (sglist == NULL) {
> ipr_trace;
> return NULL;
> }
> -
> - scatterlist = sglist->scatterlist;
> - sg_init_table(scatterlist, num_elem);
> -
> sglist->order = order;
> - sglist->num_sg = num_elem;
> -
> - /* Allocate a bunch of sg elements */
> - for (i = 0; i < num_elem; i++) {
> - page = alloc_pages(GFP_KERNEL, order);
> - if (!page) {
> - ipr_trace;
> -
> - /* Free up what we already allocated */
> - for (j = i - 1; j >= 0; j--)
> - __free_pages(sg_page(&scatterlist[j]), order);
> - kfree(sglist);
> - return NULL;
> - }
> -
> - sg_set_page(&scatterlist[i], page, 0, 0);
> + sglist->scatterlist = sgl_alloc_order(buf_len, order, false, GFP_KERNEL,
> + &sglist->num_sg);
> + if (!sglist->scatterlist) {
> + kfree(sglist);
> + return NULL;
> }
>
> return sglist;
> @@ -3882,11 +3853,7 @@ static struct ipr_sglist *ipr_alloc_ucode_buffer(int
> buf_len)
> **/
> static void ipr_free_ucode_buffer(struct ipr_sglist *sglist)
> {
> - int i;
> -
> - for (i = 0; i < sglist->num_sg; i++)
> - __free_pages(sg_page(&sglist->scatterlist[i]), sglist->order);
> -
> + sgl_free_order(sglist->scatterlist, sglist->order);
> kfree(sglist);
> }
>
> diff --git a/drivers/scsi/ipr.h b/drivers/scsi/ipr.h
> index c7f0e9e3cd7d..93570734cbfb 100644
> --- a/drivers/scsi/ipr.h
> +++ b/drivers/scsi/ipr.h
> @@ -1454,7 +1454,7 @@ struct ipr_sglist {
> u32 num_sg;
> u32 num_dma_sg;
> u32 buffer_len;
> - struct scatterlist scatterlist[1];
> + struct scatterlist *scatterlist;
> };
>
> enum ipr_sdt_state {
>
Not sure if this is a valid conversion.
Originally the driver would allocate a single buffer; with this buffer
we have two distinct buffers.
Given that this is used to download the microcode I'm not sure if this
isn't a hardware-dependent structure which requires a single buffer
including the sglist.
Brian, can you shed some light here?
Cheers,
Hannes
--
Dr. Hannes ReineckeTeamlead Storage & Networking
h...@suse.de +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
Re: [PATCH v2 5/8] target: Use sgl_alloc_order() and sgl_free()
On 10/17/2017 12:49 AM, Bart Van Assche wrote:
> Use the sgl_alloc_order() and sgl_free() functions instead of open
> coding these functions.
>
> Signed-off-by: Bart Van Assche
> Cc: Nicholas A. Bellinger
> Cc: Christoph Hellwig
> Cc: Hannes Reinecke
> Cc: Sagi Grimberg
> ---
> drivers/target/Kconfig | 1 +
> drivers/target/target_core_transport.c | 46
> +++---
> 2 files changed, 5 insertions(+), 42 deletions(-)
>
> diff --git a/drivers/target/Kconfig b/drivers/target/Kconfig
> index e2bc99980f75..4c44d7bed01a 100644
> --- a/drivers/target/Kconfig
> +++ b/drivers/target/Kconfig
> @@ -5,6 +5,7 @@ menuconfig TARGET_CORE
> select CONFIGFS_FS
> select CRC_T10DIF
> select BLK_SCSI_REQUEST # only for scsi_command_size_tbl..
> + select SGL_ALLOC
> default n
> help
> Say Y or M here to enable the TCM Storage Engine and ConfigFS enabled
> diff --git a/drivers/target/target_core_transport.c
> b/drivers/target/target_core_transport.c
> index 836d552b0385..9bbd08be9d60 100644
> --- a/drivers/target/target_core_transport.c
> +++ b/drivers/target/target_core_transport.c
> @@ -2293,13 +2293,7 @@ static void target_complete_ok_work(struct work_struct
> *work)
>
> void target_free_sgl(struct scatterlist *sgl, int nents)
> {
> - struct scatterlist *sg;
> - int count;
> -
> - for_each_sg(sgl, sg, nents, count)
> - __free_page(sg_page(sg));
> -
> - kfree(sgl);
> + sgl_free(sgl);
> }
> EXPORT_SYMBOL(target_free_sgl);
>
> @@ -2423,42 +2417,10 @@ int
> target_alloc_sgl(struct scatterlist **sgl, unsigned int *nents, u32 length,
>bool zero_page, bool chainable)
> {
> - struct scatterlist *sg;
> - struct page *page;
> - gfp_t zero_flag = (zero_page) ? __GFP_ZERO : 0;
> - unsigned int nalloc, nent;
> - int i = 0;
> -
> - nalloc = nent = DIV_ROUND_UP(length, PAGE_SIZE);
> - if (chainable)
> - nalloc++;
> - sg = kmalloc_array(nalloc, sizeof(struct scatterlist), GFP_KERNEL);
> - if (!sg)
> - return -ENOMEM;
> + gfp_t gfp = GFP_KERNEL | (zero_page ? __GFP_ZERO : 0);
>
> - sg_init_table(sg, nalloc);
> -
> - while (length) {
> - u32 page_len = min_t(u32, length, PAGE_SIZE);
> - page = alloc_page(GFP_KERNEL | zero_flag);
> - if (!page)
> - goto out;
> -
> - sg_set_page(&sg[i], page, page_len, 0);
> - length -= page_len;
> - i++;
> - }
> - *sgl = sg;
> - *nents = nent;
> - return 0;
> -
> -out:
> - while (i > 0) {
> - i--;
> - __free_page(sg_page(&sg[i]));
> - }
> - kfree(sg);
> - return -ENOMEM;
> + *sgl = sgl_alloc_order(length, 0, chainable, gfp, nents);
> + return *sgl ? 0 : -ENOMEM;
> }
> EXPORT_SYMBOL(target_alloc_sgl);
>
> The calling convention from target_alloc_sgl() is decidedly dodgy.
Can't we convert it into returning the sgl, and remove the first parameter?
Cheers,
Hannes
--
Dr. Hannes ReineckeTeamlead Storage & Networking
h...@suse.de +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)
Re: [PATCH v2 4/8] nvmet/rdma: Use sgl_alloc() and sgl_free()
On 10/17/2017 12:49 AM, Bart Van Assche wrote: > Use the sgl_alloc() and sgl_free() functions instead of open coding > these functions. > > Signed-off-by: Bart Van Assche > Reviewed-by: Johannes Thumshirn > Cc: Keith Busch > Cc: Christoph Hellwig > Cc: Sagi Grimberg > --- > drivers/nvme/target/Kconfig | 1 + > drivers/nvme/target/rdma.c | 63 > +++-- > 2 files changed, 5 insertions(+), 59 deletions(-) > Reviewed-by: Hannes Reinecke Cheers, Hannes -- Dr. Hannes ReineckeTeamlead Storage & Networking h...@suse.de +49 911 74053 688 SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton HRB 21284 (AG Nürnberg)
Re: [PATCH v2 3/8] nvmet/fc: Use sgl_alloc() and sgl_free()
On 10/17/2017 12:49 AM, Bart Van Assche wrote: > Use the sgl_alloc() and sgl_free() functions instead of open coding > these functions. > > Signed-off-by: Bart Van Assche > Reviewed-by: Johannes Thumshirn > Cc: Keith Busch > Cc: Christoph Hellwig > Cc: James Smart > Cc: Sagi Grimberg > --- > drivers/nvme/target/Kconfig | 1 + > drivers/nvme/target/fc.c| 36 ++-- > 2 files changed, 3 insertions(+), 34 deletions(-) > Reviewed-by: Hannes Reinecke Cheers, Hannes -- Dr. Hannes ReineckeTeamlead Storage & Networking h...@suse.de +49 911 74053 688 SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton HRB 21284 (AG Nürnberg)
Re: [PATCH v2 1/8] lib/scatterlist: Introduce sgl_alloc() and sgl_free()
On 10/17/2017 12:49 AM, Bart Van Assche wrote: > Many kernel drivers contain code that allocates and frees both a > scatterlist and the pages that populate that scatterlist. > Introduce functions in lib/scatterlist.c that perform these tasks > instead of duplicating this functionality in multiple drivers. > Only include these functions in the build if CONFIG_SGL_ALLOC=y > to avoid that the kernel size increases if this functionality is > not used. > > Signed-off-by: Bart Van Assche > --- > include/linux/scatterlist.h | 10 + > lib/Kconfig | 4 ++ > lib/scatterlist.c | 105 > > 3 files changed, 119 insertions(+) > Reviewed-by: Hannes Reinecke Cheers, Hannes -- Dr. Hannes ReineckeTeamlead Storage & Networking h...@suse.de +49 911 74053 688 SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton HRB 21284 (AG Nürnberg)
Re: [PATCH v2 2/8] crypto: scompress - use sgl_alloc() and sgl_free()
On 10/17/2017 12:49 AM, Bart Van Assche wrote: > Use the sgl_alloc() and sgl_free() functions instead of open coding > these functions. > > Signed-off-by: Bart Van Assche > Cc: Ard Biesheuvel > Cc: Herbert Xu > --- > crypto/Kconfig | 1 + > crypto/scompress.c | 51 ++- > 2 files changed, 3 insertions(+), 49 deletions(-) > Reviewed-by: Hannes Reinecke Cheers, Hannes -- Dr. Hannes ReineckeTeamlead Storage & Networking h...@suse.de +49 911 74053 688 SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton HRB 21284 (AG Nürnberg)
New Linux accelerators discussion list [was: Re: Fostering linux community collaboration on hardware accelerators]
On 17/10/17 01:07, Jonathan Cameron wrote: So as ever with a linux community focusing on a particular topic, the obvious solution is a mailing list. There are a number of options on how do this. 1) Ask one of the industry bodies to host? Who? 2) Put together a compelling argument for linux-accelerat...@vger.kernel.org as probably the most generic location for such a list. Happy to offer linux-accelerat...@lists.ozlabs.org, which I can get set up immediately (and if we want patchwork, patchwork.ozlabs.org is available as always, no matter where the list is hosted). That would be great! Thanks for doing this. Much easier to find out what such a list is useful for by the practical option of having a list and see what people do with it. [+ LKML] We now have a mailing list: List: linux-accelerat...@lists.ozlabs.org Info: https://lists.ozlabs.org/listinfo/linux-accelerators Archives: https://lists.ozlabs.org/pipermail/linux-accelerators I haven't set up Patchwork as yet, but if people think that's a good idea I can get that done too. Andrew -- Andrew Donnellan OzLabs, ADL Canberra andrew.donnel...@au1.ibm.com IBM Australia Limited
[PATCH] hwrng/xgene-rng: Convert timers to use timer_setup()
In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.
