* [Intel-gfx] ✓ Fi.CI.BAT: success for series starting with [1/2] drm/i915: Add a few asserts around handling of i915_request_is_active() (rev6)
From: Patchwork @ 2020-05-29 17:46 UTC (permalink / raw)
To: Chris Wilson; +Cc: intel-gfx
In-Reply-To: <20200529085809.23691-1-chris@chris-wilson.co.uk>
== Series Details ==
Series: series starting with [1/2] drm/i915: Add a few asserts around handling of i915_request_is_active() (rev6)
URL : https://patchwork.freedesktop.org/series/77781/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_8554 -> Patchwork_17820
====================================================
Summary
-------
**SUCCESS**
No regressions found.
External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17820/index.html
Changes
-------
No changes found
Participating hosts (49 -> 42)
------------------------------
Missing (7): fi-ilk-m540 fi-hsw-4200u fi-byt-squawks fi-bsw-cyan fi-bdw-samus fi-byt-clapper fi-skl-6600u
Build changes
-------------
* Linux: CI_DRM_8554 -> Patchwork_17820
CI-20190529: 20190529
CI_DRM_8554: ac5c0538eb79074e97a7a5c03c285f339290d961 @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_5684: bd399f5eb8263bb4a84ae6a5bb1a13d329e0515d @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
Patchwork_17820: 319babfa21e7433aad375377305d5ad092615dfa @ git://anongit.freedesktop.org/gfx-ci/linux
== Linux commits ==
319babfa21e7 drm/i915: Check for awaits on still currently executing requests
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17820/index.html
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply
* Re: [Linux-kernel-mentees] [fixup v2] kunit: use --build_dir=.kunit as default
From: Vitor Massaru Iha @ 2020-05-29 17:45 UTC (permalink / raw)
To: Shuah Khan
Cc: linux-kernel-mentees, Brendan Higgins, Linux Kernel Mailing List,
open list:KERNEL SELFTEST FRAMEWORK, KUnit Development
In-Reply-To: <9b12d44c-938e-b57f-d3ed-509b8e1b07bc@linuxfoundation.org>
On Thu, May 28, 2020 at 2:28 PM Shuah Khan <skhan@linuxfoundation.org> wrote:
>
> On 5/26/20 6:34 PM, Vitor Massaru Iha wrote:
> > To make KUnit easier to use, and to avoid overwriting object and
> > .config files, the default KUnit build directory is set to .kunit
> >
> > Fixed up minor merge conflicts - Shuah Khan <skhan@linuxfoundation.org>
> >
> > Fixed this identation error exchanging spaces for tabs between lines
> > 248 and 252:
> >
> > tools/testing/kunit/kunit.py run --defconfig
> >
> > File "tools/testing/kunit/kunit.py", line 254
> > if not linux:
> > ^
> > TabError: inconsistent use of tabs and spaces in indentation
> >
> > Signed-off-by: Vitor Massaru Iha <vitor@massaru.org>
> > Reviewed-by: Brendan Higgins <brendanhiggins@google.com>
> > Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
> > Link: https://bugzilla.kernel.org/show_bug.cgi?id=205221
>
> Vitor,
>
> This patch doesn't apply. Can you change to subject to be unique
> Fix TabError and do the patch on top of linux-kselftest kunit
>
> I think you did this, but I can't apply this on top:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git/commit/?h=kunit
>
Sure, I'll do that.
> thanks,
> -- Shuah
_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees
^ permalink raw reply
* Re: [fixup v2] kunit: use --build_dir=.kunit as default
From: Vitor Massaru Iha @ 2020-05-29 17:45 UTC (permalink / raw)
To: Shuah Khan
Cc: KUnit Development, open list:KERNEL SELFTEST FRAMEWORK,
Linux Kernel Mailing List, Brendan Higgins, linux-kernel-mentees
In-Reply-To: <9b12d44c-938e-b57f-d3ed-509b8e1b07bc@linuxfoundation.org>
On Thu, May 28, 2020 at 2:28 PM Shuah Khan <skhan@linuxfoundation.org> wrote:
>
> On 5/26/20 6:34 PM, Vitor Massaru Iha wrote:
> > To make KUnit easier to use, and to avoid overwriting object and
> > .config files, the default KUnit build directory is set to .kunit
> >
> > Fixed up minor merge conflicts - Shuah Khan <skhan@linuxfoundation.org>
> >
> > Fixed this identation error exchanging spaces for tabs between lines
> > 248 and 252:
> >
> > tools/testing/kunit/kunit.py run --defconfig
> >
> > File "tools/testing/kunit/kunit.py", line 254
> > if not linux:
> > ^
> > TabError: inconsistent use of tabs and spaces in indentation
> >
> > Signed-off-by: Vitor Massaru Iha <vitor@massaru.org>
> > Reviewed-by: Brendan Higgins <brendanhiggins@google.com>
> > Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
> > Link: https://bugzilla.kernel.org/show_bug.cgi?id=205221
>
> Vitor,
>
> This patch doesn't apply. Can you change to subject to be unique
> Fix TabError and do the patch on top of linux-kselftest kunit
>
> I think you did this, but I can't apply this on top:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git/commit/?h=kunit
>
Sure, I'll do that.
> thanks,
> -- Shuah
^ permalink raw reply
* [PATCH 2/2] linux-ti-staging: update to 5.4.40 and 07.00.00.004 tag
From: Dan Murphy @ 2020-05-29 17:45 UTC (permalink / raw)
To: meta-ti, denys; +Cc: Dan Murphy
In-Reply-To: <20200529174505.27706-1-dmurphy@ti.com>
Signed-off-by: Dan Murphy <dmurphy@ti.com>
---
recipes-kernel/linux/linux-ti-staging-rt_5.4.bb | 2 +-
recipes-kernel/linux/linux-ti-staging_5.4.bb | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/recipes-kernel/linux/linux-ti-staging-rt_5.4.bb b/recipes-kernel/linux/linux-ti-staging-rt_5.4.bb
index 7a2492a9..6f9f4619 100644
--- a/recipes-kernel/linux/linux-ti-staging-rt_5.4.bb
+++ b/recipes-kernel/linux/linux-ti-staging-rt_5.4.bb
@@ -6,5 +6,5 @@ FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}-5.4:"
BRANCH = "ti-rt-linux-5.4.y"
-SRCREV = "83062031887b4a57d8e7761d56b6a08b931c8ad4"
+SRCREV = "8591cbda22d3b1a85bb8fae1195ce06323949bd3"
PV = "5.4.40+git${SRCPV}"
diff --git a/recipes-kernel/linux/linux-ti-staging_5.4.bb b/recipes-kernel/linux/linux-ti-staging_5.4.bb
index 27495512..e705c0ad 100644
--- a/recipes-kernel/linux/linux-ti-staging_5.4.bb
+++ b/recipes-kernel/linux/linux-ti-staging_5.4.bb
@@ -61,7 +61,7 @@ S = "${WORKDIR}/git"
BRANCH = "ti-linux-5.4.y"
-SRCREV = "7f8dc0f36a11569b8e901d8a466b5fcb4f7f01b8"
+SRCREV = "bf0b43568bb5b401970127e59bbbc7921cebe6d2"
PV = "5.4.40+git${SRCPV}"
# Append to the MACHINE_KERNEL_PR so that a new SRCREV will cause a rebuild
--
2.26.2
^ permalink raw reply related
* Re: [PATCH] spi: bcm2835: Enable shared interrupt support
From: Florian Fainelli @ 2020-05-29 17:46 UTC (permalink / raw)
To: Lukas Wunner, Nicolas Saenz Julienne
Cc: Scott Branden, Ray Jui, linux-kernel, linux-spi, Mark Brown,
bcm-kernel-feedback-list, linux-rpi-kernel, Martin Sperl,
linux-arm-kernel
In-Reply-To: <20200529174358.som3snunfxch6phi@wunner.de>
On 5/29/20 10:43 AM, Lukas Wunner wrote:
> On Thu, May 28, 2020 at 08:58:04PM +0200, Nicolas Saenz Julienne wrote:
>> --- a/drivers/spi/spi-bcm2835.c
>> +++ b/drivers/spi/spi-bcm2835.c
>> @@ -379,6 +379,10 @@ static irqreturn_t bcm2835_spi_interrupt(int irq, void *dev_id)
>> if (bs->tx_len && cs & BCM2835_SPI_CS_DONE)
>> bcm2835_wr_fifo_blind(bs, BCM2835_SPI_FIFO_SIZE);
>>
>> + /* check if we got interrupt enabled */
>> + if (!(bcm2835_rd(bs, BCM2835_SPI_CS) & BCM2835_SPI_CS_INTR))
>> + return IRQ_NONE;
>> +
>> /* Read as many bytes as possible from FIFO */
>> bcm2835_rd_fifo(bs);
>> /* Write as many bytes as possible to FIFO */
>
> This definitely looks wrong. The check whether the interrupt is enabled
> should be moved *before* the conditional calls to bcm2835_rd_fifo_blind()
> and bcm2835_wr_fifo_blind(), i.e. to the top of the function.
>
> Otherwise if an interrupt is raised by another SPI controller,
> this function may perform read/write accesses to the FIFO and
> interfere with an ongoing transfer in DMA or poll mode.
>
> Also, instead of performing an MMIO read, just use the "cs" variable
> which was assigned at the top of the function.
>
> The code comment should explain that the check is necessary because the
> interrupt may be shared with other controllers on the BCM2711.
>
> Finally, it would be nice if the check would be optimized away when
> compiling for pre-RasPi4 products, maybe something like:
>
> + if (IS_ENABLED(CONFIG_ARM_LPAE) && !(cs & BCM2835_SPI_CS_INTR))
> + return IRQ_NONE;
Rather than keying this off ARM_LPAE or any other option, this should be
keyed off a compatible string, that way we can even conditionally pass
IRQF_SHARED to the interrupt handler if we care so much about performance.
--
Florian
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply
* [PATCH] sh: Implement __get_user_u64() required for 64-bit get_user()
From: John Paul Adrian Glaubitz @ 2020-05-29 17:45 UTC (permalink / raw)
To: linux-sh
Cc: Rich Felker, Yoshinori Sato, Geert Uytterhoeven, Michael Karcher,
linux-kernel, John Paul Adrian Glaubitz
In-Reply-To: <20200529174540.4189874-1-glaubitz@physik.fu-berlin.de>
Trying to build the kernel with CONFIG_INFINIBAND_USER_ACCESS enabled fails
ERROR: "__get_user_unknown" [drivers/infiniband/core/ib_uverbs.ko] undefined!
with on SH since the kernel misses a 64-bit implementation of get_user().
Implement the missing 64-bit get_user() as __get_user_u64(), matching the
already existing __put_user_u64() which implements the 64-bit put_user().
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
---
arch/sh/include/asm/uaccess_32.h | 49 ++++++++++++++++++++++++++++++++
1 file changed, 49 insertions(+)
diff --git a/arch/sh/include/asm/uaccess_32.h b/arch/sh/include/asm/uaccess_32.h
index 624cf55acc27..8bc1cb50f8bf 100644
--- a/arch/sh/include/asm/uaccess_32.h
+++ b/arch/sh/include/asm/uaccess_32.h
@@ -26,6 +26,9 @@ do { \
case 4: \
__get_user_asm(x, ptr, retval, "l"); \
break; \
+ case 8: \
+ __get_user_u64(x, ptr, retval); \
+ break; \
default: \
__get_user_unknown(); \
break; \
@@ -66,6 +69,52 @@ do { \
extern void __get_user_unknown(void);
+#if defined(CONFIG_CPU_LITTLE_ENDIAN)
+#define __get_user_u64(x, addr, err) \
+({ \
+__asm__ __volatile__( \
+ "1:\n\t" \
+ "mov.l %2,%R1\n\t" \
+ "mov.l %T2,%S1\n\t" \
+ "2:\n" \
+ ".section .fixup,\"ax\"\n" \
+ "3:\n\t" \
+ "mov #0, %1\n\t" \
+ "mov.l 4f, %0\n\t" \
+ "jmp @%0\n\t" \
+ " mov %3, %0\n\t" \
+ ".balign 4\n" \
+ "4: .long 2b\n\t" \
+ ".previous\n" \
+ ".section __ex_table,\"a\"\n\t" \
+ ".long 1b, 3b\n\t" \
+ ".previous" \
+ :"=&r" (err), "=&r" (x) \
+ :"m" (__m(addr)), "i" (-EFAULT), "0" (err)); })
+#else
+#define __get_user_u64(x, addr, err) \
+({ \
+__asm__ __volatile__( \
+ "1:\n\t" \
+ "mov.l %2,%S1\n\t" \
+ "mov.l %T2,%R1\n\t" \
+ "2:\n" \
+ ".section .fixup,\"ax\"\n" \
+ "3:\n\t" \
+ "mov #0, %1\n\t" \
+ "mov.l 4f, %0\n\t" \
+ "jmp @%0\n\t" \
+ " mov %3, %0\n\t" \
+ ".balign 4\n" \
+ "4: .long 2b\n\t" \
+ ".previous\n" \
+ ".section __ex_table,\"a\"\n\t" \
+ ".long 1b, 3b\n\t" \
+ ".previous" \
+ :"=&r" (err), "=&r" (x) \
+ :"m" (__m(addr)), "i" (-EFAULT), "0" (err)); })
+#endif
+
#define __put_user_size(x,ptr,size,retval) \
do { \
retval = 0; \
--
2.27.0.rc2
^ permalink raw reply related
* Re: [PATCH] spi: bcm2835: Enable shared interrupt support
From: Florian Fainelli @ 2020-05-29 17:46 UTC (permalink / raw)
To: Lukas Wunner, Nicolas Saenz Julienne
Cc: Mark Brown, Ray Jui, Scott Branden, bcm-kernel-feedback-list,
Martin Sperl, linux-spi, linux-rpi-kernel, linux-arm-kernel,
linux-kernel
In-Reply-To: <20200529174358.som3snunfxch6phi@wunner.de>
On 5/29/20 10:43 AM, Lukas Wunner wrote:
> On Thu, May 28, 2020 at 08:58:04PM +0200, Nicolas Saenz Julienne wrote:
>> --- a/drivers/spi/spi-bcm2835.c
>> +++ b/drivers/spi/spi-bcm2835.c
>> @@ -379,6 +379,10 @@ static irqreturn_t bcm2835_spi_interrupt(int irq, void *dev_id)
>> if (bs->tx_len && cs & BCM2835_SPI_CS_DONE)
>> bcm2835_wr_fifo_blind(bs, BCM2835_SPI_FIFO_SIZE);
>>
>> + /* check if we got interrupt enabled */
>> + if (!(bcm2835_rd(bs, BCM2835_SPI_CS) & BCM2835_SPI_CS_INTR))
>> + return IRQ_NONE;
>> +
>> /* Read as many bytes as possible from FIFO */
>> bcm2835_rd_fifo(bs);
>> /* Write as many bytes as possible to FIFO */
>
> This definitely looks wrong. The check whether the interrupt is enabled
> should be moved *before* the conditional calls to bcm2835_rd_fifo_blind()
> and bcm2835_wr_fifo_blind(), i.e. to the top of the function.
