| Message ID | 20210505202618.2663889-1-clabbe@baylibre.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | crypto: start to fix ixp4xx | expand |
On Wed, May 05, 2021 at 08:26:07PM +0000, Corentin Labbe wrote: > Hello > > Loading the ixp4xx crypto driver exhibits lots of error. > All algorithm fail selftests with different reasons. > This series start to fixes some of thoses problem. > > Corentin Labbe (11): > crypto: ixp4xx: dma_unmap the correct address > crypto: ixp4xx: update IV after requests > crypto: ixp4xx: fallback when having more than one SG > crypto: ixp4xx: convert unsigned to unsigned int > crypto: ixp4xx: convert all printk to dev_xxx > crypto: ixp4xx: whitespace fixes > crypto: ixp4xx: Do not initialize static to NULL > crypto: ixp4xx: remove brackets from single statement > crypto: ixp4xx: Correct functions alignment > MAINTAINERS: add ixp4xx_crypto to the right arch list > MAINTAINERS: add myself as maintainer of ixp4xx_crypto > > MAINTAINERS | 7 + > drivers/crypto/Kconfig | 5 + > drivers/crypto/ixp4xx_crypto.c | 277 +++++++++++++++++++++------------ > 3 files changed, 188 insertions(+), 101 deletions(-) All applied. Thanks.
On Wed, May 5, 2021 at 10:26 PM Corentin Labbe <clabbe@baylibre.com> wrote: > Loading the ixp4xx crypto driver exhibits lots of error. > All algorithm fail selftests with different reasons. > This series start to fixes some of thoses problem. Excellent! Thanks for taking over this Corentin!! FWIW: Acked-by: Linus Walleij <linus.walleij@linaro.org> If I merge my 3 additional patches through ARM SoC will it work out or do I need to think about some clever merging strategy? Yours, Linus Walleij
On Tue, May 18, 2021 at 01:53:56AM +0200, Linus Walleij wrote: > On Wed, May 5, 2021 at 10:26 PM Corentin Labbe <clabbe@baylibre.com> wrote: > > > Loading the ixp4xx crypto driver exhibits lots of error. > > All algorithm fail selftests with different reasons. > > This series start to fixes some of thoses problem. > > Excellent! Thanks for taking over this Corentin!! > FWIW: > Acked-by: Linus Walleij <linus.walleij@linaro.org> > > If I merge my 3 additional patches through ARM SoC > will it work out or do I need to think about some clever > merging strategy? Well if your patches don't touch the drivers/crypto then there shouldn't be any conflicts. Thanks,
On Tue, May 18, 2021 at 3:19 AM Herbert Xu <herbert@gondor.apana.org.au> wrote: > On Tue, May 18, 2021 at 01:53:56AM +0200, Linus Walleij wrote: > > On Wed, May 5, 2021 at 10:26 PM Corentin Labbe <clabbe@baylibre.com> wrote: > > > > > Loading the ixp4xx crypto driver exhibits lots of error. > > > All algorithm fail selftests with different reasons. > > > This series start to fixes some of thoses problem. > > > > Excellent! Thanks for taking over this Corentin!! > > FWIW: > > Acked-by: Linus Walleij <linus.walleij@linaro.org> > > > > If I merge my 3 additional patches through ARM SoC > > will it work out or do I need to think about some clever > > merging strategy? > > Well if your patches don't touch the drivers/crypto then there > shouldn't be any conflicts. Sadly they do, it's these: https://lore.kernel.org/linux-crypto/20210510213634.600866-1-linus.walleij@linaro.org/ https://lore.kernel.org/linux-crypto/20210510213634.600866-2-linus.walleij@linaro.org/ https://lore.kernel.org/linux-crypto/20210510213634.600866-3-linus.walleij@linaro.org/ I guess we could turn it around: I see if I can apply these two on top of the crypto tree and resend then we hope there is no collision in the ARM SoC tree instead. I need to respin the series anyway so shall we try this? Yours, Linus Walleij
On Tue, May 18, 2021 at 11:36:01PM +0200, Linus Walleij wrote: > > Sadly they do, it's these: > https://lore.kernel.org/linux-crypto/20210510213634.600866-1-linus.walleij@linaro.org/ > https://lore.kernel.org/linux-crypto/20210510213634.600866-2-linus.walleij@linaro.org/ > https://lore.kernel.org/linux-crypto/20210510213634.600866-3-linus.walleij@linaro.org/ > > I guess we could turn it around: I see if I can apply these two > on top of the crypto tree and resend then we hope there is no > collision in the ARM SoC tree instead. > > I need to respin the series anyway so shall we try this? Yes I think those changes are probably best routed through crypto as the code changes are quite extensive so the potential for non- trivial conflicts is high. Thanks,