Cc: Matt Mackall
Cc: Herbert Xu
Cc: linux-crypto@vger.kernel.org
Signed-off-by: Kees Cook
---
drivers/char/hw_random/xgene-rng.c | 8 +++-
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/drivers/char/hw_random/xgene-rng.c
b/drivers/char/hw_random/xgene-rng.c
index 3c77645405e5..71755790c32b 100644
--- a/drivers/char/hw_random/xgene-rng.c
+++ b/drivers/char/hw_random/xgene-rng.c
@@ -100,9 +100,9 @@ struct xgene_rng_dev {
struct clk *clk;
};
-static void xgene_rng_expired_timer(unsigned long arg)
+static void xgene_rng_expired_timer(struct timer_list *t)
{
- struct xgene_rng_dev *ctx = (struct xgene_rng_dev *) arg;
+ struct xgene_rng_dev *ctx = from_timer(ctx, t, failure_timer);
/* Clear failure counter as timer expired */
disable_irq(ctx->irq);
@@ -113,8 +113,6 @@ static void xgene_rng_expired_timer(unsigned long arg)
static void xgene_rng_start_timer(struct xgene_rng_dev *ctx)
{
- ctx->failure_timer.data = (unsigned long) ctx;
- ctx->failure_timer.function = xgene_rng_expired_timer;
ctx->failure_timer.expires = jiffies + 120 * HZ;
add_timer(&ctx->failure_timer);
}
@@ -292,7 +290,7 @@ static int xgene_rng_init(struct hwrng *rng)
struct xgene_rng_dev *ctx = (struct xgene_rng_dev *) rng->priv;
ctx->failure_cnt = 0;
- init_timer(&ctx->failure_timer);
+ timer_setup(&ctx->failure_timer, xgene_rng_expired_timer, 0);
ctx->revision = readl(ctx->csr_base + RNG_EIP_REV);
--
2.7.4
--
Kees Cook
Pixel Security
Re: vmx ghash buggy on ppc64le
Tue, Sep 19, 2017 at 11:18:45AM +0800, Herbert Xu wrote: > Hi: > > I have received a report that ghash on ppc64le does not interoperate > with other implementations of ghash, e.g., on x86-64. > > https://bugzilla.redhat.com/show_bug.cgi?id=1490972 > > Could you guys take a look at this and see if this is a bug in > the mainline vmx driver too? > > Thanks! > -- > Email: Herbert Xu > Home Page: http://gondor.apana.org.au/~herbert/ > PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt > Hi Herbert, It seems that this bug is affected by the fact that ctr mode is using blkcipher for the fallback. I just sent a patch [1] to make it use skcipher and, since then, ipsec with ikev2 and ghash are working fine for ppc64 and ppc64le. Best regards, -- Paulo Flabiano Smorigo IBM Linux Technology Center [1] https://lkml.org/lkml/2017/10/16/911
[PATCH] crypto: vmx - Use skcipher for ctr fallback
Signed-off-by: Paulo Flabiano Smorigo
---
drivers/crypto/vmx/aes_ctr.c | 33 +
1 file changed, 17 insertions(+), 16 deletions(-)
diff --git a/drivers/crypto/vmx/aes_ctr.c b/drivers/crypto/vmx/aes_ctr.c
index 17d8421..fc60d00 100644
--- a/drivers/crypto/vmx/aes_ctr.c
+++ b/drivers/crypto/vmx/aes_ctr.c
@@ -27,21 +27,23 @@
#include
#include
#include
+#include
+
#include "aesp8-ppc.h"
struct p8_aes_ctr_ctx {
- struct crypto_blkcipher *fallback;
+ struct crypto_skcipher *fallback;
struct aes_key enc_key;
};
static int p8_aes_ctr_init(struct crypto_tfm *tfm)
{
const char *alg = crypto_tfm_alg_name(tfm);
- struct crypto_blkcipher *fallback;
+ struct crypto_skcipher *fallback;
struct p8_aes_ctr_ctx *ctx = crypto_tfm_ctx(tfm);
- fallback =
- crypto_alloc_blkcipher(alg, 0, CRYPTO_ALG_NEED_FALLBACK);
+ fallback = crypto_alloc_skcipher(alg, 0,
+ CRYPTO_ALG_ASYNC | CRYPTO_ALG_NEED_FALLBACK);
if (IS_ERR(fallback)) {
printk(KERN_ERR
"Failed to allocate transformation for '%s': %ld\n",
@@ -49,11 +51,11 @@ static int p8_aes_ctr_init(struct crypto_tfm *tfm)
return PTR_ERR(fallback);
}
printk(KERN_INFO "Using '%s' as fallback implementation.\n",
- crypto_tfm_alg_driver_name((struct crypto_tfm *) fallback));
+ crypto_skcipher_driver_name(fallback));
- crypto_blkcipher_set_flags(
+ crypto_skcipher_set_flags(
fallback,
- crypto_blkcipher_get_flags((struct crypto_blkcipher *)tfm));
+ crypto_skcipher_get_flags((struct crypto_skcipher *)tfm));
ctx->fallback = fallback;
return 0;
@@ -64,7 +66,7 @@ static void p8_aes_ctr_exit(struct crypto_tfm *tfm)
struct p8_aes_ctr_ctx *ctx = crypto_tfm_ctx(tfm);
if (ctx->fallback) {
- crypto_free_blkcipher(ctx->fallback);
+ crypto_free_skcipher(ctx->fallback);
ctx->fallback = NULL;
}
}
@@ -83,7 +85,7 @@ static int p8_aes_ctr_setkey(struct crypto_tfm *tfm, const u8
*key,
pagefault_enable();
preempt_enable();
- ret += crypto_blkcipher_setkey(ctx->fallback, key, keylen);
+ ret += crypto_skcipher_setkey(ctx->fallback, key, keylen);
return ret;
}
@@ -117,15 +119,14 @@ static int p8_aes_ctr_crypt(struct blkcipher_desc *desc,
struct blkcipher_walk walk;
struct p8_aes_ctr_ctx *ctx =
crypto_tfm_ctx(crypto_blkcipher_tfm(desc->tfm));
- struct blkcipher_desc fallback_desc = {
- .tfm = ctx->fallback,
- .info = desc->info,
- .flags = desc->flags
- };
if (in_interrupt()) {
- ret = crypto_blkcipher_encrypt(&fallback_desc, dst, src,
- nbytes);
+ SKCIPHER_REQUEST_ON_STACK(req, ctx->fallback);
+ skcipher_request_set_tfm(req, ctx->fallback);
+ skcipher_request_set_callback(req, desc->flags, NULL, NULL);
+ skcipher_request_set_crypt(req, src, dst, nbytes, desc->info);
+ ret = crypto_skcipher_encrypt(req);
+ skcipher_request_zero(req);
} else {
blkcipher_walk_init(&walk, dst, src, nbytes);
ret = blkcipher_walk_virt_block(desc, &walk, AES_BLOCK_SIZE);
--
2.9.4
[PATCH v2 3/8] nvmet/fc: Use sgl_alloc() and sgl_free()
Use the sgl_alloc() and sgl_free() functions instead of open coding
these functions.
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Cc: Keith Busch
Cc: Christoph Hellwig
Cc: James Smart
Cc: Sagi Grimberg
---
drivers/nvme/target/Kconfig | 1 +
drivers/nvme/target/fc.c| 36 ++--
2 files changed, 3 insertions(+), 34 deletions(-)
diff --git a/drivers/nvme/target/Kconfig b/drivers/nvme/target/Kconfig
index 03e4ab65fe77..4d9715630e21 100644
--- a/drivers/nvme/target/Kconfig
+++ b/drivers/nvme/target/Kconfig
@@ -39,6 +39,7 @@ config NVME_TARGET_FC
tristate "NVMe over Fabrics FC target driver"
depends on NVME_TARGET
depends on HAS_DMA
+ select SGL_ALLOC
help
This enables the NVMe FC target support, which allows exporting NVMe
devices over FC.
diff --git a/drivers/nvme/target/fc.c b/drivers/nvme/target/fc.c
index 58e010bdda3e..20f96038c272 100644
--- a/drivers/nvme/target/fc.c
+++ b/drivers/nvme/target/fc.c
@@ -1683,31 +1683,12 @@ static int
nvmet_fc_alloc_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
{
struct scatterlist *sg;
- struct page *page;
unsigned int nent;
- u32 page_len, length;
- int i = 0;
- length = fod->total_length;
- nent = DIV_ROUND_UP(length, PAGE_SIZE);
- sg = kmalloc_array(nent, sizeof(struct scatterlist), GFP_KERNEL);
+ sg = sgl_alloc(fod->total_length, GFP_KERNEL, &nent);
if (!sg)
goto out;
- sg_init_table(sg, nent);
-
- while (length) {
- page_len = min_t(u32, length, PAGE_SIZE);
-
- page = alloc_page(GFP_KERNEL);
- if (!page)
- goto out_free_pages;
-
- sg_set_page(&sg[i], page, page_len, 0);
- length -= page_len;
- i++;
- }
-
fod->data_sg = sg;
fod->data_sg_cnt = nent;
fod->data_sg_cnt = fc_dma_map_sg(fod->tgtport->dev, sg, nent,
@@ -1717,14 +1698,6 @@ nvmet_fc_alloc_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
return 0;
-out_free_pages:
- while (i > 0) {
- i--;
- __free_page(sg_page(&sg[i]));
- }
- kfree(sg);
- fod->data_sg = NULL;
- fod->data_sg_cnt = 0;
out:
return NVME_SC_INTERNAL;
}
@@ -1732,18 +1705,13 @@ nvmet_fc_alloc_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
static void
nvmet_fc_free_tgt_pgs(struct nvmet_fc_fcp_iod *fod)
{
- struct scatterlist *sg;
- int count;
-
if (!fod->data_sg || !fod->data_sg_cnt)
return;
fc_dma_unmap_sg(fod->tgtport->dev, fod->data_sg, fod->data_sg_cnt,
((fod->io_dir == NVMET_FCP_WRITE) ?