>
> Otherwise if an interrupt is raised by another SPI controller,
> this function may perform read/write accesses to the FIFO and
> interfere with an ongoing transfer in DMA or poll mode.
>
> Also, instead of performing an MMIO read, just use the "cs" variable
> which was assigned at the top of the function.
>
> The code comment should explain that the check is necessary because the
> interrupt may be shared with other controllers on the BCM2711.
>
> Finally, it would be nice if the check would be optimized away when
> compiling for pre-RasPi4 products, maybe something like:
>
> + if (IS_ENABLED(CONFIG_ARM_LPAE) && !(cs & BCM2835_SPI_CS_INTR))
> + return IRQ_NONE;
Rather than keying this off ARM_LPAE or any other option, this should be
keyed off a compatible string, that way we can even conditionally pass
IRQF_SHARED to the interrupt handler if we care so much about performance.
--
Florian
^ permalink raw reply
* [RESEND] sh: Implement __get_user_u64() required for 64-bit get_user()
From: John Paul Adrian Glaubitz @ 2020-05-29 17:45 UTC (permalink / raw)
To: linux-sh
Cc: Rich Felker, Yoshinori Sato, Geert Uytterhoeven, Michael Karcher,
linux-kernel
Hi!
This is my attempt of implementing a 64-bit get_user() for SH to address the
build problem when CONFIG_INFINIBAND_USER_ACCESS is enabled.
I have carefully looked at the existing implementations of __get_user_asm(),
__put_user_asm() and the 64-bit __put_user_u64() to come up with the 64-bit
__get_user_u64().
I'm admittedly not an expert when it comes to writing GCC contraints, so the
code might be completely wrong. However, it builds fine without warnings
and fixes the aforementioned issue for me.
Hopefully someone from the more experienced group of kernel developers can
review my code and help me get it into proper shape for submission.
Resent because I forgot to add a subject for the first cover text.
Thanks,
Adrian
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@debian.org
`. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
^ permalink raw reply
* [PATCH] sh: Implement __get_user_u64() required for 64-bit get_user()
From: John Paul Adrian Glaubitz @ 2020-05-29 17:45 UTC (permalink / raw)
To: linux-sh
Cc: Rich Felker, Yoshinori Sato, Geert Uytterhoeven, Michael Karcher,
linux-kernel, John Paul Adrian Glaubitz
In-Reply-To: <20200529174540.4189874-1-glaubitz@physik.fu-berlin.de>
Trying to build the kernel with CONFIG_INFINIBAND_USER_ACCESS enabled fails
ERROR: "__get_user_unknown" [drivers/infiniband/core/ib_uverbs.ko] undefined!
with on SH since the kernel misses a 64-bit implementation of get_user().
Implement the missing 64-bit get_user() as __get_user_u64(), matching the
already existing __put_user_u64() which implements the 64-bit put_user().
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
---
arch/sh/include/asm/uaccess_32.h | 49 ++++++++++++++++++++++++++++++++
1 file changed, 49 insertions(+)
diff --git a/arch/sh/include/asm/uaccess_32.h b/arch/sh/include/asm/uaccess_32.h
index 624cf55acc27..8bc1cb50f8bf 100644
--- a/arch/sh/include/asm/uaccess_32.h
+++ b/arch/sh/include/asm/uaccess_32.h
@@ -26,6 +26,9 @@ do { \
case 4: \
__get_user_asm(x, ptr, retval, "l"); \
break; \
+ case 8: \
+ __get_user_u64(x, ptr, retval); \
+ break; \
default: \
__get_user_unknown(); \
break; \
@@ -66,6 +69,52 @@ do { \
extern void __get_user_unknown(void);
+#if defined(CONFIG_CPU_LITTLE_ENDIAN)
+#define __get_user_u64(x, addr, err) \
+({ \
+__asm__ __volatile__( \
+ "1:\n\t" \
+ "mov.l %2,%R1\n\t" \
+ "mov.l %T2,%S1\n\t" \
+ "2:\n" \
+ ".section .fixup,\"ax\"\n" \
+ "3:\n\t" \
+ "mov #0, %1\n\t" \
+ "mov.l 4f, %0\n\t" \
+ "jmp @%0\n\t" \
+ " mov %3, %0\n\t" \
+ ".balign 4\n" \
+ "4: .long 2b\n\t" \
+ ".previous\n" \
+ ".section __ex_table,\"a\"\n\t" \
+ ".long 1b, 3b\n\t" \
+ ".previous" \
+ :"=&r" (err), "=&r" (x) \
+ :"m" (__m(addr)), "i" (-EFAULT), "0" (err)); })
+#else
+#define __get_user_u64(x, addr, err) \
+({ \
+__asm__ __volatile__( \
+ "1:\n\t" \
+ "mov.l %2,%S1\n\t" \
+ "mov.l %T2,%R1\n\t" \
+ "2:\n" \
+ ".section .fixup,\"ax\"\n" \
+ "3:\n\t" \
+ "mov #0, %1\n\t" \
+ "mov.l 4f, %0\n\t" \
+ "jmp @%0\n\t" \
+ " mov %3, %0\n\t" \
+ ".balign 4\n" \
+ "4: .long 2b\n\t" \
+ ".previous\n" \
+ ".section __ex_table,\"a\"\n\t" \
+ ".long 1b, 3b\n\t" \
+ ".previous" \
+ :"=&r" (err), "=&r" (x) \
+ :"m" (__m(addr)), "i" (-EFAULT), "0" (err)); })
+#endif
+
#define __put_user_size(x,ptr,size,retval) \
do { \
retval = 0; \
--
2.27.0.rc2
^ permalink raw reply related
* [RESEND] sh: Implement __get_user_u64() required for 64-bit get_user()
From: John Paul Adrian Glaubitz @ 2020-05-29 17:45 UTC (permalink / raw)
To: linux-sh
Cc: Rich Felker, Yoshinori Sato, Geert Uytterhoeven, Michael Karcher,
linux-kernel
Hi!
This is my attempt of implementing a 64-bit get_user() for SH to address the
build problem when CONFIG_INFINIBAND_USER_ACCESS is enabled.
I have carefully looked at the existing implementations of __get_user_asm(),
__put_user_asm() and the 64-bit __put_user_u64() to come up with the 64-bit
__get_user_u64().
I'm admittedly not an expert when it comes to writing GCC contraints, so the
code might be completely wrong. However, it builds fine without warnings
and fixes the aforementioned issue for me.
Hopefully someone from the more experienced group of kernel developers can
review my code and help me get it into proper shape for submission.
Resent because I forgot to add a subject for the first cover text.
Thanks,
Adrian
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@debian.org
`. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
^ permalink raw reply
* Re: [PATCH] usb: dwc2: Fix shutdown callback in platform
From: Frank Mori Hess @ 2020-05-29 17:44 UTC (permalink / raw)
To: Doug Anderson
Cc: Minas Harutyunyan, John Youn, Felipe Balbi, Greg Kroah-Hartman,
linux-usb@vger.kernel.org, Felipe Balbi, Heiko Stübner,
# 4.0+
In-Reply-To: <CAD=FV=UCMqyX92o9m7H40E3sHzAFieHSu3TUY953VqNb-vuPPg@mail.gmail.com>
On Fri, May 29, 2020 at 12:37 PM Doug Anderson <dianders@chromium.org> wrote:
>
> I'm not sure I understand. Are you saying that you'll just add
> shutdown callbacks to all the drivers using this IRQ and call
> disable_irq() there? That seems like the best solution to me.
I don't get it. A hypothetical machine could have literally anything
sharing the IRQ line, right? If it is important to call disable_irq
during shutdown (I have no idea if it is) then shouldn't the kernel
core just disable all irqs after calling all the driver shutdown
callbacks?
Anyways, my screaming interrupt occurs after a a new kernel has been
booted with kexec. In this case, it doesn't matter if the old kernel
called disable_irq or not. As soon as the new kernel re-enables the
interrupt line, the kernel immediately disables it again with a
backtrace due to the unhandled screaming interrupt. That's why the
dwc2 hardware needs to have its interrupts turned off when the old
kernel is shutdown.
--
Frank
^ permalink raw reply
* [PATCH 1/2] u-boot-ti-staging: update to 07.00.00.004 tag
From: Dan Murphy @ 2020-05-29 17:45 UTC (permalink / raw)
To: meta-ti, denys; +Cc: Dan Murphy
Signed-off-by: Dan Murphy <dmurphy@ti.com>
---
recipes-bsp/u-boot/u-boot-ti-staging_2020.01.bb | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/recipes-bsp/u-boot/u-boot-ti-staging_2020.01.bb b/recipes-bsp/u-boot/u-boot-ti-staging_2020.01.bb
index 7a5f4bcb..cfa500a8 100644
--- a/recipes-bsp/u-boot/u-boot-ti-staging_2020.01.bb
+++ b/recipes-bsp/u-boot/u-boot-ti-staging_2020.01.bb
@@ -1,7 +1,7 @@
require u-boot-ti.inc
-PR = "r10"
+PR = "r11"
BRANCH = "ti-u-boot-2020.01"
-SRCREV = "34fcb949aa53c57d13af1be920c15c5a2e68a019"
+SRCREV = "aa92fa7968d0b3ead6b3513ade990efe12ed5971"
--
2.26.2
^ permalink raw reply related
* [PATCH net-next v3 2/7] dpaa2-eth: Distribute ingress frames based on VLAN prio
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Radulescu, Ioana Ciornei
In-Reply-To: <20200529174345.27537-1-ioana.ciornei@nxp.com>
From: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Configure static ingress classification based on VLAN PCP field.
If the DPNI doesn't have enough traffic classes to accommodate all
priority levels, the lowest ones end up on TC 0 (default on miss).
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
Changes in v3:
- revert to explicitly cast mask to u16 * to not get into
sparse warnings
.../net/ethernet/freescale/dpaa2/dpaa2-eth.c | 116 ++++++++++++++++
.../net/ethernet/freescale/dpaa2/dpaa2-eth.h | 1 +
.../net/ethernet/freescale/dpaa2/dpni-cmd.h | 34 +++++
drivers/net/ethernet/freescale/dpaa2/dpni.c | 131 ++++++++++++++++++
drivers/net/ethernet/freescale/dpaa2/dpni.h | 36 +++++
5 files changed, 318 insertions(+)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index 01263e247d39..b3722ad60fae 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -2696,6 +2696,118 @@ static void update_tx_fqids(struct dpaa2_eth_priv *priv)
priv->enqueue = dpaa2_eth_enqueue_qd;
}
+/* Configure ingress classification based on VLAN PCP */
+static int set_vlan_qos(struct dpaa2_eth_priv *priv)
+{
+ struct device *dev = priv->net_dev->dev.parent;
+ struct dpkg_profile_cfg kg_cfg = {0};
+ struct dpni_qos_tbl_cfg qos_cfg = {0};
+ struct dpni_rule_cfg key_params;
+ void *dma_mem, *key, *mask;
+ u8 key_size = 2; /* VLAN TCI field */
+ int i, pcp, err;
+
+ /* VLAN-based classification only makes sense if we have multiple
+ * traffic classes.