DMA_FROM_DEVICE : DMA_TO_DEVICE));
- for_each_sg(fod->data_sg, sg, fod->data_sg_cnt, count)
- __free_page(sg_page(sg));
- kfree(fod->data_sg);
+ sgl_free(fod->data_sg);
fod->data_sg = NULL;
fod->data_sg_cnt = 0;
}
--
2.14.2
[PATCH v2 7/8] scsi/pmcraid: Remove an unused structure member
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Cc: linux-s...@vger.kernel.org
Cc: Martin K. Petersen
Cc: Anil Ravindranath
---
drivers/scsi/pmcraid.h | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/scsi/pmcraid.h b/drivers/scsi/pmcraid.h
index 8bfac72a242b..44da91712115 100644
--- a/drivers/scsi/pmcraid.h
+++ b/drivers/scsi/pmcraid.h
@@ -542,7 +542,6 @@ struct pmcraid_sglist {
u32 order;
u32 num_sg;
u32 num_dma_sg;
- u32 buffer_len;
struct scatterlist scatterlist[1];
};
--
2.14.2
[PATCH v2 6/8] scsi/ipr: Use sgl_alloc_order() and sgl_free_order()
Use the sgl_alloc_order() and sgl_free_order() functions instead
of open coding these functions.
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Cc: linux-s...@vger.kernel.org
Cc: Martin K. Petersen
Cc: Brian King
---
drivers/scsi/Kconfig | 1 +
drivers/scsi/ipr.c | 49 -
drivers/scsi/ipr.h | 2 +-
3 files changed, 10 insertions(+), 42 deletions(-)
diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
index 41366339b950..d11e75e76195 100644
--- a/drivers/scsi/Kconfig
+++ b/drivers/scsi/Kconfig
@@ -1058,6 +1058,7 @@ config SCSI_IPR
depends on PCI && SCSI && ATA
select FW_LOADER
select IRQ_POLL
+ select SGL_ALLOC
---help---
This driver supports the IBM Power Linux family RAID adapters.
This includes IBM pSeries 5712, 5703, 5709, and 570A, as well
diff --git a/drivers/scsi/ipr.c b/drivers/scsi/ipr.c
index f838bd73befa..6fed5db6255e 100644
--- a/drivers/scsi/ipr.c
+++ b/drivers/scsi/ipr.c
@@ -3815,10 +3815,8 @@ static struct device_attribute ipr_iopoll_weight_attr = {
**/
static struct ipr_sglist *ipr_alloc_ucode_buffer(int buf_len)
{
- int sg_size, order, bsize_elem, num_elem, i, j;
+ int sg_size, order;
struct ipr_sglist *sglist;
- struct scatterlist *scatterlist;
- struct page *page;
/* Get the minimum size per scatter/gather element */
sg_size = buf_len / (IPR_MAX_SGLIST - 1);
@@ -3826,45 +3824,18 @@ static struct ipr_sglist *ipr_alloc_ucode_buffer(int
buf_len)
/* Get the actual size per element */
order = get_order(sg_size);
- /* Determine the actual number of bytes per element */
- bsize_elem = PAGE_SIZE * (1 << order);
-
- /* Determine the actual number of sg entries needed */
- if (buf_len % bsize_elem)
- num_elem = (buf_len / bsize_elem) + 1;
- else
- num_elem = buf_len / bsize_elem;
-
/* Allocate a scatter/gather list for the DMA */
- sglist = kzalloc(sizeof(struct ipr_sglist) +
-(sizeof(struct scatterlist) * (num_elem - 1)),
-GFP_KERNEL);
-
+ sglist = kzalloc(sizeof(struct ipr_sglist), GFP_KERNEL);
if (sglist == NULL) {
ipr_trace;
return NULL;
}
-
- scatterlist = sglist->scatterlist;
- sg_init_table(scatterlist, num_elem);
-
sglist->order = order;
- sglist->num_sg = num_elem;
-
- /* Allocate a bunch of sg elements */
- for (i = 0; i < num_elem; i++) {
- page = alloc_pages(GFP_KERNEL, order);
- if (!page) {
- ipr_trace;
-
- /* Free up what we already allocated */
- for (j = i - 1; j >= 0; j--)
- __free_pages(sg_page(&scatterlist[j]), order);
- kfree(sglist);
- return NULL;
- }
-
- sg_set_page(&scatterlist[i], page, 0, 0);
+ sglist->scatterlist = sgl_alloc_order(buf_len, order, false, GFP_KERNEL,
+ &sglist->num_sg);
+ if (!sglist->scatterlist) {
+ kfree(sglist);
+ return NULL;
}
return sglist;
@@ -3882,11 +3853,7 @@ static struct ipr_sglist *ipr_alloc_ucode_buffer(int
buf_len)
**/
static void ipr_free_ucode_buffer(struct ipr_sglist *sglist)
{
- int i;
-
- for (i = 0; i < sglist->num_sg; i++)
- __free_pages(sg_page(&sglist->scatterlist[i]), sglist->order);
-
+ sgl_free_order(sglist->scatterlist, sglist->order);
kfree(sglist);
}
diff --git a/drivers/scsi/ipr.h b/drivers/scsi/ipr.h
index c7f0e9e3cd7d..93570734cbfb 100644
--- a/drivers/scsi/ipr.h
+++ b/drivers/scsi/ipr.h
@@ -1454,7 +1454,7 @@ struct ipr_sglist {
u32 num_sg;
u32 num_dma_sg;
u32 buffer_len;
- struct scatterlist scatterlist[1];
+ struct scatterlist *scatterlist;
};
enum ipr_sdt_state {
--
2.14.2
[PATCH v2 5/8] target: Use sgl_alloc_order() and sgl_free()
Use the sgl_alloc_order() and sgl_free() functions instead of open
coding these functions.
Signed-off-by: Bart Van Assche
Cc: Nicholas A. Bellinger
Cc: Christoph Hellwig
Cc: Hannes Reinecke
Cc: Sagi Grimberg
---
drivers/target/Kconfig | 1 +
drivers/target/target_core_transport.c | 46 +++---
2 files changed, 5 insertions(+), 42 deletions(-)
diff --git a/drivers/target/Kconfig b/drivers/target/Kconfig
index e2bc99980f75..4c44d7bed01a 100644
--- a/drivers/target/Kconfig
+++ b/drivers/target/Kconfig
@@ -5,6 +5,7 @@ menuconfig TARGET_CORE
select CONFIGFS_FS
select CRC_T10DIF
select BLK_SCSI_REQUEST # only for scsi_command_size_tbl..
+ select SGL_ALLOC
default n
help
Say Y or M here to enable the TCM Storage Engine and ConfigFS enabled
diff --git a/drivers/target/target_core_transport.c
b/drivers/target/target_core_transport.c
index 836d552b0385..9bbd08be9d60 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -2293,13 +2293,7 @@ static void target_complete_ok_work(struct work_struct
*work)
void target_free_sgl(struct scatterlist *sgl, int nents)
{
- struct scatterlist *sg;
- int count;
-
- for_each_sg(sgl, sg, nents, count)
- __free_page(sg_page(sg));
-
- kfree(sgl);
+ sgl_free(sgl);
}
EXPORT_SYMBOL(target_free_sgl);
@@ -2423,42 +2417,10 @@ int
target_alloc_sgl(struct scatterlist **sgl, unsigned int *nents, u32 length,
bool zero_page, bool chainable)
{
- struct scatterlist *sg;
- struct page *page;
- gfp_t zero_flag = (zero_page) ? __GFP_ZERO : 0;
- unsigned int nalloc, nent;
- int i = 0;
-
- nalloc = nent = DIV_ROUND_UP(length, PAGE_SIZE);
- if (chainable)
- nalloc++;
- sg = kmalloc_array(nalloc, sizeof(struct scatterlist), GFP_KERNEL);
- if (!sg)
- return -ENOMEM;
+ gfp_t gfp = GFP_KERNEL | (zero_page ? __GFP_ZERO : 0);
- sg_init_table(sg, nalloc);
-
- while (length) {
- u32 page_len = min_t(u32, length, PAGE_SIZE);
- page = alloc_page(GFP_KERNEL | zero_flag);
- if (!page)
- goto out;
-
- sg_set_page(&sg[i], page, page_len, 0);
- length -= page_len;
- i++;
- }
- *sgl = sg;
- *nents = nent;
- return 0;
-
-out:
- while (i > 0) {
- i--;
- __free_page(sg_page(&sg[i]));
- }
- kfree(sg);
- return -ENOMEM;
+ *sgl = sgl_alloc_order(length, 0, chainable, gfp, nents);
+ return *sgl ? 0 : -ENOMEM;
}
EXPORT_SYMBOL(target_alloc_sgl);
--
2.14.2
[PATCH v2 2/8] crypto: scompress - use sgl_alloc() and sgl_free()
Use the sgl_alloc() and sgl_free() functions instead of open coding
these functions.
Signed-off-by: Bart Van Assche
Cc: Ard Biesheuvel
Cc: Herbert Xu
---
crypto/Kconfig | 1 +
crypto/scompress.c | 51 ++-
2 files changed, 3 insertions(+), 49 deletions(-)
diff --git a/crypto/Kconfig b/crypto/Kconfig
index 0a121f9ddf8e..a0667dd284ff 100644
--- a/crypto/Kconfig
+++ b/crypto/Kconfig
@@ -105,6 +105,7 @@ config CRYPTO_KPP
config CRYPTO_ACOMP2
tristate
select CRYPTO_ALGAPI2
+ select SGL_ALLOC
config CRYPTO_ACOMP
tristate
diff --git a/crypto/scompress.c b/crypto/scompress.c
index 2075e2c4e7df..968bbcf65c94 100644
--- a/crypto/scompress.c
+++ b/crypto/scompress.c
@@ -140,53 +140,6 @@ static int crypto_scomp_init_tfm(struct crypto_tfm *tfm)
return ret;
}
-static void crypto_scomp_sg_free(struct scatterlist *sgl)
-{
- int i, n;
- struct page *page;
-
- if (!sgl)
- return;
-
- n = sg_nents(sgl);
- for_each_sg(sgl, sgl, n, i) {
- page = sg_page(sgl);
- if (page)
- __free_page(page);
- }
-
- kfree(sgl);
-}
-
-static struct scatterlist *crypto_scomp_sg_alloc(size_t size, gfp_t gfp)
-{
- struct scatterlist *sgl;
- struct page *page;
- int i, n;
-
- n = ((size - 1) >> PAGE_SHIFT) + 1;
-
- sgl = kmalloc_array(n, sizeof(struct scatterlist), gfp);
- if (!sgl)
- return NULL;
-
- sg_init_table(sgl, n);
-
- for (i = 0; i < n; i++) {
- page = alloc_page(gfp);
- if (!page)
- goto err;
- sg_set_page(sgl + i, page, PAGE_SIZE, 0);
- }
-
- return sgl;
-
-err:
- sg_mark_end(sgl + i);
- crypto_scomp_sg_free(sgl);
- return NULL;
-}
-
static int scomp_acomp_comp_decomp(struct acomp_req *req, int dir)
{
struct crypto_acomp *tfm = crypto_acomp_reqtfm(req);
@@ -220,7 +173,7 @@ static int scomp_acomp_comp_decomp(struct acomp_req *req,
int dir)
scratch_dst, &req->dlen, *ctx);
if (!ret) {
if (!req->dst) {
- req->dst = crypto_scomp_sg_alloc(req->dlen, GFP_ATOMIC);
+ req->dst = sgl_alloc(req->dlen, GFP_ATOMIC, NULL);
if (!req->dst)
goto out;
}
@@ -274,7 +227,7 @@ int crypto_init_scomp_ops_async(struct crypto_tfm *tfm)
crt->compress = scomp_acomp_compress;
crt->decompress = scomp_acomp_decompress;
- crt->dst_free = crypto_scomp_sg_free;
+ crt->dst_free = sgl_free;
crt->reqsize = sizeof(void *);
return 0;
--
2.14.2
[PATCH v2 8/8] scsi/pmcraid: Use sgl_alloc_order() and sgl_free_order()
Use the sgl_alloc_order() and sgl_free_order() functions instead
of open coding these functions.
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Cc: linux-s...@vger.kernel.org
Cc: Martin K. Petersen
Cc: Anil Ravindranath
---
drivers/scsi/Kconfig | 1 +
drivers/scsi/pmcraid.c | 43 ---
drivers/scsi/pmcraid.h | 2 +-
3 files changed, 6 insertions(+), 40 deletions(-)
diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
index d11e75e76195..632200ec36a6 100644
--- a/drivers/scsi/Kconfig
+++ b/drivers/scsi/Kconfig
@@ -1576,6 +1576,7 @@ config ZFCP
config SCSI_PMCRAID
tristate "PMC SIERRA Linux MaxRAID adapter support"
depends on PCI && SCSI && NET
+ select SGL_ALLOC
---help---
This driver supports the PMC SIERRA MaxRAID adapters.
diff --git a/drivers/scsi/pmcraid.c b/drivers/scsi/pmcraid.c
index b4d6cd8cd1ad..95fc0352f9bb 100644
--- a/drivers/scsi/pmcraid.c
+++ b/drivers/scsi/pmcraid.c
@@ -3232,12 +3232,7 @@ static int pmcraid_build_ioadl(
*/
static void pmcraid_free_sglist(struct pmcraid_sglist *sglist)
{
- int i;
-
- for (i = 0; i < sglist->num_sg; i++)
- __free_pages(sg_page(&(sglist->scatterlist[i])),
-sglist->order);
-
+ sgl_free_order(sglist->scatterlist, sglist->order);
kfree(sglist);
}
@@ -3254,50 +3249,20 @@ static void pmcraid_free_sglist(struct pmcraid_sglist
*sglist)
static struct pmcraid_sglist *pmcraid_alloc_sglist(int buflen)
{
struct pmcraid_sglist *sglist;
- struct scatterlist *scatterlist;
- struct page *page;
- int num_elem, i, j;
int sg_size;
int order;
- int bsize_elem;
sg_size = buflen / (PMCRAID_MAX_IOADLS - 1);
order = (sg_size > 0) ? get_order(sg_size) : 0;
- bsize_elem = PAGE_SIZE * (1 << order);
-
- /* Determine the actual number of sg entries needed */
- if (buflen % bsize_elem)
- num_elem = (buflen / bsize_elem) + 1;
- else
- num_elem = buflen / bsize_elem;
/* Allocate a scatter/gather list for the DMA */
- sglist = kzalloc(sizeof(struct pmcraid_sglist) +
-(sizeof(struct scatterlist) * (num_elem - 1)),
-GFP_KERNEL);
-
+ sglist = kzalloc(sizeof(struct pmcraid_sglist), GFP_KERNEL);
if (sglist == NULL)
return NULL;
- scatterlist = sglist->scatterlist;
- sg_init_table(scatterlist, num_elem);
sglist->order = order;
- sglist->num_sg = num_elem;
- sg_size = buflen;
-
- for (i = 0; i < num_elem; i++) {
- page = alloc_pages(GFP_KERNEL|GFP_DMA|__GFP_ZERO, order);
- if (!page) {
- for (j = i - 1; j >= 0; j--)
- __free_pages(sg_page(&scatterlist[j]), order);
- kfree(sglist);
- return NULL;
- }
-
- sg_set_page(&scatterlist[i], page,
- sg_size < bsize_elem ? sg_size : bsize_elem, 0);
- sg_size -= bsize_elem;
- }
+ sgl_alloc_order(buflen, order, false,
+ GFP_KERNEL | GFP_DMA | __GFP_ZERO, &sglist->num_sg);
return sglist;
}
diff --git a/drivers/scsi/pmcraid.h b/drivers/scsi/pmcraid.h
index 44da91712115..754ef30927e2 100644
--- a/drivers/scsi/pmcraid.h
+++ b/drivers/scsi/pmcraid.h
@@ -542,7 +542,7 @@ struct pmcraid_sglist {
u32 order;
u32 num_sg;
u32 num_dma_sg;
- struct scatterlist scatterlist[1];
+ struct scatterlist *scatterlist;
};
/* page D0 inquiry data of focal point resource */
--
2.14.2
[PATCH v2 1/8] lib/scatterlist: Introduce sgl_alloc() and sgl_free()
Many kernel drivers contain code that allocates and frees both a
scatterlist and the pages that populate that scatterlist.
Introduce functions in lib/scatterlist.c that perform these tasks
instead of duplicating this functionality in multiple drivers.
Only include these functions in the build if CONFIG_SGL_ALLOC=y
to avoid that the kernel size increases if this functionality is
not used.
Signed-off-by: Bart Van Assche
---
include/linux/scatterlist.h | 10 +
lib/Kconfig | 4 ++
lib/scatterlist.c | 105
3 files changed, 119 insertions(+)
diff --git a/include/linux/scatterlist.h b/include/linux/scatterlist.h
index 4b3286ac60c8..3642511918d5 100644
--- a/include/linux/scatterlist.h
+++ b/include/linux/scatterlist.h
@@ -266,6 +266,16 @@ int sg_alloc_table_from_pages(struct sg_table *sgt,
unsigned long offset, unsigned long size,
gfp_t gfp_mask);
+#ifdef CONFIG_SGL_ALLOC
+struct scatterlist *sgl_alloc_order(unsigned long long length,
+ unsigned int order, bool chainable,
+ gfp_t gfp, unsigned int *nent_p);
+struct scatterlist *sgl_alloc(unsigned long long length, gfp_t gfp,
+ unsigned int *nent_p);
+void sgl_free_order(struct scatterlist *sgl, int order);
+void sgl_free(struct scatterlist *sgl);
+#endif /* CONFIG_SGL_ALLOC */
+
size_t sg_copy_buffer(struct scatterlist *sgl, unsigned int nents, void *buf,
size_t buflen, off_t skip, bool to_buffer);
diff --git a/lib/Kconfig b/lib/Kconfig
index b1445b22a6de..8396c4cfa1ab 100644
--- a/lib/Kconfig
+++ b/lib/Kconfig
@@ -413,6 +413,10 @@ config HAS_DMA
depends on !NO_DMA
default y
+config SGL_ALLOC
+ bool
+ default n
+
config DMA_NOOP_OPS
bool
depends on HAS_DMA && (!64BIT || ARCH_DMA_ADDR_T_64BIT)
diff --git a/lib/scatterlist.c b/lib/scatterlist.c
index be7b4dd6b68d..e2b5a78ceb44 100644
--- a/lib/scatterlist.c
+++ b/lib/scatterlist.c
@@ -433,6 +433,111 @@ int sg_alloc_table_from_pages(struct sg_table *sgt,
}
EXPORT_SYMBOL(sg_alloc_table_from_pages);
+#ifdef CONFIG_SGL_ALLOC
+
+/**
+ * sgl_alloc_order - allocate a scatterlist and its pages
+ * @length: Length in bytes of the scatterlist. Must be at least one
+ * @order: Second argument for alloc_pages()
+ * @chainable: Whether or not to allocate an extra element in the scatterlist
+ * for scatterlist chaining purposes
+ * @gfp: Memory allocation flags
+ * @nent_p: [out] Number of entries in the scatterlist that have pages
+ *
+ * Returns: A pointer to an initialized scatterlist or %NULL upon failure.
+ */
+struct scatterlist *sgl_alloc_order(unsigned long long length,
+ unsigned int order, bool chainable,
+ gfp_t gfp, unsigned int *nent_p)
+{
+ struct scatterlist *sgl, *sg;
+ struct page *page;
+ unsigned int nent, nalloc;
+ u32 elem_len;
+
+ nent = round_up(length, PAGE_SIZE << order) >> (PAGE_SHIFT + order);
+ /* Check for integer overflow */
+ if (length > (nent << (PAGE_SHIFT + order)))
+ return NULL;
+ nalloc = nent;
+ if (chainable) {
+ /* Check for integer overflow */
+ if (nalloc + 1 < nalloc)
+ return NULL;
+ nalloc++;
+ }
+ sgl = kmalloc_array(nalloc, sizeof(struct scatterlist),
+ (gfp & ~GFP_DMA) | __GFP_ZERO);
+ if (!sgl)
+ return NULL;
+
+ sg_init_table(sgl, nent);
+ sg = sgl;
+ while (length) {
+ elem_len = min_t(u64, length, PAGE_SIZE << order);
+ page = alloc_pages(gfp, order);
+ if (!page) {
+ sgl_free(sgl);
+ return NULL;
+ }
+
+ sg_set_page(sg, page, elem_len, 0);
+ length -= elem_len;
+ sg = sg_next(sg);
+ }
+ WARN_ON_ONCE(sg);
+ if (nent_p)
+ *nent_p = nent;
+ return sgl;
+}
+EXPORT_SYMBOL(sgl_alloc_order);
+
+/**
+ * sgl_alloc - allocate a scatterlist and its pages
+ * @length: Length in bytes of the scatterlist
+ * @gfp: Memory allocation flags
+ * @nent_p: [out] Number of entries in the scatterlist
+ *
+ * Returns: A pointer to an initialized scatterlist or %NULL upon failure.