+ * Also, we need to extract just the 3-bit PCP field from the VLAN
+ * header and we can only do that by using a mask
+ */
+ if (dpaa2_eth_tc_count(priv) == 1 || !dpaa2_eth_fs_mask_enabled(priv)) {
+ dev_dbg(dev, "VLAN-based QoS classification not supported\n");
+ return -EOPNOTSUPP;
+ }
+
+ dma_mem = kzalloc(DPAA2_CLASSIFIER_DMA_SIZE, GFP_KERNEL);
+ if (!dma_mem)
+ return -ENOMEM;
+
+ kg_cfg.num_extracts = 1;
+ kg_cfg.extracts[0].type = DPKG_EXTRACT_FROM_HDR;
+ kg_cfg.extracts[0].extract.from_hdr.prot = NET_PROT_VLAN;
+ kg_cfg.extracts[0].extract.from_hdr.type = DPKG_FULL_FIELD;
+ kg_cfg.extracts[0].extract.from_hdr.field = NH_FLD_VLAN_TCI;
+
+ err = dpni_prepare_key_cfg(&kg_cfg, dma_mem);
+ if (err) {
+ dev_err(dev, "dpni_prepare_key_cfg failed\n");
+ goto out_free_tbl;
+ }
+
+ /* set QoS table */
+ qos_cfg.default_tc = 0;
+ qos_cfg.discard_on_miss = 0;
+ qos_cfg.key_cfg_iova = dma_map_single(dev, dma_mem,
+ DPAA2_CLASSIFIER_DMA_SIZE,
+ DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, qos_cfg.key_cfg_iova)) {
+ dev_err(dev, "QoS table DMA mapping failed\n");
+ err = -ENOMEM;
+ goto out_free_tbl;
+ }
+
+ err = dpni_set_qos_table(priv->mc_io, 0, priv->mc_token, &qos_cfg);
+ if (err) {
+ dev_err(dev, "dpni_set_qos_table failed\n");
+ goto out_unmap_tbl;
+ }
+
+ /* Add QoS table entries */
+ key = kzalloc(key_size * 2, GFP_KERNEL);
+ if (!key) {
+ err = -ENOMEM;
+ goto out_unmap_tbl;
+ }
+ mask = key + key_size;
+ *(u16 *)mask = cpu_to_be16(VLAN_PRIO_MASK);
+
+ key_params.key_iova = dma_map_single(dev, key, key_size * 2,
+ DMA_TO_DEVICE);
+ if (dma_mapping_error(dev, key_params.key_iova)) {
+ dev_err(dev, "Qos table entry DMA mapping failed\n");
+ err = -ENOMEM;
+ goto out_free_key;
+ }
+
+ key_params.mask_iova = key_params.key_iova + key_size;
+ key_params.key_size = key_size;
+
+ /* We add rules for PCP-based distribution starting with highest
+ * priority (VLAN PCP = 7). If this DPNI doesn't have enough traffic
+ * classes to accommodate all priority levels, the lowest ones end up
+ * on TC 0 which was configured as default
+ */
+ for (i = dpaa2_eth_tc_count(priv) - 1, pcp = 7; i >= 0; i--, pcp--) {
+ *(u16 *)key = cpu_to_be16(pcp << VLAN_PRIO_SHIFT);
+ dma_sync_single_for_device(dev, key_params.key_iova,
+ key_size * 2, DMA_TO_DEVICE);
+
+ err = dpni_add_qos_entry(priv->mc_io, 0, priv->mc_token,
+ &key_params, i, i);
+ if (err) {
+ dev_err(dev, "dpni_add_qos_entry failed\n");
+ dpni_clear_qos_table(priv->mc_io, 0, priv->mc_token);
+ goto out_unmap_key;
+ }
+ }
+
+ priv->vlan_cls_enabled = true;
+
+ /* Table and key memory is not persistent, clean everything up after
+ * configuration is finished
+ */
+out_unmap_key:
+ dma_unmap_single(dev, key_params.key_iova, key_size * 2, DMA_TO_DEVICE);
+out_free_key:
+ kfree(key);
+out_unmap_tbl:
+ dma_unmap_single(dev, qos_cfg.key_cfg_iova, DPAA2_CLASSIFIER_DMA_SIZE,
+ DMA_TO_DEVICE);
+out_free_tbl:
+ kfree(dma_mem);
+
+ return err;
+}
+
/* Configure the DPNI object this interface is associated with */
static int setup_dpni(struct fsl_mc_device *ls_dev)
{
@@ -2758,6 +2870,10 @@ static int setup_dpni(struct fsl_mc_device *ls_dev)
goto close;
}
+ err = set_vlan_qos(priv);
+ if (err && err != -EOPNOTSUPP)
+ goto close;
+
priv->cls_rules = devm_kzalloc(dev, sizeof(struct dpaa2_eth_cls_rule) *
dpaa2_eth_fs_count(priv), GFP_KERNEL);
if (!priv->cls_rules) {
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 580ad5fd7bd8..7856f69bcf36 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -427,6 +427,7 @@ struct dpaa2_eth_priv {
u64 rx_cls_fields;
struct dpaa2_eth_cls_rule *cls_rules;
u8 rx_cls_enabled;
+ u8 vlan_cls_enabled;
struct bpf_prog *xdp_prog;
#ifdef CONFIG_DEBUG_FS
struct dpaa2_debugfs dbg;
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h b/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
index d9b6918807af..0048e856f85e 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
@@ -59,6 +59,10 @@
#define DPNI_CMDID_SET_RX_TC_DIST DPNI_CMD(0x235)
+#define DPNI_CMDID_SET_QOS_TBL DPNI_CMD(0x240)
+#define DPNI_CMDID_ADD_QOS_ENT DPNI_CMD(0x241)
+#define DPNI_CMDID_REMOVE_QOS_ENT DPNI_CMD(0x242)
+#define DPNI_CMDID_CLR_QOS_TBL DPNI_CMD(0x243)
#define DPNI_CMDID_ADD_FS_ENT DPNI_CMD(0x244)
#define DPNI_CMDID_REMOVE_FS_ENT DPNI_CMD(0x245)
#define DPNI_CMDID_CLR_FS_ENT DPNI_CMD(0x246)
@@ -567,4 +571,34 @@ struct dpni_cmd_remove_fs_entry {
__le64 mask_iova;
};
+#define DPNI_DISCARD_ON_MISS_SHIFT 0
+#define DPNI_DISCARD_ON_MISS_SIZE 1
+
+struct dpni_cmd_set_qos_table {
+ __le32 pad;
+ u8 default_tc;
+ /* only the LSB */
+ u8 discard_on_miss;
+ __le16 pad1[21];
+ __le64 key_cfg_iova;
+};
+
+struct dpni_cmd_add_qos_entry {
+ __le16 pad;
+ u8 tc_id;
+ u8 key_size;
+ __le16 index;
+ __le16 pad1;
+ __le64 key_iova;
+ __le64 mask_iova;
+};
+
+struct dpni_cmd_remove_qos_entry {
+ u8 pad[3];
+ u8 key_size;
+ __le32 pad1;
+ __le64 key_iova;
+ __le64 mask_iova;
+};
+
#endif /* _FSL_DPNI_CMD_H */
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni.c b/drivers/net/ethernet/freescale/dpaa2/dpni.c
index dd54e6953aeb..78fa325407ca 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni.c
@@ -1786,3 +1786,134 @@ int dpni_remove_fs_entry(struct fsl_mc_io *mc_io,
/* send command to mc*/
return mc_send_command(mc_io, &cmd);
}
+
+/**
+ * dpni_set_qos_table() - Set QoS mapping table
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @token: Token of DPNI object
+ * @cfg: QoS table configuration
+ *
+ * This function and all QoS-related functions require that
+ *'max_tcs > 1' was set at DPNI creation.
+ *
+ * warning: Before calling this function, call dpkg_prepare_key_cfg() to
+ * prepare the key_cfg_iova parameter
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
+int dpni_set_qos_table(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ const struct dpni_qos_tbl_cfg *cfg)
+{
+ struct dpni_cmd_set_qos_table *cmd_params;
+ struct fsl_mc_command cmd = { 0 };
+
+ /* prepare command */
+ cmd.header = mc_encode_cmd_header(DPNI_CMDID_SET_QOS_TBL,
+ cmd_flags,
+ token);
+ cmd_params = (struct dpni_cmd_set_qos_table *)cmd.params;
+ cmd_params->default_tc = cfg->default_tc;
+ cmd_params->key_cfg_iova = cpu_to_le64(cfg->key_cfg_iova);
+ dpni_set_field(cmd_params->discard_on_miss, DISCARD_ON_MISS,
+ cfg->discard_on_miss);
+
+ /* send command to mc*/
+ return mc_send_command(mc_io, &cmd);
+}
+
+/**
+ * dpni_add_qos_entry() - Add QoS mapping entry (to select a traffic class)
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @token: Token of DPNI object
+ * @cfg: QoS rule to add
+ * @tc_id: Traffic class selection (0-7)
+ * @index: Location in the QoS table where to insert the entry.
+ * Only relevant if MASKING is enabled for QoS classification on
+ * this DPNI, it is ignored for exact match.
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
+int dpni_add_qos_entry(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ const struct dpni_rule_cfg *cfg,
+ u8 tc_id,
+ u16 index)
+{
+ struct dpni_cmd_add_qos_entry *cmd_params;
+ struct fsl_mc_command cmd = { 0 };
+
+ /* prepare command */
+ cmd.header = mc_encode_cmd_header(DPNI_CMDID_ADD_QOS_ENT,
+ cmd_flags,
+ token);
+ cmd_params = (struct dpni_cmd_add_qos_entry *)cmd.params;
+ cmd_params->tc_id = tc_id;
+ cmd_params->key_size = cfg->key_size;
+ cmd_params->index = cpu_to_le16(index);
+ cmd_params->key_iova = cpu_to_le64(cfg->key_iova);
+ cmd_params->mask_iova = cpu_to_le64(cfg->mask_iova);
+
+ /* send command to mc*/
+ return mc_send_command(mc_io, &cmd);
+}
+
+/**
+ * dpni_remove_qos_entry() - Remove QoS mapping entry
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @token: Token of DPNI object
+ * @cfg: QoS rule to remove
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
+int dpni_remove_qos_entry(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ const struct dpni_rule_cfg *cfg)
+{
+ struct dpni_cmd_remove_qos_entry *cmd_params;
+ struct fsl_mc_command cmd = { 0 };
+
+ /* prepare command */
+ cmd.header = mc_encode_cmd_header(DPNI_CMDID_REMOVE_QOS_ENT,
+ cmd_flags,
+ token);
+ cmd_params = (struct dpni_cmd_remove_qos_entry *)cmd.params;
+ cmd_params->key_size = cfg->key_size;
+ cmd_params->key_iova = cpu_to_le64(cfg->key_iova);
+ cmd_params->mask_iova = cpu_to_le64(cfg->mask_iova);
+
+ /* send command to mc*/
+ return mc_send_command(mc_io, &cmd);
+}
+
+/**
+ * dpni_clear_qos_table() - Clear all QoS mapping entries
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @token: Token of DPNI object
+ *
+ * Following this function call, all frames are directed to
+ * the default traffic class (0)
+ *
+ * Return: '0' on Success; Error code otherwise.
+ */
+int dpni_clear_qos_table(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token)
+{
+ struct fsl_mc_command cmd = { 0 };
+
+ /* prepare command */
+ cmd.header = mc_encode_cmd_header(DPNI_CMDID_CLR_QOS_TBL,
+ cmd_flags,
+ token);
+
+ /* send command to mc*/
+ return mc_send_command(mc_io, &cmd);
+}
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni.h b/drivers/net/ethernet/freescale/dpaa2/dpni.h
index ee0711d06b3a..8c7ac20bf1a7 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni.h
@@ -715,6 +715,26 @@ int dpni_set_rx_hash_dist(struct fsl_mc_io *mc_io,
u16 token,
const struct dpni_rx_dist_cfg *cfg);
+/**
+ * struct dpni_qos_tbl_cfg - Structure representing QOS table configuration
+ * @key_cfg_iova: I/O virtual address of 256 bytes DMA-able memory filled with
+ * key extractions to be used as the QoS criteria by calling
+ * dpkg_prepare_key_cfg()
+ * @discard_on_miss: Set to '1' to discard frames in case of no match (miss);
+ * '0' to use the 'default_tc' in such cases
+ * @default_tc: Used in case of no-match and 'discard_on_miss'= 0
+ */
+struct dpni_qos_tbl_cfg {
+ u64 key_cfg_iova;
+ int discard_on_miss;
+ u8 default_tc;
+};
+
+int dpni_set_qos_table(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ const struct dpni_qos_tbl_cfg *cfg);
+
/**
* enum dpni_dest - DPNI destination types
* @DPNI_DEST_NONE: Unassigned destination; The queue is set in parked mode and
@@ -961,6 +981,22 @@ int dpni_remove_fs_entry(struct fsl_mc_io *mc_io,
u8 tc_id,
const struct dpni_rule_cfg *cfg);
+int dpni_add_qos_entry(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ const struct dpni_rule_cfg *cfg,
+ u8 tc_id,
+ u16 index);
+
+int dpni_remove_qos_entry(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ const struct dpni_rule_cfg *cfg);
+
+int dpni_clear_qos_table(struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token);
+
int dpni_get_api_version(struct fsl_mc_io *mc_io,
u32 cmd_flags,
u16 *major_ver,
--
2.17.1
^ permalink raw reply related
* [PATCH net-next v3 7/7] dpaa2-eth: Keep congestion group taildrop enabled when PFC on
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Ciornei
In-Reply-To: <20200529174345.27537-1-ioana.ciornei@nxp.com>
Leave congestion group taildrop enabled for all traffic classes
when PFC is enabled. Notification threshold is low enough such
that it will be hit first and this also ensures that FQs on
traffic classes which are not PFC enabled won't drain the buffer
pool.