+ */
+struct scatterlist *sgl_alloc(unsigned long long length, gfp_t gfp,
+ unsigned int *nent_p)
+{
+ return sgl_alloc_order(length, 0, false, gfp, nent_p);
+}
+EXPORT_SYMBOL(sgl_alloc);
+
+/**
+ * sgl_free_order - free a scatterlist and its pages
+ * @sgl: Scatterlist with one or more elements
+ * @order: Second argument for __free_pages()
+ */
+void sgl_free_order(struct scatterlist *sgl, int order)
+{
+ struct scatterlist *sg;
+ struct page *page;
+
+
[PATCH v2 0/8] Introduce sgl_alloc() and sgl_free()
Hello Jens, As you know there are multiple drivers that both allocate a scatter/gather list and populate that list with pages. This patch series moves the code for allocating and freeing such scatterlists from these drivers into lib/scatterlist.c. Please consider this patch series for kernel v4.15. Notes: - Only the ib_srpt driver changes have been tested but none of the other drivers have been retested. - The next step is to introduce a caching mechanism for these scatterlists and make the nvmet/rdma and SCSI target drivers use that caching mechanism since for these drivers sgl allocation occurs in the hot path. Changes compared to v1: - Moved the sgl_alloc*() and sgl_free() functions from a new source file into lib/scatterlist.c. - Changed the argument order for the sgl_alloc*() functions such that the (pointer to) the output argument comes last. Thanks, Bart. Bart Van Assche (8): lib/scatterlist: Introduce sgl_alloc() and sgl_free() crypto: scompress - use sgl_alloc() and sgl_free() nvmet/fc: Use sgl_alloc() and sgl_free() nvmet/rdma: Use sgl_alloc() and sgl_free() target: Use sgl_alloc_order() and sgl_free() scsi/ipr: Use sgl_alloc_order() and sgl_free_order() scsi/pmcraid: Remove an unused structure member scsi/pmcraid: Use sgl_alloc_order() and sgl_free_order() crypto/Kconfig | 1 + crypto/scompress.c | 51 +--- drivers/nvme/target/Kconfig| 2 + drivers/nvme/target/fc.c | 36 +-- drivers/nvme/target/rdma.c | 63 ++-- drivers/scsi/Kconfig | 2 + drivers/scsi/ipr.c | 49 +++ drivers/scsi/ipr.h | 2 +- drivers/scsi/pmcraid.c | 43 ++ drivers/scsi/pmcraid.h | 3 +- drivers/target/Kconfig | 1 + drivers/target/target_core_transport.c | 46 ++- include/linux/scatterlist.h| 10 lib/Kconfig| 4 ++ lib/scatterlist.c | 105 + 15 files changed, 151 insertions(+), 267 deletions(-) -- 2.14.2
[PATCH v2 4/8] nvmet/rdma: Use sgl_alloc() and sgl_free()
Use the sgl_alloc() and sgl_free() functions instead of open coding
these functions.
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Cc: Keith Busch
Cc: Christoph Hellwig
Cc: Sagi Grimberg
---
drivers/nvme/target/Kconfig | 1 +
drivers/nvme/target/rdma.c | 63 +++--
2 files changed, 5 insertions(+), 59 deletions(-)
diff --git a/drivers/nvme/target/Kconfig b/drivers/nvme/target/Kconfig
index 4d9715630e21..5f4f8b16685f 100644
--- a/drivers/nvme/target/Kconfig
+++ b/drivers/nvme/target/Kconfig
@@ -29,6 +29,7 @@ config NVME_TARGET_RDMA
tristate "NVMe over Fabrics RDMA target support"
depends on INFINIBAND
depends on NVME_TARGET
+ select SGL_ALLOC
help
This enables the NVMe RDMA target support, which allows exporting NVMe
devices over RDMA.
diff --git a/drivers/nvme/target/rdma.c b/drivers/nvme/target/rdma.c
index 76d2bb793afe..363d44c10b68 100644
--- a/drivers/nvme/target/rdma.c
+++ b/drivers/nvme/target/rdma.c
@@ -185,59 +185,6 @@ nvmet_rdma_put_rsp(struct nvmet_rdma_rsp *rsp)
spin_unlock_irqrestore(&rsp->queue->rsps_lock, flags);
}
-static void nvmet_rdma_free_sgl(struct scatterlist *sgl, unsigned int nents)
-{
- struct scatterlist *sg;
- int count;
-
- if (!sgl || !nents)
- return;
-
- for_each_sg(sgl, sg, nents, count)
- __free_page(sg_page(sg));
- kfree(sgl);
-}
-
-static int nvmet_rdma_alloc_sgl(struct scatterlist **sgl, unsigned int *nents,
- u32 length)
-{
- struct scatterlist *sg;
- struct page *page;
- unsigned int nent;
- int i = 0;
-
- nent = DIV_ROUND_UP(length, PAGE_SIZE);
- sg = kmalloc_array(nent, sizeof(struct scatterlist), GFP_KERNEL);
- if (!sg)
- goto out;
-
- sg_init_table(sg, nent);
-
- while (length) {
- u32 page_len = min_t(u32, length, PAGE_SIZE);
-
- page = alloc_page(GFP_KERNEL);
- if (!page)
- goto out_free_pages;
-
- sg_set_page(&sg[i], page, page_len, 0);
- length -= page_len;
- i++;
- }
- *sgl = sg;
- *nents = nent;
- return 0;
-
-out_free_pages:
- while (i > 0) {
- i--;
- __free_page(sg_page(&sg[i]));
- }
- kfree(sg);
-out:
- return NVME_SC_INTERNAL;
-}
-
static int nvmet_rdma_alloc_cmd(struct nvmet_rdma_device *ndev,
struct nvmet_rdma_cmd *c, bool admin)
{
@@ -484,7 +431,7 @@ static void nvmet_rdma_release_rsp(struct nvmet_rdma_rsp
*rsp)
}
if (rsp->req.sg != &rsp->cmd->inline_sg)
- nvmet_rdma_free_sgl(rsp->req.sg, rsp->req.sg_cnt);
+ sgl_free(rsp->req.sg);
if (unlikely(!list_empty_careful(&queue->rsp_wr_wait_list)))
nvmet_rdma_process_wr_wait_list(queue);
@@ -620,16 +567,14 @@ static u16 nvmet_rdma_map_sgl_keyed(struct nvmet_rdma_rsp
*rsp,
u32 len = get_unaligned_le24(sgl->length);
u32 key = get_unaligned_le32(sgl->key);
int ret;
- u16 status;
/* no data command? */
if (!len)
return 0;
- status = nvmet_rdma_alloc_sgl(&rsp->req.sg, &rsp->req.sg_cnt,
- len);
- if (status)
- return status;
+ rsp->req.sg = sgl_alloc(len, GFP_KERNEL, &rsp->req.sg_cnt);
+ if (!rsp->req.sg)
+ return NVME_SC_INTERNAL;
ret = rdma_rw_ctx_init(&rsp->rw, cm_id->qp, cm_id->port_num,
rsp->req.sg, rsp->req.sg_cnt, 0, addr, key,
--
2.14.2
Re: [PATCH] staging: ccree: fix boolreturn.cocci warning
On Mon, Oct 16, 2017 at 03:39:57PM +0530, suni...@techveda.org wrote:
> From: Suniel Mahesh
>
> This fixes the following coccinelle warning:
> WARNING: return of 0/1 in function 'ssi_is_hw_key' with return type bool.
Perhaps
Coccinelle emits WARNING: return of 0/1 in function 'ssi_is_hw_key' with return
type bool.
Return 'false' instead of 0.
> Signed-off-by: Suniel Mahesh
> ---
> Note:
> - Patch was tested and built(ARCH=arm) on latest
> linux-next.
> - No build issues reported, however it was not
> tested on real hardware.
> - Please discard this changeset, if this is not
> helping the code look better.
> ---
> drivers/staging/ccree/ssi_cipher.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/staging/ccree/ssi_cipher.h
> b/drivers/staging/ccree/ssi_cipher.h
> index c9a83df..f499962 100644
> --- a/drivers/staging/ccree/ssi_cipher.h
> +++ b/drivers/staging/ccree/ssi_cipher.h
> @@ -75,7 +75,7 @@ struct arm_hw_key_info {
>
> static inline bool ssi_is_hw_key(struct crypto_tfm *tfm)
> {
> - return 0;
> + return false;
> }
Hope this helps,
Tobin.
Re: [PATCH] staging: ccree: Fix bool comparison
On Mon, Oct 16, 2017 at 03:38:11PM +0530, suni...@techveda.org wrote: > From: Suniel Mahesh > > Bool tests don't need comparisons. This commit log could be a bit longer. You may like to read Documentation/process/submitting-patches.rst (section 2). > This fixes the following coccinelle warning: > WARNING: Comparison of bool to 0/1 > > Signed-off-by: Suniel Mahesh > --- > Note: > - Patch was tested and built(ARCH=arm) on latest > linux-next. Building is _not_ testing. > - No build issues reported, however it was not > tested on real hardware. This is implicit if you state 'builds on ARM' > - Please discard this changeset, if this is not > helping the code look better. This is implicit also ;) > --- > drivers/staging/ccree/ssi_request_mgr.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/staging/ccree/ssi_request_mgr.c > b/drivers/staging/ccree/ssi_request_mgr.c > index 2e0df57..942afe2 100644 > --- a/drivers/staging/ccree/ssi_request_mgr.c > +++ b/drivers/staging/ccree/ssi_request_mgr.c > @@ -272,7 +272,7 @@ int send_request( > unsigned int max_required_seq_len = (total_seq_len + > ((ssi_req->ivgen_dma_addr_len == 0) ? 0 > : > SSI_IVPOOL_SEQ_LEN) + > - ((is_dout == 0) ? 1 : 0)); > + (!is_dout ? 1 : 0)); Perhaps - ((is_dout == 0) ? 1 : 0)); + (is_dout ? 0 : 1) Good luck, Tobin.
Re: md5sum (from libkcapi) fails on kernel 4.9 but not on 4.13
Am Montag, 16. Oktober 2017, 08:53:00 CEST schrieb Christophe LEROY: Hi Christophe, > Hi Stephan, > > I get an issue with md5sum of a big file with kernel 4.9. It don't get > that issue with kernel 4.13. The key to the difference in libkcapi is the following code: static void _kcapi_handle_flags(struct kcapi_handle *handle) ... /* older interfaces only processed 16 pages in a row */ handle->flags.alg_max_pages = _kcapi_kernver_ge(handle, 4, 11, 0) ? UINT_MAX : ALG_MAX_PAGES; ... This check is mainly for skcipher but it affects all cipher types. Thus, on older kernels, only 16 pages are injected into the kernel with splice and then it is reverted to sendmsg. Again, this logic is not so much of an issue for algif_hash, but rather for algif_skcipher and algif_aead. > > When I do an strace, I see a difference in the calls: at the end of the > file, with 4.9 md5sum uses sendmsg() for the last block while with 4.13 > it uses splice() as for all the previous blocks. > > The problem is that the last block has a size over 32kbytes, which is > the maximum size the talitos driver accepts to hash at once. > It looks like sendmsg() sends the entire block to the crypto driver > while splice() calls the crypto driver with blocks of page size. I understand that your driver has this 32kb limit. But I am not sure what the difference between the handling of the sendmsg and the splice invocation is. According to the strace, the splice accepted 256kb of data. Looking into the hash_sendpage function implementing the backend of this splice system call, it simply performs an update invocation with this buffer size. I.e. it invokes your driver with 256kb of data. The sendmsg first performs a copy_from_user with the maximum limit of 16 * PAGE_SIZE (see the limit variable in hash_sendmsg) and then invokes the update function. So, I am not sure why the sendmsg call chokes where the sendpage call succeeds. If you tamper with the code shown above from libkcapi and set alg_max_pages to a low value, the library reverts to sendmsg after the given number of pages. Ciao Stephan
Re: [PATCH 0/2] crypto: s5p-sss: Add HASH support for Exynos
Hi all, sorry for error in subject line, it should be [PATCH v6 0/2] and so on, I can resend this if needed ? On 16.10.2017 19:43, Kamil Konieczny wrote: > First patch cleans up spaces in defines, second adds HASH support for Exynos. > Changes: -- Best regards, Kamil Konieczny Samsung R&D Institute Poland
[PATCH 2/2] Add HASH support for Exynos
Add support for MD5, SHA1, SHA256 hash algorithms for Exynos HW. It uses the crypto framework asynchronous hash api. It is based on omap-sham.c driver. S5P has some HW differencies and is not implemented. Modifications in s5p-sss: - Add hash supporting structures and functions. - Modify irq handler to handle both aes and hash signals. - Resize resource end in probe if EXYNOS_HASH is enabled in Kconfig. - Add new copyright line and new author. - Tested on Odroid-U3 with Exynos 4412 CPU, kernel 4.13-rc6 with crypto run-time self test testmgr and with tcrypt module with: modprobe tcrypt sec=1 mode=N where N=402, 403, 404 (MD5, SHA1, SHA256). Modifications in drivers/crypto/Kconfig: - Add new CRYPTO_DEV_EXYNOS_HASH, depend on !EXYNOS_RNG and CRYPTO_DEV_S5P - Select sw algorithms MD5, SHA1 and SHA256 in EXYNOS_HASH as they are nedded for fallback. Signed-off-by: Kamil Konieczny --- drivers/crypto/Kconfig | 14 + drivers/crypto/s5p-sss.c | 1406 +- 2 files changed, 1410 insertions(+), 10 deletions(-) diff --git a/drivers/crypto/Kconfig b/drivers/crypto/Kconfig index 4b75084fabad..dea4d33d9c7f 100644 --- a/drivers/crypto/Kconfig +++ b/drivers/crypto/Kconfig @@ -439,6 +439,20 @@ config CRYPTO_DEV_S5P Select this to offload Samsung S5PV210 or S5PC110, Exynos from AES algorithms execution. +config CRYPTO_DEV_EXYNOS_HASH + bool "Support for Samsung Exynos HASH accelerator" + depends on CRYPTO_DEV_S5P + depends on !CRYPTO_DEV_EXYNOS_RNG && CRYPTO_DEV_EXYNOS_RNG!=m + select CRYPTO_SHA1 + select CRYPTO_MD5 + select CRYPTO_SHA256 + help + Select this to offload Exynos from HASH MD5/SHA1/SHA256. + This will select software SHA1, MD5 and SHA256 as they are + needed for small and zero-size messages. + HASH algorithms will be disabled if EXYNOS_RNG + is enabled due to hw conflict. + config CRYPTO_DEV_NX bool "Support for IBM PowerPC Nest (NX) cryptographic acceleration" depends on PPC64 diff --git a/drivers/crypto/s5p-sss.c b/drivers/crypto/s5p-sss.c index dfae1865c384..b5a7e49f9285 100644 --- a/drivers/crypto/s5p-sss.c +++ b/drivers/crypto/s5p-sss.c @@ -1,18 +1,21 @@ /* * Cryptographic API. * - * Support for Samsung S5PV210 HW acceleration. + * Support for Samsung S5PV210 and Exynos HW acceleration. * * Copyright (C) 2011 NetUP Inc. All rights reserved. + * Copyright (c) 2017 Samsung Electronics Co., Ltd. All rights reserved. * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as published * by the Free Software Foundation. * + * Hash part based on omap-sham.c driver. */ #include #include +#include #include #include #include @@ -30,28 +33,41 @@ #include #include +#include +#include +#include +#include + #define _SBF(s, v) ((v) << (s)) /* Feed control registers */ #define SSS_REG_FCINTSTAT 0x +#define SSS_FCINTSTAT_HPARTINT BIT(7) +#define SSS_FCINTSTAT_HDONEINT BIT(5) #define SSS_FCINTSTAT_BRDMAINT BIT(3) #define SSS_FCINTSTAT_BTDMAINT BIT(2) #define SSS_FCINTSTAT_HRDMAINT BIT(1) #define SSS_FCINTSTAT_PKDMAINT BIT(0) #define SSS_REG_FCINTENSET 0x0004 +#define SSS_FCINTENSET_HPARTINTENSET BIT(7) +#define SSS_FCINTENSET_HDONEINTENSET BIT(5) #define SSS_FCINTENSET_BRDMAINTENSET BIT(3) #define SSS_FCINTENSET_BTDMAINTENSET BIT(2) #define SSS_FCINTENSET_HRDMAINTENSET BIT(1) #define SSS_FCINTENSET_PKDMAINTENSET BIT(0) #define SSS_REG_FCINTENCLR 0x0008 +#define SSS_FCINTENCLR_HPARTINTENCLR BIT(7) +#define SSS_FCINTENCLR_HDONEINTENCLR BIT(5) #define SSS_FCINTENCLR_BRDMAINTENCLR BIT(3) #define SSS_FCINTENCLR_BTDMAINTENCLR BIT(2) #define SSS_FCINTENCLR_HRDMAINTENCLR BIT(1) #define SSS_FCINTENCLR_PKDMAINTENCLR BIT(0) #define SSS_REG_FCINTPEND 0x000C +#define SSS_FCINTPEND_HPARTINTPBIT(7) +#define SSS_FCINTPEND_HDONEINTPBIT(5) #define SSS_FCINTPEND_BRDMAINTPBIT(3) #define SSS_FCINTPEND_BTDMAINTPBIT(2) #define SSS_FCINTPEND_HRDMAINTPBIT(1) @@ -72,6 +88,7 @@ #define SSS_HASHIN_INDEPENDENT _SBF(0, 0x00) #define SSS_HASHIN_CIPHER_INPUT_SBF(0, 0x01) #define SSS_HASHIN_CIPHER_OUTPUT _SBF(0, 0x02) +#define SSS_HASHIN_MASK_SBF(0, 0x03) #define SSS_REG_FCBRDMAS 0x0020 #define SSS_REG_FCBRDMAL 0x0024 @@ -146,9 +163,80 @@ #define AES_KEY_LEN16 #define CRYPTO_QUEUE_LEN 1 +/* HASH registers */ +#define SSS_REG_HASH_CTRL 0x00 + +#define SSS_HASH_USER_IV_ENBIT(5) +#define SSS_HASH_INIT_BIT BIT(4) +#define SSS_HASH_ENGINE_
[PATCH 1/2] change spaces into tabs in defines
Change spaces into tabs in defines. Signed-off-by: Kamil Konieczny --- drivers/crypto/s5p-sss.c | 190 +++ 1 file changed, 95 insertions(+), 95 deletions(-) diff --git a/drivers/crypto/s5p-sss.c b/drivers/crypto/s5p-sss.c index 7ac657f46d15..dfae1865c384 100644 --- a/drivers/crypto/s5p-sss.c +++ b/drivers/crypto/s5p-sss.c @@ -30,98 +30,98 @@ #include #include -#define _SBF(s, v) ((v) << (s)) +#define _SBF(s, v) ((v) << (s)) /* Feed control registers */ -#define SSS_REG_FCINTSTAT 0x -#define SSS_FCINTSTAT_BRDMAINT BIT(3) -#define SSS_FCINTSTAT_BTDMAINT BIT(2) -#define SSS_FCINTSTAT_HRDMAINT BIT(1) -#define SSS_FCINTSTAT_PKDMAINT BIT(0) - -#define SSS_REG_FCINTENSET 0x0004 -#define SSS_FCINTENSET_BRDMAINTENSETBIT(3) -#define SSS_FCINTENSET_BTDMAINTENSETBIT(2) -#define SSS_FCINTENSET_HRDMAINTENSETBIT(1) -#define SSS_FCINTENSET_PKDMAINTENSETBIT(0) - -#define SSS_REG_FCINTENCLR 0x0008 -#define SSS_FCINTENCLR_BRDMAINTENCLRBIT(3) -#define SSS_FCINTENCLR_BTDMAINTENCLRBIT(2) -#define SSS_FCINTENCLR_HRDMAINTENCLRBIT(1) -#define SSS_FCINTENCLR_PKDMAINTENCLRBIT(0) - -#define SSS_REG_FCINTPEND 0x000C -#define SSS_FCINTPEND_BRDMAINTP BIT(3) -#define SSS_FCINTPEND_BTDMAINTP BIT(2) -#define SSS_FCINTPEND_HRDMAINTP BIT(1) -#define SSS_FCINTPEND_PKDMAINTP BIT(0) - -#define SSS_REG_FCFIFOSTAT 0x0010 -#define SSS_FCFIFOSTAT_BRFIFOFULBIT(7) -#define SSS_FCFIFOSTAT_BRFIFOEMPBIT(6) -#define SSS_FCFIFOSTAT_BTFIFOFULBIT(5) -#define SSS_FCFIFOSTAT_BTFIFOEMPBIT(4) -#define SSS_FCFIFOSTAT_HRFIFOFULBIT(3) -#define SSS_FCFIFOSTAT_HRFIFOEMPBIT(2) -#define SSS_FCFIFOSTAT_PKFIFOFULBIT(1) -#define SSS_FCFIFOSTAT_PKFIFOEMPBIT(0) - -#define SSS_REG_FCFIFOCTRL 0x0014 -#define SSS_FCFIFOCTRL_DESSEL BIT(2) -#define SSS_HASHIN_INDEPENDENT _SBF(0, 0x00) -#define SSS_HASHIN_CIPHER_INPUT _SBF(0, 0x01) -#define