FQ taildrop threshold is kept disabled as long as any form of
flow control is on. Since FQ taildrop works with bytes, not number
of frames, we can't guarantee it will not interfere with the
congestion notification mechanism for all frame sizes.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
Changes in v3:
- new patch
.../ethernet/freescale/dpaa2/dpaa2-eth-dcb.c | 8 +++--
.../net/ethernet/freescale/dpaa2/dpaa2-eth.c | 29 ++++++++++++++-----
.../net/ethernet/freescale/dpaa2/dpaa2-eth.h | 7 ++++-
3 files changed, 34 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c
index 7ee07872af4d..83dee575c2fa 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c
@@ -63,6 +63,7 @@ static int dpaa2_eth_dcbnl_ieee_setpfc(struct net_device *net_dev,
{
struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
struct dpni_link_cfg link_cfg = {0};
+ bool tx_pause;
int err;
if (pfc->mbc || pfc->delay)
@@ -75,8 +76,8 @@ static int dpaa2_eth_dcbnl_ieee_setpfc(struct net_device *net_dev,
/* We allow PFC configuration even if it won't have any effect until
* general pause frames are enabled
*/
- if (!dpaa2_eth_rx_pause_enabled(priv->link_state.options) ||
- !dpaa2_eth_tx_pause_enabled(priv->link_state.options))
+ tx_pause = dpaa2_eth_tx_pause_enabled(priv->link_state.options);
+ if (!dpaa2_eth_rx_pause_enabled(priv->link_state.options) || !tx_pause)
netdev_warn(net_dev, "Pause support must be enabled in order for PFC to work!\n");
link_cfg.rate = priv->link_state.rate;
@@ -97,6 +98,9 @@ static int dpaa2_eth_dcbnl_ieee_setpfc(struct net_device *net_dev,
return err;
memcpy(&priv->pfc, pfc, sizeof(priv->pfc));
+ priv->pfc_enabled = !!pfc->pfc_en;
+
+ dpaa2_eth_set_rx_taildrop(priv, tx_pause, priv->pfc_enabled);
return 0;
}
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index 1deccd377c5a..b4b573ed3653 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -1287,18 +1287,22 @@ static void disable_ch_napi(struct dpaa2_eth_priv *priv)
}
}
-static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv,
- bool tx_pause)
+void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv,
+ bool tx_pause, bool pfc)
{
struct dpni_taildrop td = {0};
struct dpaa2_eth_fq *fq;
int i, err;
+ /* FQ taildrop: threshold is in bytes, per frame queue. Enabled if
+ * flow control is disabled (as it might interfere with either the
+ * buffer pool depletion trigger for pause frames or with the group
+ * congestion trigger for PFC frames)
+ */
td.enable = !tx_pause;
- if (priv->rx_td_enabled == td.enable)
- return;
+ if (priv->rx_fqtd_enabled == td.enable)
+ goto set_cgtd;
- /* FQ taildrop: threshold is in bytes, per frame queue */
td.threshold = DPAA2_ETH_FQ_TAILDROP_THRESH;
td.units = DPNI_CONGESTION_UNIT_BYTES;
@@ -1316,9 +1320,20 @@ static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv,
}
}
+ priv->rx_fqtd_enabled = td.enable;
+
+set_cgtd:
/* Congestion group taildrop: threshold is in frames, per group
* of FQs belonging to the same traffic class
+ * Enabled if general Tx pause disabled or if PFCs are enabled
+ * (congestion group threhsold for PFC generation is lower than the
+ * CG taildrop threshold, so it won't interfere with it; we also
+ * want frames in non-PFC enabled traffic classes to be kept in check)
*/
+ td.enable = !tx_pause || (tx_pause && pfc);
+ if (priv->rx_cgtd_enabled == td.enable)
+ return;
+
td.threshold = DPAA2_ETH_CG_TAILDROP_THRESH(priv);
td.units = DPNI_CONGESTION_UNIT_FRAMES;
for (i = 0; i < dpaa2_eth_tc_count(priv); i++) {
@@ -1332,7 +1347,7 @@ static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv,
}
}
- priv->rx_td_enabled = td.enable;
+ priv->rx_cgtd_enabled = td.enable;
}
static int link_state_update(struct dpaa2_eth_priv *priv)
@@ -1353,7 +1368,7 @@ static int link_state_update(struct dpaa2_eth_priv *priv)
* only when pause frame generation is disabled.
*/
tx_pause = dpaa2_eth_tx_pause_enabled(state.options);
- dpaa2_eth_set_rx_taildrop(priv, tx_pause);
+ dpaa2_eth_set_rx_taildrop(priv, tx_pause, priv->pfc_enabled);
/* When we manage the MAC/PHY using phylink there is no need
* to manually update the netif_carrier.
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 31b7b9b52da0..2d7ada0f0dbd 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -436,7 +436,8 @@ struct dpaa2_eth_priv {
struct dpaa2_eth_drv_stats __percpu *percpu_extras;
u16 mc_token;
- u8 rx_td_enabled;
+ u8 rx_fqtd_enabled;
+ u8 rx_cgtd_enabled;
struct dpni_link_state link_state;
bool do_link_poll;
@@ -448,6 +449,7 @@ struct dpaa2_eth_priv {
struct dpaa2_eth_cls_rule *cls_rules;
u8 rx_cls_enabled;
u8 vlan_cls_enabled;
+ u8 pfc_enabled;
#ifdef CONFIG_FSL_DPAA2_ETH_DCB
u8 dcbx_mode;
struct ieee_pfc pfc;
@@ -584,6 +586,9 @@ int dpaa2_eth_cls_key_size(u64 key);
int dpaa2_eth_cls_fld_off(int prot, int field);
void dpaa2_eth_cls_trim_rule(void *key_mem, u64 fields);
+void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv,
+ bool tx_pause, bool pfc);
+
extern const struct dcbnl_rtnl_ops dpaa2_eth_dcbnl_ops;
#endif /* __DPAA2_H */
--
2.17.1
^ permalink raw reply related
* [PATCH net-next v3 6/7] dpaa2-eth: Add PFC support through DCB ops
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Ciornei
In-Reply-To: <20200529174345.27537-1-ioana.ciornei@nxp.com>
Add support in dpaa2-eth for PFC (Priority Flow Control)
through the DCB ops.
Instruct the hardware to respond to received PFC frames.
Current firmware doesn't allow us to selectively enable PFC
on the Rx side for some priorities only, so we will react to
all incoming PFC frames (and stop transmitting on the traffic
classes specified in the frame).
Also, configure the hardware to generate PFC frames based on Rx
congestion notifications. When a certain number of frames accumulate in
the ingress queues corresponding to a traffic class, priority flow
control frames are generated for that TC.
The number of PFC traffic classes available can be queried through
lldptool. Also, which of those traffic classes have PFC enabled is also
controlled through the same dcbnl_rtnl_ops callbacks.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
Changes in v3:
- new patch
drivers/net/ethernet/freescale/dpaa2/Kconfig | 10 ++
drivers/net/ethernet/freescale/dpaa2/Makefile | 1 +
.../ethernet/freescale/dpaa2/dpaa2-eth-dcb.c | 146 ++++++++++++++++++
.../net/ethernet/freescale/dpaa2/dpaa2-eth.c | 9 ++
.../net/ethernet/freescale/dpaa2/dpaa2-eth.h | 18 +++
.../net/ethernet/freescale/dpaa2/dpni-cmd.h | 25 +++
drivers/net/ethernet/freescale/dpaa2/dpni.c | 46 ++++++
drivers/net/ethernet/freescale/dpaa2/dpni.h | 61 ++++++++
8 files changed, 316 insertions(+)
create mode 100644 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c
diff --git a/drivers/net/ethernet/freescale/dpaa2/Kconfig b/drivers/net/ethernet/freescale/dpaa2/Kconfig
index c6fb8e4021ac..feea797cde02 100644
--- a/drivers/net/ethernet/freescale/dpaa2/Kconfig
+++ b/drivers/net/ethernet/freescale/dpaa2/Kconfig
@@ -9,6 +9,16 @@ config FSL_DPAA2_ETH
The driver manages network objects discovered on the Freescale
MC bus.
+if FSL_DPAA2_ETH
+config FSL_DPAA2_ETH_DCB
+ bool "Data Center Bridging (DCB) Support"
+ default n
+ depends on DCB
+ help
+ Enable Priority-Based Flow Control (PFC) support for DPAA2 Ethernet
+ devices.
+endif
+
config FSL_DPAA2_PTP_CLOCK
tristate "Freescale DPAA2 PTP Clock"
depends on FSL_DPAA2_ETH && PTP_1588_CLOCK_QORIQ
diff --git a/drivers/net/ethernet/freescale/dpaa2/Makefile b/drivers/net/ethernet/freescale/dpaa2/Makefile
index 69184ca3b7b9..6e7f33c956bf 100644
--- a/drivers/net/ethernet/freescale/dpaa2/Makefile
+++ b/drivers/net/ethernet/freescale/dpaa2/Makefile
@@ -7,6 +7,7 @@ obj-$(CONFIG_FSL_DPAA2_ETH) += fsl-dpaa2-eth.o
obj-$(CONFIG_FSL_DPAA2_PTP_CLOCK) += fsl-dpaa2-ptp.o
fsl-dpaa2-eth-objs := dpaa2-eth.o dpaa2-ethtool.o dpni.o dpaa2-mac.o dpmac.o
+fsl-dpaa2-eth-${CONFIG_FSL_DPAA2_ETH_DCB} += dpaa2-eth-dcb.o
fsl-dpaa2-eth-${CONFIG_DEBUG_FS} += dpaa2-eth-debugfs.o
fsl-dpaa2-ptp-objs := dpaa2-ptp.o dprtc.o
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c
new file mode 100644
index 000000000000..7ee07872af4d
--- /dev/null
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c
@@ -0,0 +1,146 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause)
+/* Copyright 2020 NXP */
+
+#include "dpaa2-eth.h"
+
+static int dpaa2_eth_dcbnl_ieee_getpfc(struct net_device *net_dev,
+ struct ieee_pfc *pfc)
+{
+ struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
+
+ if (!(priv->link_state.options & DPNI_LINK_OPT_PFC_PAUSE))
+ return 0;
+
+ memcpy(pfc, &priv->pfc, sizeof(priv->pfc));
+ pfc->pfc_cap = dpaa2_eth_tc_count(priv);
+
+ return 0;
+}
+
+static inline bool is_prio_enabled(u8 pfc_en, u8 tc)
+{
+ return !!(pfc_en & (1 << tc));
+}
+
+static int set_pfc_cn(struct dpaa2_eth_priv *priv, u8 pfc_en)
+{
+ struct dpni_congestion_notification_cfg cfg = {0};
+ int i, err;
+
+ cfg.notification_mode = DPNI_CONG_OPT_FLOW_CONTROL;
+ cfg.units = DPNI_CONGESTION_UNIT_FRAMES;
+ cfg.message_iova = 0ULL;
+ cfg.message_ctx = 0ULL;
+
+ for (i = 0; i < dpaa2_eth_tc_count(priv); i++) {
+ if (is_prio_enabled(pfc_en, i)) {
+ cfg.threshold_entry = DPAA2_ETH_CN_THRESH_ENTRY(priv);
+ cfg.threshold_exit = DPAA2_ETH_CN_THRESH_EXIT(priv);
+ } else {
+ /* For priorities not set in the pfc_en mask, we leave
+ * the congestion thresholds at zero, which effectively
+ * disables generation of PFC frames for them
+ */
+ cfg.threshold_entry = 0;
+ cfg.threshold_exit = 0;
+ }
+
+ err = dpni_set_congestion_notification(priv->mc_io, 0,
+ priv->mc_token,
+ DPNI_QUEUE_RX, i, &cfg);
+ if (err) {
+ netdev_err(priv->net_dev,
+ "dpni_set_congestion_notification failed\n");
+ return err;
+ }
+ }
+
+ return 0;
+}
+
+static int dpaa2_eth_dcbnl_ieee_setpfc(struct net_device *net_dev,
+ struct ieee_pfc *pfc)
+{
+ struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
+ struct dpni_link_cfg link_cfg = {0};
+ int err;
+
+ if (pfc->mbc || pfc->delay)
+ return -EOPNOTSUPP;
+
+ /* If same PFC enabled mask, nothing to do */
+ if (priv->pfc.pfc_en == pfc->pfc_en)
+ return 0;
+
+ /* We allow PFC configuration even if it won't have any effect until
+ * general pause frames are enabled
+ */
+ if (!dpaa2_eth_rx_pause_enabled(priv->link_state.options) ||
+ !dpaa2_eth_tx_pause_enabled(priv->link_state.options))
+ netdev_warn(net_dev, "Pause support must be enabled in order for PFC to work!\n");
+
+ link_cfg.rate = priv->link_state.rate;
+ link_cfg.options = priv->link_state.options;
+ if (pfc->pfc_en)
+ link_cfg.options |= DPNI_LINK_OPT_PFC_PAUSE;
+ else
+ link_cfg.options &= ~DPNI_LINK_OPT_PFC_PAUSE;
+ err = dpni_set_link_cfg(priv->mc_io, 0, priv->mc_token, &link_cfg);
+ if (err) {
+ netdev_err(net_dev, "dpni_set_link_cfg failed\n");
+ return err;
+ }
+
+ /* Configure congestion notifications for the enabled priorities */
+ err = set_pfc_cn(priv, pfc->pfc_en);
+ if (err)
+ return err;
+
+ memcpy(&priv->pfc, pfc, sizeof(priv->pfc));
+
+ return 0;
+}
+
+static u8 dpaa2_eth_dcbnl_getdcbx(struct net_device *net_dev)
+{
+ struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
+
+ return priv->dcbx_mode;
+}
+
+static u8 dpaa2_eth_dcbnl_setdcbx(struct net_device *net_dev, u8 mode)
+{
+ struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
+
+ return (mode != (priv->dcbx_mode)) ? 1 : 0;
+}
+
+static u8 dpaa2_eth_dcbnl_getcap(struct net_device *net_dev, int capid, u8 *cap)
+{
+ struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
+
+ switch (capid) {
+ case DCB_CAP_ATTR_PFC:
+ *cap = true;
+ break;
+ case DCB_CAP_ATTR_PFC_TCS:
+ *cap = 1 << (dpaa2_eth_tc_count(priv) - 1);
+ break;
+ case DCB_CAP_ATTR_DCBX:
+ *cap = priv->dcbx_mode;
+ break;
+ default:
+ *cap = false;
+ break;
+ }
+
+ return 0;
+}
+
+const struct dcbnl_rtnl_ops dpaa2_eth_dcbnl_ops = {
+ .ieee_getpfc = dpaa2_eth_dcbnl_ieee_getpfc,
+ .ieee_setpfc = dpaa2_eth_dcbnl_ieee_setpfc,
+ .getdcbx = dpaa2_eth_dcbnl_getdcbx,
+ .setdcbx = dpaa2_eth_dcbnl_setdcbx,
+ .getcap = dpaa2_eth_dcbnl_getcap,
+};
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index 8cbbebefdcfd..1deccd377c5a 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -3844,6 +3844,15 @@ static int dpaa2_eth_probe(struct fsl_mc_device *dpni_dev)
if (err)
goto err_alloc_rings;
+#ifdef CONFIG_FSL_DPAA2_ETH_DCB
+ if (dpaa2_eth_has_pause_support(priv) && priv->vlan_cls_enabled) {
+ priv->dcbx_mode = DCB_CAP_DCBX_HOST | DCB_CAP_DCBX_VER_IEEE;
+ net_dev->dcbnl_ops = &dpaa2_eth_dcbnl_ops;
+ } else {
+ dev_dbg(dev, "PFC not supported\n");
+ }
+#endif
+
err = setup_irqs(dpni_dev);
if (err) {
netdev_warn(net_dev, "Failed to set link interrupt, fall back to polling\n");
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 02c0eea69a23..31b7b9b52da0 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -6,6 +6,7 @@
#ifndef __DPAA2_ETH_H
#define __DPAA2_ETH_H
+#include <linux/dcbnl.h>
#include <linux/netdevice.h>
#include <linux/if_vlan.h>
#include <linux/fsl/mc.h>
@@ -65,6 +66,17 @@
#define DPAA2_ETH_CG_TAILDROP_THRESH(priv) \
(1024 * dpaa2_eth_queue_count(priv) / dpaa2_eth_tc_count(priv))
+/* Congestion group notification threshold: when this many frames accumulate
+ * on the Rx queues belonging to the same TC, the MAC is instructed to send
+ * PFC frames for that TC.