SSS_HASHIN_CIPHER_OUTPUT_SBF(0, 0x02) - -#define SSS_REG_FCBRDMAS0x0020 -#define SSS_REG_FCBRDMAL0x0024 -#define SSS_REG_FCBRDMAC0x0028 -#define SSS_FCBRDMAC_BYTESWAP BIT(1) -#define SSS_FCBRDMAC_FLUSH BIT(0) - -#define SSS_REG_FCBTDMAS0x0030 -#define SSS_REG_FCBTDMAL0x0034 -#define SSS_REG_FCBTDMAC0x0038 -#define SSS_FCBTDMAC_BYTESWAP BIT(1) -#define SSS_FCBTDMAC_FLUSH BIT(0) - -#define SSS_REG_FCHRDMAS0x0040 -#define SSS_REG_FCHRDMAL0x0044 -#define SSS_REG_FCHRDMAC0x0048 -#define SSS_FCHRDMAC_BYTESWAP BIT(1) -#define SSS_FCHRDMAC_FLUSH BIT(0) - -#define SSS_REG_FCPKDMAS0x0050 -#define SSS_REG_FCPKDMAL0x0054 -#define SSS_REG_FCPKDMAC0x0058 -#define SSS_FCPKDMAC_BYTESWAP BIT(3) -#define SSS_FCPKDMAC_DESCENDBIT(2) -#define SSS_FCPKDMAC_TRANSMIT BIT(1) -#define SSS_FCPKDMAC_FLUSH BIT(0) - -#define SSS_REG_FCPKDMAO0x005C +#define SSS_REG_FCINTSTAT 0x +#define SSS_FCINTSTAT_BRDMAINT BIT(3) +#define SSS_FCINTSTAT_BTDMAINT BIT(2) +#define SSS_FCINTSTAT_HRDMAINT BIT(1) +#define SSS_FCINTSTAT_PKDMAINT BIT(0) + +#define SSS_REG_FCINTENSET 0x0004 +#define SSS_FCINTENSET_BRDMAINTENSET BIT(3) +#define SSS_FCINTENSET_BTDMAINTENSET BIT(2) +#define SSS_FCINTENSET_HRDMAINTENSET BIT(1) +#define SSS_FCINTENSET_PKDMAINTENSET BIT(0) + +#define SSS_REG_FCINTENCLR 0x0008 +#define SSS_FCINTENCLR_BRDMAINTENCLR BIT(3) +#define SSS_FCINTENCLR_BTDMAINTENCLR BIT(2) +#define SSS_FCINTENCLR_HRDMAINTENCLR BIT(1) +#define SSS_FCINTENCLR_PKDMAINTENCLR BIT(0) + +#define SSS_REG_FCINTPEND 0x000C +#define SSS_FCINTPEND_BRDMAINTPBIT(3) +#define SSS_FCINTPEND_BTDMAINTPBIT(2) +#define SSS_FCINTPEND_HRDMAINTPBIT(1) +#define SSS_FCINTPEND_PKDMAINTPBIT(0) + +#define SSS_REG_FCFIFOSTAT 0x0010 +#define SSS_FCFIFOSTAT_BRFIFOFUL BIT(7) +#define SSS_FCFIFOSTAT_BRFIFOEMP BIT(6) +#define SSS_FCFIFOSTAT_BTFIFOFUL BIT(5) +#define SSS_FCFIFOSTAT_BTFIFOEMP BIT(4) +#define SSS_FCFIFOSTAT_HRFIFOFUL BIT(3) +#define SSS_FCFIFOSTAT_HRFIFOEMP BIT(2) +#define SSS_FCFIFOSTAT_PKFIFOFUL BIT(1) +#define SSS_FCFIFOSTAT_PKFIFOEMP BIT(0) + +#define SSS_REG_FCFIFOCTRL 0x0014 +#define SSS_FCFIFOCTRL_DESSEL BIT(2) +#define SSS_HASHIN_INDEPENDENT _SBF(0, 0x00) +#define SSS_HASHIN_CIPHER_INPUT_SBF(0, 0x01) +#define SSS_HASHIN_CIPHER_OUTPUT _SBF(0, 0x02
[PATCH 0/2] crypto: s5p-sss: Add HASH support for Exynos
First patch cleans up spaces in defines, second adds HASH support for Exynos. Changes: version 6: - fixes suggested by Vladimir Zapolskiy: change HASH_OP enum into bool, fix comments, change int into unsigned int in several functions, change some functions to return void, remove unnecessary parentheses in s5p_hash_import, replace rctx with ctx for request context, drop some dd vars and use tctx->dd instead, simplify s5p_hash_rx, s5p_hash_copy_result and s5p_hash_set_flow, change int final into bool final, reoder some declarations, split patch into two - rewrite and fix while loop in s5p_hash_copy_sg_lists - rewrite while loop in s5p_hash_prepare_sgs version 5: - fix suggested by Krzysztof Kozlowski: change defines HASH_OP into enum, fix comments version 4: - fixes suggested by Krzysztof Kozlowski: reformat comments, convert context flags into two bool vars, drop SSS_ALIGNED, change name of SSS_DMA_ALIGN and SSS_DMA_ALIGN_MASK, split assignments into separate lines, use IS_ENABLED in place of ifdef, remove sss_hash_algs_info and simplify register and deregister HASH algs version 3: - many fixes suggested by Krzysztof Kozlowski: comments, uppercases in const, remove unused defines, remove unused variable bs, constify aes_variant, remove global var use_hash, remove WARN_ON, improve hash_import(), change goto label into 'out' in s5p_hash_handle_queue(), reorder variable declarations, add spinlock to protect clearing HASH_FLAGS_BUSY - simplify code: replace one-line functions s5p_hash_update_req(), s5p_hash_final_req() with call to s5p_hash_xmit_dma(), and delete them - replace call to s5p_hash_hw_init() into s5p_ahash_dma_init() and delete it - fix clearing shash flag CRYPTO_TFM_REQ_MAY_SLEEP - fix s5p_hash_set_flow() version 2: - change patch format so number of lines drops - change in Kconfig as suggested by Krzysztof Kozlowski, add EXYNOS_HASH subsection - change #ifndef EXYNOS_RNG into #ifdef CRYPTO_DEV_EXYNOS_HASH - remove style fixups in aes, as they should go in separate patch - remove FLOW_LOG, FLOW_DUMP macros and its uses - remove #if 0 ... endif - remove unused function hash_wait and its defines - fix compiler warning in dev_dbg - remove some comments - other minor fixes in comments Kamil Konieczny (2): change spaces into tabs in defines Add HASH support for Exynos drivers/crypto/Kconfig | 14 + drivers/crypto/s5p-sss.c | 1596 +++--- 2 files changed, 1505 insertions(+), 105 deletions(-) -- 2.14.1.536.g6867272d5b56
Re: [PATCH v5] crypto: s5p-sss: Add HASH support for Exynos
Hello Vladimir,
>>[snip]
>> -static struct s5p_aes_dev *s5p_dev;
>> +enum hash_op {
>> +HASH_OP_UPDATE = 1,
>> +HASH_OP_FINAL = 2
>> +};
>
> If this type is not going to be extended, then I'd rather suggest to
> use a boolean correspondent field in the struct s5p_hash_reqctx instead
> of a new introduced type. [...]
In prevoious mail I wrote 'op_final', but this will actually be 'op_update'.
--
Best regards,
Kamil Konieczny
Samsung R&D Institute Poland
Re: [PATCH v5] crypto: s5p-sss: Add HASH support for Exynos
Hello Vladimir,
thank you for review. I will apply most of your suggestions,
in a few places it is not possible, see comments below.
Your review helped me to improve two functions.
On 12.10.2017 13:45, Vladimir Zapolskiy wrote:
> Hello Kamil,
>
> thank you for the change, please find below a number of minor
> review comments.
>
> On 10/09/2017 02:12 PM, Kamil Konieczny wrote:
>> Add support for MD5, SHA1, SHA256 hash algorithms for Exynos HW.
>> It uses the crypto framework asynchronous hash api.
>> It is based on omap-sham.c driver.
>> S5P has some HW differencies and is not implemented.
>>
>
> [snip]
>
>>
>> /* Feed control registers */
>> #define SSS_REG_FCINTSTAT 0x
>> +#define SSS_FCINTSTAT_HPARTINT BIT(7)
>> +#define SSS_FCINTSTAT_HDONEINT BIT(5)
>
> ^
> Please use the same style of whitespaces as it is found around.
>
> Generally I do agree that the tabs are better here, please feel free
> to send a preceding change, which changes the spacing to tabs, otherwise
> use space symbols in this change.
ok, I will break this patch into two, first one change spaces into tabs.
>> #define SSS_FCINTSTAT_BRDMAINT BIT(3)
>> #define SSS_FCINTSTAT_BTDMAINT BIT(2)
>> #define SSS_FCINTSTAT_HRDMAINT BIT(1)
>> #define SSS_FCINTSTAT_PKDMAINT BIT(0)
>>
>> #define SSS_REG_FCINTENSET 0x0004
>> +#define SSS_FCINTENSET_HPARTINTENSETBIT(7)
>> +#define SSS_FCINTENSET_HDONEINTENSETBIT(5)
>
> Same as above.
ok with this one and following
>> [skip]
> [snip]
>
>> struct s5p_aes_reqctx {
>> @@ -195,6 +284,19 @@ struct s5p_aes_ctx {
>> * protects against concurrent access to these fields.
>> * @lock: Lock for protecting both access to device hardware registers
>> * and fields related to current request (including the busy
>> field).
>> + * @res:Resources for hash.
>> + * @io_hash_base: Per-variant offset for HASH block IO memory.
>> + * @hash_lock: Lock for protecting hash_req, hash_queue and hash_flags
>> + * variable.
>> + * @hash_tasklet: New HASH request scheduling job.
>> + * @xmit_buf: Buffer for current HASH request transfer into SSS block.
>> + * @hash_flags: Flags for current HASH op.
>> + * @hash_queue: Async hash queue.
>> + * @hash_req: Current request sending to SSS HASH block.
>> + * @hash_sg_iter: Scatterlist transferred through DMA into SSS HASH block.
>> + * @hash_sg_cnt: Counter for hash_sg_iter.
>> + *
>> + * @use_hash: true if HASH algs enabled
>
> You may want to reorder the lines to get the same order as in the struct.
ok
>> */
>> struct s5p_aes_dev {
>> struct device *dev;
>> @@ -215,16 +317,88 @@ struct s5p_aes_dev {
>> struct crypto_queue queue;
>> boolbusy;
>> spinlock_t lock;
>> +
>> +struct resource *res;
>> +void __iomem*io_hash_base;
>> +
>> +spinlock_t hash_lock; /* protect hash_ vars */
>> +unsigned long hash_flags;
>> +struct crypto_queue hash_queue;
>> +struct tasklet_struct hash_tasklet;
>> +
>> +u8 xmit_buf[BUFLEN];
>> +struct ahash_request*hash_req;
>> +struct scatterlist *hash_sg_iter;
>> +int hash_sg_cnt;
>
> Please change the type to 'unsigned int'.
>
>> +
>> +booluse_hash;
>> };
>>
>> -static struct s5p_aes_dev *s5p_dev;
>> +enum hash_op {
>> +HASH_OP_UPDATE = 1,
>> +HASH_OP_FINAL = 2
>> +};
>
> If this type is not going to be extended, then I'd rather suggest to
> use a boolean correspondent field in the struct s5p_hash_reqctx instead
> of a new introduced type.
>
I will change this into 'bool op_final'.
>> +
>> +/**
>> + * struct s5p_hash_reqctx - HASH request context
>> + * @dev:Associated device
>> + * @op: Current request operation (OP_UPDATE or OP_FINAL)
>
> See a comment above.
>
>> + * @digcnt: Number of bytes processed by HW (without buffer[] ones)
>> + * @digest: Digest message or IV for partial result
>> + * @bufcnt: Number of bytes holded in buffer[]
>> + * @nregs: Number of HW registers for digest or IV read/write
>> + * @engine: Bits for selecting type of HASH in SSS block
>> + * @sg: sg for DMA transfer
>> + * @sg_len: Length of sg for DMA transfer
>> + * @sgl[]: sg for joining buffer and req->src scatterlist
>> + * @skip: Skip offset in req->src for current op
>> + * @total: Total number of bytes for current request
>> + * @finup: Keep state for finup or final.
>> + * @error: Keep track of error.
>> + * @buffer[]: For byte(s) from end of req->src in UPDATE op
>> + */
>> +struct s5p_hash_reqctx {
>> +struct s5p_aes_de
[PATCH 1/2] x86/crypto/sha256-mb: fix panic due to unaligned access
struct sha256_ctx_mgr allocated in sha256_mb_mod_init() via kzalloc()
and later passed in sha256_mb_flusher_mgr_flush_avx2() function where
instructions vmovdqa used to access the struct. vmovdqa requires
16-bytes aligned argument, but nothing guarantees that struct
sha256_ctx_mgr will have that alignment. Unaligned vmovdqa will
generate GP fault.
Fix this by replacing vmovdqa with vmovdqu which doesn't have alignment
requirements.
Fixes: a377c6b1876e ("crypto: sha256-mb - submit/flush routines for AVX2")
Reported-by: Josh Poimboeuf
Signed-off-by: Andrey Ryabinin
Cc:
---
arch/x86/crypto/sha256-mb/sha256_mb_mgr_flush_avx2.S | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/arch/x86/crypto/sha256-mb/sha256_mb_mgr_flush_avx2.S
b/arch/x86/crypto/sha256-mb/sha256_mb_mgr_flush_avx2.S
index 8fe6338bcc84..16c4ccb1f154 100644
--- a/arch/x86/crypto/sha256-mb/sha256_mb_mgr_flush_avx2.S
+++ b/arch/x86/crypto/sha256-mb/sha256_mb_mgr_flush_avx2.S
@@ -155,8 +155,8 @@ LABEL skip_ %I
.endr
# Find min length
- vmovdqa _lens+0*16(state), %xmm0
- vmovdqa _lens+1*16(state), %xmm1
+ vmovdqu _lens+0*16(state), %xmm0
+ vmovdqu _lens+1*16(state), %xmm1
vpminud %xmm1, %xmm0, %xmm2 # xmm2 has {D,C,B,A}
vpalignr $8, %xmm2, %xmm3, %xmm3# xmm3 has {x,x,D,C}
@@ -176,8 +176,8 @@ LABEL skip_ %I
vpsubd %xmm2, %xmm0, %xmm0
vpsubd %xmm2, %xmm1, %xmm1
- vmovdqa %xmm0, _lens+0*16(state)
- vmovdqa %xmm1, _lens+1*16(state)
+ vmovdqu %xmm0, _lens+0*16(state)
+ vmovdqu %xmm1, _lens+1*16(state)
# "state" and "args" are the same address, arg1
# len is arg2
@@ -234,8 +234,8 @@ ENTRY(sha256_mb_mgr_get_comp_job_avx2)
jc .return_null
# Find min length
- vmovdqa _lens(state), %xmm0
- vmovdqa _lens+1*16(state), %xmm1
+ vmovdqu _lens(state), %xmm0
+ vmovdqu _lens+1*16(state), %xmm1
vpminud %xmm1, %xmm0, %xmm2 # xmm2 has {D,C,B,A}
vpalignr $8, %xmm2, %xmm3, %xmm3# xmm3 has {x,x,D,C}
--
2.13.6
[PATCH 2/2] x86/crypto/sha1-mb: fix panic due to unaligned access
struct sha1_ctx_mgr allocated in sha1_mb_mod_init() via kzalloc()
and later passed in sha1_mb_flusher_mgr_flush_avx2() function where
instructions vmovdqa used to access the struct. vmovdqa requires
16-bytes aligned argument, but nothing guarantees that struct
sha1_ctx_mgr will have that alignment. Unaligned vmovdqa will
generate GP fault.
Fix this by replacing vmovdqa with vmovdqu which doesn't have alignment
requirements.
Fixes: 2249cbb53ead ("crypto: sha-mb - SHA1 multibuffer submit and flush
routines for AVX2")
Signed-off-by: Andrey Ryabinin
Cc:
---
arch/x86/crypto/sha1-mb/sha1_mb_mgr_flush_avx2.S | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/arch/x86/crypto/sha1-mb/sha1_mb_mgr_flush_avx2.S
b/arch/x86/crypto/sha1-mb/sha1_mb_mgr_flush_avx2.S
index 93b945597ecf..7cfba738f104 100644
--- a/arch/x86/crypto/sha1-mb/sha1_mb_mgr_flush_avx2.S
+++ b/arch/x86/crypto/sha1-mb/sha1_mb_mgr_flush_avx2.S
@@ -157,8 +157,8 @@ LABEL skip_ %I
.endr
# Find min length
- vmovdqa _lens+0*16(state), %xmm0
- vmovdqa _lens+1*16(state), %xmm1
+ vmovdqu _lens+0*16(state), %xmm0
+ vmovdqu _lens+1*16(state), %xmm1
vpminud %xmm1, %xmm0, %xmm2 # xmm2 has {D,C,B,A}
vpalignr $8, %xmm2, %xmm3, %xmm3 # xmm3 has {x,x,D,C}
@@ -178,8 +178,8 @@ LABEL skip_ %I
vpsubd %xmm2, %xmm0, %xmm0
vpsubd %xmm2, %xmm1, %xmm1
- vmovdqa %xmm0, _lens+0*16(state)
- vmovdqa %xmm1, _lens+1*16(state)
+ vmovdqu %xmm0, _lens+0*16(state)
+ vmovdqu %xmm1, _lens+1*16(state)
# "state" and "args" are the same address, arg1
# len is arg2
@@ -235,8 +235,8 @@ ENTRY(sha1_mb_mgr_get_comp_job_avx2)
jc .return_null
# Find min length
- vmovdqa _lens(state), %xmm0
- vmovdqa _lens+1*16(state), %xmm1
+ vmovdqu _lens(state), %xmm0
+ vmovdqu _lens+1*16(state), %xmm1
vpminud %xmm1, %xmm0, %xmm2# xmm2 has {D,C,B,A}
vpalignr $8, %xmm2, %xmm3, %xmm3 # xmm3 has {x,x,D,C}
--
2.13.6
Re: Fostering linux community collaboration on hardware accelerators
> > So as ever with a linux community focusing on a particular topic, the > > obvious solution is a mailing list. There are a number of options on how > > do this. > > > > 1) Ask one of the industry bodies to host? Who? > > > > 2) Put together a compelling argument for linux-accelerat...@vger.kernel.org > > as probably the most generic location for such a list. > > Happy to offer linux-accelerat...@lists.ozlabs.org, which I can get set > up immediately (and if we want patchwork, patchwork.ozlabs.org is > available as always, no matter where the list is hosted). > That would be great! Thanks for doing this. Much easier to find out what such a list is useful for by the practical option of having a list and see what people do with it. Thanks, Jonathan > > More open questions are > > 1) Scope? > > * Would anyone ever use such an overarching list? > > * Are we better off with the usual adhoc list of 'interested parties' + > > lkml? > > * Do we actually need to define the full scope - are we better with a > > vague > > definition? > > I think a list with a broad and vaguely defined scope is a good idea - > it would certainly be helpful to us to be able to follow what other > contributors are doing that could be relevant to our CAPI and OpenCAPI work. > > > > > 2) Is there an existing community we can use to discuss these issues? > > (beyond the obvious firehose of LKML). > > > > 3) Who else to approach for input on these general questions? > > > > In parallel to this there are elements such as git / patchwork etc but > > they can all be done as they are needed. > > > > Thanks > > > > -- > > Jonathan Cameron > > Huawei > > >
[PATCH] staging: ccree: fix boolreturn.cocci warning
From: Suniel Mahesh
This fixes the following coccinelle warning:
WARNING: return of 0/1 in function 'ssi_is_hw_key' with return type bool.
Signed-off-by: Suniel Mahesh
---
Note:
- Patch was tested and built(ARCH=arm) on latest
linux-next.
- No build issues reported, however it was not
tested on real hardware.
- Please discard this changeset, if this is not
helping the code look better.
---
drivers/staging/ccree/ssi_cipher.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/staging/ccree/ssi_cipher.h
b/drivers/staging/ccree/ssi_cipher.h
index c9a83df..f499962 100644
--- a/drivers/staging/ccree/ssi_cipher.h
+++ b/drivers/staging/ccree/ssi_cipher.h
@@ -75,7 +75,7 @@ struct arm_hw_key_info {
static inline bool ssi_is_hw_key(struct crypto_tfm *tfm)
{
- return 0;
+ return false;
}
#endif /* CRYPTO_TFM_REQ_HW_KEY */
--
1.9.1
[PATCH] staging: ccree: Fix bool comparison
From: Suniel Mahesh Bool tests don't need comparisons. This fixes the following coccinelle warning: WARNING: Comparison of bool to 0/1 Signed-off-by: Suniel Mahesh --- Note: - Patch was tested and built(ARCH=arm) on latest linux-next. - No build issues reported, however it was not tested on real hardware. - Please discard this changeset, if this is not helping the code look better. --- drivers/staging/ccree/ssi_request_mgr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/staging/ccree/ssi_request_mgr.c b/drivers/staging/ccree/ssi_request_mgr.c index 2e0df57..942afe2 100644 --- a/drivers/staging/ccree/ssi_request_mgr.c +++ b/drivers/staging/ccree/ssi_request_mgr.c @@ -272,7 +272,7 @@ int send_request( unsigned int max_required_seq_len = (total_seq_len + ((ssi_req->ivgen_dma_addr_len == 0) ? 0 : SSI_IVPOOL_SEQ_LEN) + - ((is_dout == 0) ? 1 : 0)); + (!is_dout ? 1 : 0)); #if defined(CONFIG_PM_RUNTIME) || defined(CONFIG_PM_SLEEP) rc = ssi_power_mgr_runtime_get(dev); -- 1.9.1