+ * When number of pending frames drops below exit threshold transmission of
+ * PFC frames is stopped.
+ */
+#define DPAA2_ETH_CN_THRESH_ENTRY(priv) \
+ (DPAA2_ETH_CG_TAILDROP_THRESH(priv) / 2)
+#define DPAA2_ETH_CN_THRESH_EXIT(priv) \
+ (DPAA2_ETH_CN_THRESH_ENTRY(priv) * 3 / 4)
+
/* Maximum number of buffers that can be acquired/released through a single
* QBMan command
*/
@@ -436,6 +448,10 @@ struct dpaa2_eth_priv {
struct dpaa2_eth_cls_rule *cls_rules;
u8 rx_cls_enabled;
u8 vlan_cls_enabled;
+#ifdef CONFIG_FSL_DPAA2_ETH_DCB
+ u8 dcbx_mode;
+ struct ieee_pfc pfc;
+#endif
struct bpf_prog *xdp_prog;
#ifdef CONFIG_DEBUG_FS
struct dpaa2_debugfs dbg;
@@ -568,4 +584,6 @@ int dpaa2_eth_cls_key_size(u64 key);
int dpaa2_eth_cls_fld_off(int prot, int field);
void dpaa2_eth_cls_trim_rule(void *key_mem, u64 fields);
+extern const struct dcbnl_rtnl_ops dpaa2_eth_dcbnl_ops;
+
#endif /* __DPAA2_H */
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h b/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
index 0048e856f85e..fd069f67be9b 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
@@ -601,4 +601,29 @@ struct dpni_cmd_remove_qos_entry {
__le64 mask_iova;
};
+#define DPNI_DEST_TYPE_SHIFT 0
+#define DPNI_DEST_TYPE_SIZE 4
+#define DPNI_CONG_UNITS_SHIFT 4
+#define DPNI_CONG_UNITS_SIZE 2
+
+struct dpni_cmd_set_congestion_notification {
+ /* cmd word 0 */
+ u8 qtype;
+ u8 tc;
+ u8 pad[6];
+ /* cmd word 1 */
+ __le32 dest_id;
+ __le16 notification_mode;
+ u8 dest_priority;
+ /* from LSB: dest_type: 4 units:2 */
+ u8 type_units;
+ /* cmd word 2 */
+ __le64 message_iova;
+ /* cmd word 3 */
+ __le64 message_ctx;
+ /* cmd word 4 */
+ __le32 threshold_entry;
+ __le32 threshold_exit;
+};
+
#endif /* _FSL_DPNI_CMD_H */
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni.c b/drivers/net/ethernet/freescale/dpaa2/dpni.c
index 78fa325407ca..6b479ba66465 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni.c
@@ -1354,6 +1354,52 @@ int dpni_set_rx_tc_dist(struct fsl_mc_io *mc_io,
return mc_send_command(mc_io, &cmd);
}
+/**
+ * dpni_set_congestion_notification() - Set traffic class congestion
+ * notification configuration
+ * @mc_io: Pointer to MC portal's I/O object
+ * @cmd_flags: Command flags; one or more of 'MC_CMD_FLAG_'
+ * @token: Token of DPNI object
+ * @qtype: Type of queue - Rx, Tx and Tx confirm types are supported
+ * @tc_id: Traffic class selection (0-7)
+ * @cfg: Congestion notification configuration
+ *
+ * Return: '0' on Success; error code otherwise.
+ */
+int dpni_set_congestion_notification(
+ struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ enum dpni_queue_type qtype,
+ u8 tc_id,
+ const struct dpni_congestion_notification_cfg *cfg)
+{
+ struct dpni_cmd_set_congestion_notification *cmd_params;
+ struct fsl_mc_command cmd = { 0 };
+
+ /* prepare command */
+ cmd.header =
+ mc_encode_cmd_header(DPNI_CMDID_SET_CONGESTION_NOTIFICATION,
+ cmd_flags,
+ token);
+ cmd_params = (struct dpni_cmd_set_congestion_notification *)cmd.params;
+ cmd_params->qtype = qtype;
+ cmd_params->tc = tc_id;
+ cmd_params->dest_id = cpu_to_le32(cfg->dest_cfg.dest_id);
+ cmd_params->notification_mode = cpu_to_le16(cfg->notification_mode);
+ cmd_params->dest_priority = cfg->dest_cfg.priority;
+ dpni_set_field(cmd_params->type_units, DEST_TYPE,
+ cfg->dest_cfg.dest_type);
+ dpni_set_field(cmd_params->type_units, CONG_UNITS, cfg->units);
+ cmd_params->message_iova = cpu_to_le64(cfg->message_iova);
+ cmd_params->message_ctx = cpu_to_le64(cfg->message_ctx);
+ cmd_params->threshold_entry = cpu_to_le32(cfg->threshold_entry);
+ cmd_params->threshold_exit = cpu_to_le32(cfg->threshold_exit);
+
+ /* send command to mc*/
+ return mc_send_command(mc_io, &cmd);
+}
+
/**
* dpni_set_queue() - Set queue parameters
* @mc_io: Pointer to MC portal's I/O object
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni.h b/drivers/net/ethernet/freescale/dpaa2/dpni.h
index 8c7ac20bf1a7..e874d8084142 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni.h
@@ -513,6 +513,11 @@ int dpni_get_statistics(struct fsl_mc_io *mc_io,
*/
#define DPNI_LINK_OPT_ASYM_PAUSE 0x0000000000000008ULL
+/**
+ * Enable priority flow control pause frames
+ */
+#define DPNI_LINK_OPT_PFC_PAUSE 0x0000000000000010ULL
+
/**
* struct - Structure representing DPNI link configuration
* @rate: Rate
@@ -877,6 +882,62 @@ enum dpni_congestion_point {
DPNI_CP_GROUP,
};
+/**
+ * struct dpni_dest_cfg - Structure representing DPNI destination parameters
+ * @dest_type: Destination type
+ * @dest_id: Either DPIO ID or DPCON ID, depending on the destination type
+ * @priority: Priority selection within the DPIO or DPCON channel; valid
+ * values are 0-1 or 0-7, depending on the number of priorities
+ * in that channel; not relevant for 'DPNI_DEST_NONE' option
+ */
+struct dpni_dest_cfg {
+ enum dpni_dest dest_type;
+ int dest_id;
+ u8 priority;
+};
+
+/* DPNI congestion options */
+
+/**
+ * This congestion will trigger flow control or priority flow control.
+ * This will have effect only if flow control is enabled with
+ * dpni_set_link_cfg().
+ */
+#define DPNI_CONG_OPT_FLOW_CONTROL 0x00000040
+
+/**
+ * struct dpni_congestion_notification_cfg - congestion notification
+ * configuration
+ * @units: Units type
+ * @threshold_entry: Above this threshold we enter a congestion state.
+ * set it to '0' to disable it
+ * @threshold_exit: Below this threshold we exit the congestion state.
+ * @message_ctx: The context that will be part of the CSCN message
+ * @message_iova: I/O virtual address (must be in DMA-able memory),
+ * must be 16B aligned; valid only if 'DPNI_CONG_OPT_WRITE_MEM_<X>'
+ * is contained in 'options'
+ * @dest_cfg: CSCN can be send to either DPIO or DPCON WQ channel
+ * @notification_mode: Mask of available options; use 'DPNI_CONG_OPT_<X>' values
+ */
+
+struct dpni_congestion_notification_cfg {
+ enum dpni_congestion_unit units;
+ u32 threshold_entry;
+ u32 threshold_exit;
+ u64 message_ctx;
+ u64 message_iova;
+ struct dpni_dest_cfg dest_cfg;
+ u16 notification_mode;
+};
+
+int dpni_set_congestion_notification(
+ struct fsl_mc_io *mc_io,
+ u32 cmd_flags,
+ u16 token,
+ enum dpni_queue_type qtype,
+ u8 tc_id,
+ const struct dpni_congestion_notification_cfg *cfg);
+
/**
* struct dpni_taildrop - Structure representing the taildrop
* @enable: Indicates whether the taildrop is active or not.
--
2.17.1
^ permalink raw reply related
* [PATCH net-next v3 4/7] dpaa2-eth: Add congestion group taildrop
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Radulescu, Ioana Ciornei
In-Reply-To: <20200529174345.27537-1-ioana.ciornei@nxp.com>
From: Ioana Radulescu <ruxandra.radulescu@nxp.com>
The increase in number of ingress frame queues means we now risk
depleting the buffer pool before the FQ taildrop kicks in.
Congestion group taildrop allows us to control the number of frames that
can accumulate on a group of Rx frame queues belonging to the same
traffic class. This setting coexists with the frame queue based
taildrop: whichever limit gets hit first triggers the frame drop.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
Changes in v3:
- none
.../net/ethernet/freescale/dpaa2/dpaa2-eth.c | 35 ++++++++++++++-----
.../net/ethernet/freescale/dpaa2/dpaa2-eth.h | 13 +++++--
2 files changed, 38 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index c521f4afba8e..8cbbebefdcfd 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -1287,17 +1287,20 @@ static void disable_ch_napi(struct dpaa2_eth_priv *priv)
}
}
-static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv, bool enable)
+static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv,
+ bool tx_pause)
{
struct dpni_taildrop td = {0};
struct dpaa2_eth_fq *fq;
int i, err;
- if (priv->rx_td_enabled == enable)
+ td.enable = !tx_pause;
+ if (priv->rx_td_enabled == td.enable)
return;
- td.enable = enable;
- td.threshold = DPAA2_ETH_TAILDROP_THRESH;
+ /* FQ taildrop: threshold is in bytes, per frame queue */
+ td.threshold = DPAA2_ETH_FQ_TAILDROP_THRESH;
+ td.units = DPNI_CONGESTION_UNIT_BYTES;
for (i = 0; i < priv->num_fqs; i++) {
fq = &priv->fq[i];
@@ -1308,12 +1311,28 @@ static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv, bool enable)
fq->tc, fq->flowid, &td);
if (err) {
netdev_err(priv->net_dev,
- "dpni_set_taildrop() failed\n");
- break;
+ "dpni_set_taildrop(FQ) failed\n");
+ return;
+ }
+ }
+
+ /* Congestion group taildrop: threshold is in frames, per group
+ * of FQs belonging to the same traffic class
+ */
+ td.threshold = DPAA2_ETH_CG_TAILDROP_THRESH(priv);
+ td.units = DPNI_CONGESTION_UNIT_FRAMES;
+ for (i = 0; i < dpaa2_eth_tc_count(priv); i++) {
+ err = dpni_set_taildrop(priv->mc_io, 0, priv->mc_token,
+ DPNI_CP_GROUP, DPNI_QUEUE_RX,
+ i, 0, &td);
+ if (err) {
+ netdev_err(priv->net_dev,
+ "dpni_set_taildrop(CG) failed\n");
+ return;
}
}
- priv->rx_td_enabled = enable;
+ priv->rx_td_enabled = td.enable;
}
static int link_state_update(struct dpaa2_eth_priv *priv)
@@ -1334,7 +1353,7 @@ static int link_state_update(struct dpaa2_eth_priv *priv)
* only when pause frame generation is disabled.
*/
tx_pause = dpaa2_eth_tx_pause_enabled(state.options);
- dpaa2_eth_set_rx_taildrop(priv, !tx_pause);
+ dpaa2_eth_set_rx_taildrop(priv, tx_pause);
/* When we manage the MAC/PHY using phylink there is no need
* to manually update the netif_carrier.
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 6384f6a23349..184d5d83e497 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -40,7 +40,7 @@
* frames in the Rx queues (length of the current frame is not
* taken into account when making the taildrop decision)
*/
-#define DPAA2_ETH_TAILDROP_THRESH (64 * 1024)
+#define DPAA2_ETH_FQ_TAILDROP_THRESH (64 * 1024)
/* Maximum number of Tx confirmation frames to be processed
* in a single NAPI call
@@ -52,11 +52,20 @@
* how many 64B frames fit inside the taildrop threshold and add a margin
* to accommodate the buffer refill delay.
*/
-#define DPAA2_ETH_MAX_FRAMES_PER_QUEUE (DPAA2_ETH_TAILDROP_THRESH / 64)
+#define DPAA2_ETH_MAX_FRAMES_PER_QUEUE (DPAA2_ETH_FQ_TAILDROP_THRESH / 64)
#define DPAA2_ETH_NUM_BUFS (DPAA2_ETH_MAX_FRAMES_PER_QUEUE + 256)
#define DPAA2_ETH_REFILL_THRESH \
(DPAA2_ETH_NUM_BUFS - DPAA2_ETH_BUFS_PER_CMD)
+/* Congestion group taildrop threshold: number of frames allowed to accumulate
+ * at any moment in a group of Rx queues belonging to the same traffic class.
+ * Choose value such that we don't risk depleting the buffer pool before the
+ * taildrop kicks in
+ */
+#define DPAA2_ETH_CG_TAILDROP_THRESH(priv) \
+ (DPAA2_ETH_MAX_FRAMES_PER_QUEUE * dpaa2_eth_queue_count(priv) / \
+ dpaa2_eth_tc_count(priv))
+
/* Maximum number of buffers that can be acquired/released through a single
* QBMan command
*/
--
2.17.1
^ permalink raw reply related
* [PATCH net-next v3 5/7] dpaa2-eth: Update FQ taildrop threshold and buffer pool count
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Radulescu, Ioana Ciornei
In-Reply-To: <20200529174345.27537-1-ioana.ciornei@nxp.com>
From: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Now that we have congestion group taildrop configured at all
times, we can afford to increase the frame queue taildrop
threshold; this will ensure a better response when receiving
bursts of large-sized frames.
Also decouple the buffer pool count from the Rx FQ taildrop
threshold, as above change would increase it too much. Instead,
keep the old count as a hardcoded value.
With the new limits, we try to ensure that:
* we allow enough leeway for large frame bursts (by buffering
enough of them in queues to avoid heavy dropping in case of
bursty traffic, but when overall ingress bandwidth is manageable)
* allow pending frames to be evenly spread between ingress FQs,
regardless of frame size
* avoid dropping frames due to the buffer pool being empty; this
is not a bad behaviour per se, but system overall response is
more linear and predictable when frames are dropped at frame
queue/group level.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
Changes in v3:
- none
.../net/ethernet/freescale/dpaa2/dpaa2-eth.h | 23 +++++++++----------
1 file changed, 11 insertions(+), 12 deletions(-)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 184d5d83e497..02c0eea69a23 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -36,24 +36,24 @@
/* Convert L3 MTU to L2 MFL */
#define DPAA2_ETH_L2_MAX_FRM(mtu) ((mtu) + VLAN_ETH_HLEN)
-/* Set the taildrop threshold (in bytes) to allow the enqueue of several jumbo
- * frames in the Rx queues (length of the current frame is not
- * taken into account when making the taildrop decision)
+/* Set the taildrop threshold (in bytes) to allow the enqueue of a large
+ * enough number of jumbo frames in the Rx queues (length of the current
+ * frame is not taken into account when making the taildrop decision)
*/
-#define DPAA2_ETH_FQ_TAILDROP_THRESH (64 * 1024)
+#define DPAA2_ETH_FQ_TAILDROP_THRESH (1024 * 1024)
/* Maximum number of Tx confirmation frames to be processed
* in a single NAPI call
*/
#define DPAA2_ETH_TXCONF_PER_NAPI 256
-/* Buffer quota per queue. Must be large enough such that for minimum sized
- * frames taildrop kicks in before the bpool gets depleted, so we compute
- * how many 64B frames fit inside the taildrop threshold and add a margin
- * to accommodate the buffer refill delay.
+/* Buffer qouta per channel. We want to keep in check number of ingress frames
+ * in flight: for small sized frames, congestion group taildrop may kick in
+ * first; for large sizes, Rx FQ taildrop threshold will ensure only a
+ * reasonable number of frames will be pending at any given time.
+ * Ingress frame drop due to buffer pool depletion should be a corner case only
*/
-#define DPAA2_ETH_MAX_FRAMES_PER_QUEUE (DPAA2_ETH_FQ_TAILDROP_THRESH / 64)
-#define DPAA2_ETH_NUM_BUFS (DPAA2_ETH_MAX_FRAMES_PER_QUEUE + 256)
+#define DPAA2_ETH_NUM_BUFS 1280
#define DPAA2_ETH_REFILL_THRESH \
(DPAA2_ETH_NUM_BUFS - DPAA2_ETH_BUFS_PER_CMD)
@@ -63,8 +63,7 @@
* taildrop kicks in
*/
#define DPAA2_ETH_CG_TAILDROP_THRESH(priv) \
- (DPAA2_ETH_MAX_FRAMES_PER_QUEUE * dpaa2_eth_queue_count(priv) / \
- dpaa2_eth_tc_count(priv))
+ (1024 * dpaa2_eth_queue_count(priv) / dpaa2_eth_tc_count(priv))
/* Maximum number of buffers that can be acquired/released through a single
* QBMan command
--
2.17.1
^ permalink raw reply related
* [PATCH net-next v3 0/7] dpaa2-eth: add PFC support
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Ciornei
This patch set adds support for Priority Flow Control in DPAA2 Ethernet
devices.
The first patch make the necessary changes so that multiple
traffic classes are configured. The dequeue priority
of the maximum 8 traffic classes is configured to be equal.
The second patch adds a static distribution to said traffic
classes based on the VLAN PCP field. In the future, this could be
extended through the .setapp() DCB callback for dynamic configuration.
Also, add support for the congestion group taildrop mechanism that
allows us to control the number of frames that can accumulate on a group
of Rx frame queues belonging to the same traffic class.
The basic subset of the DCB ops is implemented so that the user can
query the number of PFC capable traffic classes, their state and
reconfigure them if necessary.
Changes in v3:
- add patches 6-7 which add the PFC functionality
- patch 2/7: revert to explicitly cast mask to u16 * to not get into
sparse warnings
Ioana Ciornei (2):
dpaa2-eth: Add PFC support through DCB ops
dpaa2-eth: Keep congestion group taildrop enabled when PFC on
Ioana Radulescu (5):
dpaa2-eth: Add support for Rx traffic classes
dpaa2-eth: Distribute ingress frames based on VLAN prio
dpaa2-eth: Add helper functions
dpaa2-eth: Add congestion group taildrop
dpaa2-eth: Update FQ taildrop threshold and buffer pool count
drivers/net/ethernet/freescale/dpaa2/Kconfig | 10 +
drivers/net/ethernet/freescale/dpaa2/Makefile | 1 +
.../ethernet/freescale/dpaa2/dpaa2-eth-dcb.c | 150 +++++++++++
.../freescale/dpaa2/dpaa2-eth-debugfs.c | 7 +-
.../net/ethernet/freescale/dpaa2/dpaa2-eth.c | 250 +++++++++++++++---
.../net/ethernet/freescale/dpaa2/dpaa2-eth.h | 69 ++++-
.../ethernet/freescale/dpaa2/dpaa2-ethtool.c | 24 +-
.../net/ethernet/freescale/dpaa2/dpni-cmd.h | 59 +++++
drivers/net/ethernet/freescale/dpaa2/dpni.c | 177 +++++++++++++
drivers/net/ethernet/freescale/dpaa2/dpni.h | 97 +++++++
10 files changed, 787 insertions(+), 57 deletions(-)
create mode 100644 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-dcb.c
--
2.17.1
^ permalink raw reply
* [PATCH net-next v3 3/7] dpaa2-eth: Add helper functions
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Radulescu, Ioana Ciornei
In-Reply-To: <20200529174345.27537-1-ioana.ciornei@nxp.com>
From: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Add convenient helper functions that determines whether Rx/Tx pause
frames are enabled based on link state flags received from firmware.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
Changes in v3:
- none
drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c | 3 +--
drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h | 11 +++++++++++
drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c | 5 ++---
3 files changed, 14 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index b3722ad60fae..c521f4afba8e 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -1333,8 +1333,7 @@ static int link_state_update(struct dpaa2_eth_priv *priv)
* Rx FQ taildrop configuration as well. We configure taildrop
* only when pause frame generation is disabled.
*/
- tx_pause = !!(state.options & DPNI_LINK_OPT_PAUSE) ^
- !!(state.options & DPNI_LINK_OPT_ASYM_PAUSE);
+ tx_pause = dpaa2_eth_tx_pause_enabled(state.options);
dpaa2_eth_set_rx_taildrop(priv, !tx_pause);
/* When we manage the MAC/PHY using phylink there is no need
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 7856f69bcf36..6384f6a23349 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -510,6 +510,17 @@ enum dpaa2_eth_rx_dist {
(dpaa2_eth_cmp_dpni_ver((priv), DPNI_PAUSE_VER_MAJOR, \
DPNI_PAUSE_VER_MINOR) >= 0)
+static inline bool dpaa2_eth_tx_pause_enabled(u64 link_options)
+{
+ return !!(link_options & DPNI_LINK_OPT_PAUSE) ^
+ !!(link_options & DPNI_LINK_OPT_ASYM_PAUSE);
+}
+
+static inline bool dpaa2_eth_rx_pause_enabled(u64 link_options)
+{
+ return !!(link_options & DPNI_LINK_OPT_PAUSE);
+}
+
static inline
unsigned int dpaa2_eth_needed_headroom(struct dpaa2_eth_priv *priv,
struct sk_buff *skb)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
index 8bf169783bea..e88269fe3de7 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
@@ -130,9 +130,8 @@ static void dpaa2_eth_get_pauseparam(struct net_device *net_dev,
return;
}
- pause->rx_pause = !!(link_options & DPNI_LINK_OPT_PAUSE);
- pause->tx_pause = pause->rx_pause ^
- !!(link_options & DPNI_LINK_OPT_ASYM_PAUSE);
+ pause->rx_pause = dpaa2_eth_rx_pause_enabled(link_options);
+ pause->tx_pause = dpaa2_eth_tx_pause_enabled(link_options);
pause->autoneg = AUTONEG_DISABLE;
}
--
2.17.1
^ permalink raw reply related
* [PATCH net-next v3 1/7] dpaa2-eth: Add support for Rx traffic classes
From: Ioana Ciornei @ 2020-05-29 17:43 UTC (permalink / raw)
To: netdev; +Cc: kuba, davem, Ioana Radulescu, Ioana Ciornei
In-Reply-To: <20200529174345.27537-1-ioana.ciornei@nxp.com>
From: Ioana Radulescu <ruxandra.radulescu@nxp.com>
The firmware reserves for each DPNI a number of RX frame queues
equal to the number of configured flows x number of configured
traffic classes.
Current driver configuration directs all incoming traffic to
FQs corresponding to TC0, leaving all other priority levels unused.
Start adding support for multiple ingress traffic classes, by
configuring the FQs associated with all priority levels, not just
TC0. All settings that are per-TC, such as those related to
hashing and flow steering, are also updated.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
---
Changes in v3:
- none
.../freescale/dpaa2/dpaa2-eth-debugfs.c | 7 +-
.../net/ethernet/freescale/dpaa2/dpaa2-eth.c | 70 +++++++++++++------
.../net/ethernet/freescale/dpaa2/dpaa2-eth.h | 4 +-
.../ethernet/freescale/dpaa2/dpaa2-ethtool.c | 19 +++--
4 files changed, 68 insertions(+), 32 deletions(-)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c
index 0a31e4268dfb..c453a23045c1 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c
@@ -81,8 +81,8 @@ static int dpaa2_dbg_fqs_show(struct seq_file *file, void *offset)
int i, err;
seq_printf(file, "FQ stats for %s:\n", priv->net_dev->name);
- seq_printf(file, "%s%16s%16s%16s%16s\n",
- "VFQID", "CPU", "Type", "Frames", "Pending frames");
+ seq_printf(file, "%s%16s%16s%16s%16s%16s\n",
+ "VFQID", "CPU", "TC", "Type", "Frames", "Pending frames");
for (i = 0; i < priv->num_fqs; i++) {
fq = &priv->fq[i];
@@ -90,9 +90,10 @@ static int dpaa2_dbg_fqs_show(struct seq_file *file, void *offset)
if (err)
fcnt = 0;
- seq_printf(file, "%5d%16d%16s%16llu%16u\n",
+ seq_printf(file, "%5d%16d%16d%16s%16llu%16u\n",
fq->fqid,
fq->target_cpu,
+ fq->tc,
fq_type_to_str(fq),
fq->stats.frames,
fcnt);
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index fe3806d54630..01263e247d39 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -1290,6 +1290,7 @@ static void disable_ch_napi(struct dpaa2_eth_priv *priv)
static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv, bool enable)
{
struct dpni_taildrop td = {0};
+ struct dpaa2_eth_fq *fq;
int i, err;
if (priv->rx_td_enabled == enable)
@@ -1299,11 +1300,12 @@ static void dpaa2_eth_set_rx_taildrop(struct dpaa2_eth_priv *priv, bool enable)
td.threshold = DPAA2_ETH_TAILDROP_THRESH;
for (i = 0; i < priv->num_fqs; i++) {
- if (priv->fq[i].type != DPAA2_RX_FQ)
+ fq = &priv->fq[i];
+ if (fq->type != DPAA2_RX_FQ)
continue;
err = dpni_set_taildrop(priv->mc_io, 0, priv->mc_token,
- DPNI_CP_QUEUE, DPNI_QUEUE_RX, 0,
- priv->fq[i].flowid, &td);
+ DPNI_CP_QUEUE, DPNI_QUEUE_RX,
+ fq->tc, fq->flowid, &td);
if (err) {
netdev_err(priv->net_dev,
"dpni_set_taildrop() failed\n");
@@ -2407,7 +2409,7 @@ static void set_fq_affinity(struct dpaa2_eth_priv *priv)
static void setup_fqs(struct dpaa2_eth_priv *priv)
{
- int i;
+ int i, j;
/* We have one TxConf FQ per Tx flow.
* The number of Tx and Rx queues is the same.
@@ -2419,10 +2421,13 @@ static void setup_fqs(struct dpaa2_eth_priv *priv)
priv->fq[priv->num_fqs++].flowid = (u16)i;
}
- for (i = 0; i < dpaa2_eth_queue_count(priv); i++) {
- priv->fq[priv->num_fqs].type = DPAA2_RX_FQ;
- priv->fq[priv->num_fqs].consume = dpaa2_eth_rx;
- priv->fq[priv->num_fqs++].flowid = (u16)i;
+ for (j = 0; j < dpaa2_eth_tc_count(priv); j++) {
+ for (i = 0; i < dpaa2_eth_queue_count(priv); i++) {
+ priv->fq[priv->num_fqs].type = DPAA2_RX_FQ;
+ priv->fq[priv->num_fqs].consume = dpaa2_eth_rx;
+ priv->fq[priv->num_fqs].tc = (u8)j;
+ priv->fq[priv->num_fqs++].flowid = (u16)i;
+ }
}
/* For each FQ, decide on which core to process incoming frames */
@@ -2789,7 +2794,7 @@ static int setup_rx_flow(struct dpaa2_eth_priv *priv,
int err;
err = dpni_get_queue(priv->mc_io, 0, priv->mc_token,
- DPNI_QUEUE_RX, 0, fq->flowid, &queue, &qid);
+ DPNI_QUEUE_RX, fq->tc, fq->flowid, &queue, &qid);
if (err) {
dev_err(dev, "dpni_get_queue(RX) failed\n");
return err;
@@ -2802,7 +2807,7 @@ static int setup_rx_flow(struct dpaa2_eth_priv *priv,
queue.destination.priority = 1;
queue.user_context = (u64)(uintptr_t)fq;
err = dpni_set_queue(priv->mc_io, 0, priv->mc_token,
- DPNI_QUEUE_RX, 0, fq->flowid,
+ DPNI_QUEUE_RX, fq->tc, fq->flowid,
DPNI_QUEUE_OPT_USER_CTX | DPNI_QUEUE_OPT_DEST,
&queue);
if (err) {
@@ -2811,6 +2816,10 @@ static int setup_rx_flow(struct dpaa2_eth_priv *priv,
}
/* xdp_rxq setup */
+ /* only once for each channel */
+ if (fq->tc > 0)
+ return 0;
+
err = xdp_rxq_info_reg(&fq->channel->xdp_rxq, priv->net_dev,
fq->flowid);
if (err) {
@@ -2948,7 +2957,7 @@ static int config_legacy_hash_key(struct dpaa2_eth_priv *priv, dma_addr_t key)
{
struct device *dev = priv->net_dev->dev.parent;
struct dpni_rx_tc_dist_cfg dist_cfg;
- int err;
+ int i, err = 0;
memset(&dist_cfg, 0, sizeof(dist_cfg));
@@ -2956,9 +2965,14 @@ static int config_legacy_hash_key(struct dpaa2_eth_priv *priv, dma_addr_t key)
dist_cfg.dist_size = dpaa2_eth_queue_count(priv);
dist_cfg.dist_mode = DPNI_DIST_MODE_HASH;
- err = dpni_set_rx_tc_dist(priv->mc_io, 0, priv->mc_token, 0, &dist_cfg);
- if (err)
- dev_err(dev, "dpni_set_rx_tc_dist failed\n");
+ for (i = 0; i < dpaa2_eth_tc_count(priv); i++) {
+ err = dpni_set_rx_tc_dist(priv->mc_io, 0, priv->mc_token,
+ i, &dist_cfg);
+ if (err) {
+ dev_err(dev, "dpni_set_rx_tc_dist failed\n");
+ break;
+ }
+ }
return err;
}
@@ -2968,7 +2982,7 @@ static int config_hash_key(struct dpaa2_eth_priv *priv, dma_addr_t key)
{
struct device *dev = priv->net_dev->dev.parent;
struct dpni_rx_dist_cfg dist_cfg;
- int err;
+ int i, err = 0;
memset(&dist_cfg, 0, sizeof(dist_cfg));
@@ -2976,9 +2990,15 @@ static int config_hash_key(struct dpaa2_eth_priv *priv, dma_addr_t key)
dist_cfg.dist_size = dpaa2_eth_queue_count(priv);
dist_cfg.enable = 1;
- err = dpni_set_rx_hash_dist(priv->mc_io, 0, priv->mc_token, &dist_cfg);
- if (err)
- dev_err(dev, "dpni_set_rx_hash_dist failed\n");
+ for (i = 0; i < dpaa2_eth_tc_count(priv); i++) {
+ dist_cfg.tc = i;
+ err = dpni_set_rx_hash_dist(priv->mc_io, 0, priv->mc_token,
+ &dist_cfg);
+ if (err) {
+ dev_err(dev, "dpni_set_rx_hash_dist failed\n");
+ break;
+ }
+ }
return err;
}
@@ -2988,7 +3008,7 @@ static int config_cls_key(struct dpaa2_eth_priv *priv, dma_addr_t key)
{
struct device *dev = priv->net_dev->dev.parent;
struct dpni_rx_dist_cfg dist_cfg;
- int err;
+ int i, err = 0;
memset(&dist_cfg, 0, sizeof(dist_cfg));
@@ -2996,9 +3016,15 @@ static int config_cls_key(struct dpaa2_eth_priv *priv, dma_addr_t key)
dist_cfg.dist_size = dpaa2_eth_queue_count(priv);
dist_cfg.enable = 1;
- err = dpni_set_rx_fs_dist(priv->mc_io, 0, priv->mc_token, &dist_cfg);
- if (err)
- dev_err(dev, "dpni_set_rx_fs_dist failed\n");
+ for (i = 0; i < dpaa2_eth_tc_count(priv); i++) {
+ dist_cfg.tc = i;
+ err = dpni_set_rx_fs_dist(priv->mc_io, 0, priv->mc_token,
+ &dist_cfg);
+ if (err) {
+ dev_err(dev, "dpni_set_rx_fs_dist failed\n");
+ break;
+ }
+ }
return err;
}
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 0581fbf1f98c..580ad5fd7bd8 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -294,7 +294,9 @@ struct dpaa2_eth_ch_stats {
/* Maximum number of queues associated with a DPNI */
#define DPAA2_ETH_MAX_TCS 8
-#define DPAA2_ETH_MAX_RX_QUEUES 16
+#define DPAA2_ETH_MAX_RX_QUEUES_PER_TC 16
+#define DPAA2_ETH_MAX_RX_QUEUES \
+ (DPAA2_ETH_MAX_RX_QUEUES_PER_TC * DPAA2_ETH_MAX_TCS)
#define DPAA2_ETH_MAX_TX_QUEUES 16
#define DPAA2_ETH_MAX_QUEUES (DPAA2_ETH_MAX_RX_QUEUES + \
DPAA2_ETH_MAX_TX_QUEUES)
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
index 049afd1d6252..8bf169783bea 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
@@ -547,7 +547,7 @@ static int do_cls_rule(struct net_device *net_dev,
dma_addr_t key_iova;
u64 fields = 0;
void *key_buf;
- int err;
+ int i, err;
if (fs->ring_cookie != RX_CLS_FLOW_DISC &&
fs->ring_cookie >= dpaa2_eth_queue_count(priv))
@@ -607,11 +607,18 @@ static int do_cls_rule(struct net_device *net_dev,
fs_act.options |= DPNI_FS_OPT_DISCARD;
else
fs_act.flow_id = fs->ring_cookie;
- err = dpni_add_fs_entry(priv->mc_io, 0, priv->mc_token, 0,
- fs->location, &rule_cfg, &fs_act);
- } else {
- err = dpni_remove_fs_entry(priv->mc_io, 0, priv->mc_token, 0,
- &rule_cfg);
+ }
+ for (i = 0; i < dpaa2_eth_tc_count(priv); i++) {
+ if (add)
+ err = dpni_add_fs_entry(priv->mc_io, 0, priv->mc_token,
+ i, fs->location, &rule_cfg,
+ &fs_act);
+ else
+ err = dpni_remove_fs_entry(priv->mc_io, 0,
+ priv->mc_token, i,
+ &rule_cfg);
+ if (err)
+ break;
}
dma_unmap_single(dev, key_iova, rule_cfg.key_size * 2, DMA_TO_DEVICE);
--
2.17.1
^ permalink raw reply related
* Re: Question: livepatch failed for new fork() task stack unreliable
From: Josh Poimboeuf @ 2020-05-29 17:44 UTC (permalink / raw)
To: Wang ShaoBo
Cc: huawei.libin, xiexiuqi, cj.chengjian, mingo, x86, linux-kernel,
live-patching, mbenes, devel, viro, esyr
In-Reply-To: <20200529101059.39885-1-bobo.shaobowang@huawei.com>
On Fri, May 29, 2020 at 06:10:59PM +0800, Wang ShaoBo wrote:
> Stack unreliable error is reported by stack_trace_save_tsk_reliable() when trying
> to insmod a hot patch for module modification, this results in frequent failures
> sometimes. We found this 'unreliable' stack is from task just fork.
For livepatch, this shouldn't actually be a failure. The patch will
just stay in the transition state until after the fork has completed.
Which should happen in a reasonable amount of time, right?
> 1) The task was not actually scheduled to excute, at this time UNWIND_HINT_EMPTY in
> ret_from_fork() has not reset unwind_hint, it's sp_reg and end field remain default value
> and end up throwing an error in unwind_next_frame() when called by arch_stack_walk_reliable();
Yes, this seems to be true for forked-but-not-yet-scheduled tasks.
I can look at fixing that. I have some ORC cleanups in progress which
are related to UNWIND_HINT_EMPTY and the end of the stack. I can add
this issue to the list of improvements.
> 2) The task has been scheduled but UNWIND_HINT_REGS not finished, at this time
> arch_stack_walk_reliable() terminates it's backtracing loop for pt_regs unknown
> and return -EINVAL because it's a user task.
Hm, do you see this problem with upstream? It seems like it should
work. arch_stack_walk_reliable() has this:
/* Success path for user tasks */
if (user_mode(regs))
return 0;
Where exactly is the error coming from?
--
Josh
^ permalink raw reply
* [superna9999:amlogic/v5.9/amlogic-fbc 3/6] include/linux/compiler.h:350:38: error: call to '__compiletime_assert_178' declared with attribute error: FIELD_PREP: value too large for the field
From: kbuild test robot @ 2020-05-29 17:44 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6067 bytes --]
tree: https://github.com/superna9999/linux amlogic/v5.9/amlogic-fbc
head: d6030aeb4da732a1667aceca86f9c820282f0ca0
commit: 78a5397a5164e673015170eb0b5d78db4153a030 [3/6] drm/meson: overlay: setup overlay for Amlogic FBC
config: arm-allyesconfig (attached as .config)
compiler: arm-linux-gnueabi-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
git checkout 78a5397a5164e673015170eb0b5d78db4153a030
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=arm
If you fix the issue, kindly add following tag as appropriate
Reported-by: kbuild test robot <lkp@intel.com>
All error/warnings (new ones prefixed by >>, old ones prefixed by <<):
In file included from include/linux/build_bug.h:5,
from include/linux/bitfield.h:10,
from drivers/gpu/drm/meson/meson_overlay.c:9:
drivers/gpu/drm/meson/meson_overlay.c: In function 'meson_overlay_setup_scaler_params':
>> include/linux/compiler.h:350:38: error: call to '__compiletime_assert_178' declared with attribute error: FIELD_PREP: value too large for the field
350 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
| ^
include/linux/compiler.h:331:4: note: in definition of macro '__compiletime_assert'
331 | prefix ## suffix(); | ^~~~~~
include/linux/compiler.h:350:2: note: in expansion of macro '_compiletime_assert'
350 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
| ^~~~~~~~~~~~~~~~~~~
include/linux/build_bug.h:39:37: note: in expansion of macro 'compiletime_assert'
39 | #define BUILD_BUG_ON_MSG(cond, msg) compiletime_assert(!(cond), msg)
| ^~~~~~~~~~~~~~~~~~
include/linux/bitfield.h:49:3: note: in expansion of macro 'BUILD_BUG_ON_MSG'
49 | BUILD_BUG_ON_MSG(__builtin_constant_p(_val) ? | ^~~~~~~~~~~~~~~~
include/linux/bitfield.h:94:3: note: in expansion of macro '__BF_FIELD_CHECK'
94 | __BF_FIELD_CHECK(_mask, 0ULL, _val, "FIELD_PREP: "); | ^~~~~~~~~~~~~~~~
drivers/gpu/drm/meson/meson_overlay.c:148:37: note: in expansion of macro 'FIELD_PREP'
148 | #define AFBC_DEC_PIXEL_BGN_H(value) FIELD_PREP(GENMASK(28, 16), value)
| ^~~~~~~~~~
>> drivers/gpu/drm/meson/meson_overlay.c:414:4: note: in expansion of macro 'AFBC_DEC_PIXEL_BGN_H'
414 | AFBC_DEC_PIXEL_BGN_H(hd_start_lines - afbc_left) |
| ^~~~~~~~~~~~~~~~~~~~
include/linux/compiler.h:350:38: error: call to '__compiletime_assert_238' declared with attribute error: FIELD_PREP: value too large for the field
350 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
| ^
include/linux/compiler.h:331:4: note: in definition of macro '__compiletime_assert'
331 | prefix ## suffix(); | ^~~~~~
include/linux/compiler.h:350:2: note: in expansion of macro '_compiletime_assert'
350 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
| ^~~~~~~~~~~~~~~~~~~
include/linux/build_bug.h:39:37: note: in expansion of macro 'compiletime_assert'
39 | #define BUILD_BUG_ON_MSG(cond, msg) compiletime_assert(!(cond), msg)
| ^~~~~~~~~~~~~~~~~~
include/linux/bitfield.h:49:3: note: in expansion of macro 'BUILD_BUG_ON_MSG'
49 | BUILD_BUG_ON_MSG(__builtin_constant_p(_val) ? | ^~~~~~~~~~~~~~~~
include/linux/bitfield.h:94:3: note: in expansion of macro '__BF_FIELD_CHECK'
94 | __BF_FIELD_CHECK(_mask, 0ULL, _val, "FIELD_PREP: "); | ^~~~~~~~~~~~~~~~
drivers/gpu/drm/meson/meson_overlay.c:62:28: note: in expansion of macro 'FIELD_PREP'
62 | #define VD_H_START(value) FIELD_PREP(GENMASK(27, 16), value)
| ^~~~~~~~~~
drivers/gpu/drm/meson/meson_overlay.c:435:40: note: in expansion of macro 'VD_H_START'
435 | priv->viu.vpp_blend_vd2_h_start_end = VD_H_START(hd_start_lines) |
| ^~~~~~~~~~
vim +/__compiletime_assert_178 +350 include/linux/compiler.h
9a8ab1c39970a49 Daniel Santos 2013-02-21 336
9a8ab1c39970a49 Daniel Santos 2013-02-21 337 #define _compiletime_assert(condition, msg, prefix, suffix) \
9a8ab1c39970a49 Daniel Santos 2013-02-21 338 __compiletime_assert(condition, msg, prefix, suffix)
9a8ab1c39970a49 Daniel Santos 2013-02-21 339
9a8ab1c39970a49 Daniel Santos 2013-02-21 340 /**
9a8ab1c39970a49 Daniel Santos 2013-02-21 341 * compiletime_assert - break build and emit msg if condition is false
9a8ab1c39970a49 Daniel Santos 2013-02-21 342 * @condition: a compile-time constant condition to check
9a8ab1c39970a49 Daniel Santos 2013-02-21 343 * @msg: a message to emit if condition is false
9a8ab1c39970a49 Daniel Santos 2013-02-21 344 *
9a8ab1c39970a49 Daniel Santos 2013-02-21 345 * In tradition of POSIX assert, this macro will break the build if the
9a8ab1c39970a49 Daniel Santos 2013-02-21 346 * supplied condition is *false*, emitting the supplied error message if the
9a8ab1c39970a49 Daniel Santos 2013-02-21 347 * compiler has support to do so.
9a8ab1c39970a49 Daniel Santos 2013-02-21 348 */
9a8ab1c39970a49 Daniel Santos 2013-02-21 349 #define compiletime_assert(condition, msg) \
af9c5d2e3b35585 Vegard Nossum 2020-04-06 @350 _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
9a8ab1c39970a49 Daniel Santos 2013-02-21 351
:::::: The code at line 350 was first introduced by commit
:::::: af9c5d2e3b355854ff0e4acfbfbfadcd5198a349 compiler.h: fix error in BUILD_BUG_ON() reporting
:::::: TO: Vegard Nossum <vegard.nossum@oracle.com>
:::::: CC: Linus Torvalds <torvalds@linux-foundation.org>
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 74061 bytes --]
^ permalink raw reply
* Re: [PATCH net-next v2] hinic: add set_channels ethtool_ops support
From: Jakub Kicinski @ 2020-05-29 17:44 UTC (permalink / raw)
To: Luo bin
Cc: davem, linux-kernel, netdev, luoxianjun, yin.yinshi,
cloud.wangxiaoyun
In-Reply-To: <20200528183633.6689-1-luobin9@huawei.com>
On Thu, 28 May 2020 18:36:33 +0000 Luo bin wrote:
> add support to change TX/RX queue number with ethtool -L
>
> Signed-off-by: Luo bin <luobin9@huawei.com>
Luo bin, your patches continue to come with Date: header being in the
past. Also suspiciously no time zone offset. Can you address this?
> +static int hinic_set_channels(struct net_device *netdev,
> + struct ethtool_channels *channels)
> +{
> + struct hinic_dev *nic_dev = netdev_priv(netdev);
> + unsigned int count = channels->combined_count;
> + int err;
> +
> + if (!count) {
> + netif_err(nic_dev, drv, netdev,
> + "Unsupported combined_count: 0\n");
> + return -EINVAL;
> + }
This check has been added to the core since the last version of you
patch:
/* ensure there is at least one RX and one TX channel */
if (!channels.combined_count &&
(!channels.rx_count || !channels.tx_count))
return -EINVAL;
> + netif_info(nic_dev, drv, netdev, "Set max combined queue number from %d to %d\n",
> + hinic_hwdev_num_qps(nic_dev->hwdev), count);
> +
> + if (netif_running(netdev)) {
> + netif_info(nic_dev, drv, netdev, "Restarting netdev\n");
> + hinic_close(netdev);
> +
> + nic_dev->hwdev->nic_cap.num_qps = count;
> +
> + err = hinic_open(netdev);
> + if (err) {
> + netif_err(nic_dev, drv, netdev,
> + "Failed to open netdev\n");
> + return -EFAULT;
> + }
> + } else {
> + nic_dev->hwdev->nic_cap.num_qps = count;
> + }
> +
> + return 0;
> }
^ permalink raw reply
* Re: [PATCH 06/30] KVM: SVM: always update CR3 in VMCB
From: Krish Sadhukhan @ 2020-05-29 17:41 UTC (permalink / raw)
To: Paolo Bonzini, linux-kernel, kvm
In-Reply-To: <20200529153934.11694-7-pbonzini@redhat.com>
On 5/29/20 8:39 AM, Paolo Bonzini wrote:
> svm_load_mmu_pgd is delaying the write of GUEST_CR3 to prepare_vmcs02
Did you mean to say enter_svm_guest_mode here ?
> as
> an optimization, but this is only correct before the nested vmentry.
> If userspace is modifying CR3 with KVM_SET_SREGS after the VM has
> already been put in guest mode, the value of CR3 will not be updated.
> Remove the optimization, which almost never triggers anyway.
> This was was added in commit 689f3bf21628 ("KVM: x86: unify callbacks
> to load paging root", 2020-03-16) just to keep the two vendor-specific
> modules closer, but we'll fix VMX too.
>
> Fixes: 689f3bf21628 ("KVM: x86: unify callbacks to load paging root")
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
> arch/x86/kvm/svm/nested.c | 6 +-----
> arch/x86/kvm/svm/svm.c | 16 +++++-----------
> 2 files changed, 6 insertions(+), 16 deletions(-)
>
> diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c
> index dcac4c3510ab..8756c9f463fd 100644
> --- a/arch/x86/kvm/svm/nested.c
> +++ b/arch/x86/kvm/svm/nested.c
> @@ -256,11 +256,7 @@ void enter_svm_guest_mode(struct vcpu_svm *svm, u64 vmcb_gpa,
> svm_set_efer(&svm->vcpu, nested_vmcb->save.efer);
> svm_set_cr0(&svm->vcpu, nested_vmcb->save.cr0);
> svm_set_cr4(&svm->vcpu, nested_vmcb->save.cr4);
> - if (npt_enabled) {
> - svm->vmcb->save.cr3 = nested_vmcb->save.cr3;
> - svm->vcpu.arch.cr3 = nested_vmcb->save.cr3;
> - } else
> - (void)kvm_set_cr3(&svm->vcpu, nested_vmcb->save.cr3);
> + (void)kvm_set_cr3(&svm->vcpu, nested_vmcb->save.cr3);
>
> /* Guest paging mode is active - reset mmu */
> kvm_mmu_reset_context(&svm->vcpu);
> diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
> index 545f63ebc720..feb96a410f2d 100644
> --- a/arch/x86/kvm/svm/svm.c
> +++ b/arch/x86/kvm/svm/svm.c
> @@ -3447,7 +3447,6 @@ static fastpath_t svm_vcpu_run(struct kvm_vcpu *vcpu)
> static void svm_load_mmu_pgd(struct kvm_vcpu *vcpu, unsigned long root)
> {
> struct vcpu_svm *svm = to_svm(vcpu);
> - bool update_guest_cr3 = true;
> unsigned long cr3;
>
> cr3 = __sme_set(root);
> @@ -3456,18 +3455,13 @@ static void svm_load_mmu_pgd(struct kvm_vcpu *vcpu, unsigned long root)
> mark_dirty(svm->vmcb, VMCB_NPT);
>
> /* Loading L2's CR3 is handled by enter_svm_guest_mode. */
> - if (is_guest_mode(vcpu))
> - update_guest_cr3 = false;
> - else if (test_bit(VCPU_EXREG_CR3, (ulong *)&vcpu->arch.regs_avail))
> - cr3 = vcpu->arch.cr3;
> - else /* CR3 is already up-to-date. */
> - update_guest_cr3 = false;
> + if (!test_bit(VCPU_EXREG_CR3, (ulong *)&vcpu->arch.regs_avail))
> + return;
> + cr3 = vcpu->arch.cr3;
> }
>
> - if (update_guest_cr3) {
> - svm->vmcb->save.cr3 = cr3;
> - mark_dirty(svm->vmcb, VMCB_CR);
> - }
> + svm->vmcb->save.cr3 = cr3;
> + mark_dirty(svm->vmcb, VMCB_CR);
> }
>
> static int is_disabled(void)
^ permalink raw reply
* Re: [PATCH] spi: bcm2835: Enable shared interrupt support
From: Lukas Wunner @ 2020-05-29 17:43 UTC (permalink / raw)
To: Nicolas Saenz Julienne
Cc: Mark Brown, Florian Fainelli, Ray Jui, Scott Branden,
bcm-kernel-feedback-list, Martin Sperl, linux-spi,
linux-rpi-kernel, linux-arm-kernel, linux-kernel
In-Reply-To: <20200528185805.28991-1-nsaenzjulienne@suse.de>
On Thu, May 28, 2020 at 08:58:04PM +0200, Nicolas Saenz Julienne wrote:
> --- a/drivers/spi/spi-bcm2835.c
> +++ b/drivers/spi/spi-bcm2835.c
> @@ -379,6 +379,10 @@ static irqreturn_t bcm2835_spi_interrupt(int irq, void *dev_id)
> if (bs->tx_len && cs & BCM2835_SPI_CS_DONE)
> bcm2835_wr_fifo_blind(bs, BCM2835_SPI_FIFO_SIZE);
>
> + /* check if we got interrupt enabled */
> + if (!(bcm2835_rd(bs, BCM2835_SPI_CS) & BCM2835_SPI_CS_INTR))
> + return IRQ_NONE;
> +
> /* Read as many bytes as possible from FIFO */
> bcm2835_rd_fifo(bs);
> /* Write as many bytes as possible to FIFO */
This definitely looks wrong. The check whether the interrupt is enabled
should be moved *before* the conditional calls to bcm2835_rd_fifo_blind()
and bcm2835_wr_fifo_blind(), i.e. to the top of the function.
Otherwise if an interrupt is raised by another SPI controller,
this function may perform read/write accesses to the FIFO and
interfere with an ongoing transfer in DMA or poll mode.
Also, instead of performing an MMIO read, just use the "cs" variable
which was assigned at the top of the function.
The code comment should explain that the check is necessary because the
interrupt may be shared with other controllers on the BCM2711.
Finally, it would be nice if the check would be optimized away when
compiling for pre-RasPi4 products, maybe something like:
+ if (IS_ENABLED(CONFIG_ARM_LPAE) && !(cs & BCM2835_SPI_CS_INTR))
+ return IRQ_NONE;
Thanks,
Lukas
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.