Message ID | 20220503085935.1533814-1-jic23@kernel.org (mailing list archive) |
---|---|
Headers | show |
Series | IIO: Fix alignment of buffers for DMA | expand |
Hi Jonathan, > From: Jonathan Cameron <jic23@kernel.org> > Sent: Tuesday, May 3, 2022 10:58 AM > To: linux-iio@vger.kernel.org > Cc: Akinobu Mita <akinobu.mita@gmail.com>; Alexandru Lazar > <alazar@startmail.com>; Tachici, Alexandru > <Alexandru.Tachici@analog.com>; Miclaus, Antoniu > <Antoniu.Miclaus@analog.com>; Charles-Antoine Couret <charles- > antoine.couret@essensium.com>; Tanislav, Cosmin > <Cosmin.Tanislav@analog.com>; Cristian Pop > <cristian.pop@analog.com>; David Lechner <david@lechnology.com>; > Ivan Mikhaylov <i.mikhaylov@yadro.com>; Jacopo Mondi > <jacopo+renesas@jmondi.org>; Jean-Baptiste Maneyrol > <jmaneyrol@invensense.com>; Lars-Peter Clausen > <lars@metafoo.de>; Marcelo Schmitt <marcelo.schmitt1@gmail.com>; > Mårten Lindahl <martenli@axis.com>; Matt Ranostay > <mranostay@gmail.com>; Hennerich, Michael > <Michael.Hennerich@analog.com>; Michael Welling > <mwelling@ieee.org>; Mugilraj Dhavachelvan > <dmugil2000@gmail.com>; Navin Sankar Velliangiri > <navin@linumiz.com>; Sa, Nuno <Nuno.Sa@analog.com>; Paul > Cercueil <paul@crapouillou.net>; Phil Reid > <preid@electromag.com.au>; Puranjay Mohan > <puranjay12@gmail.com>; Ricardo Ribalda <ribalda@kernel.org>; > Robert Jones <rjones@gateworks.com>; Rui Miguel Silva > <rui.silva@linaro.org>; Sean Nyekjaer <sean.nyekjaer@prevas.dk>; > Tomas Melin <tomas.melin@vaisala.com>; Tomislav Denis > <tomislav.denis@avl.com>; Uwe Kleine-König <u.kleine- > koenig@pengutronix.de>; Jonathan Cameron > <Jonathan.Cameron@huawei.com>; catalin.marinas@arm.com > Subject: [PATCH 00/92] IIO: Fix alignment of buffers for DMA > > [External] > > From: Jonathan Cameron <Jonathan.Cameron@huawei.com> > > Many years ago, IIO started using ____cacheline_aligned to ensure > that buffers that might be used for DMA were not sharing a cacheline > with other data that might lead to DMA safety issues. > > As it turns out, that was fine at the time, but not based on the > correct alignment requirement (though I believe it was a conservative > choice at the time). Note that on many architectures this was > introducing > unecessary padding. We should have been using > ARCH_KMALLOC_MINALIGN > as other subsystems such as crypto have done for a long time. > > Patch 1 discription contains more information but in short, there are > ARM64 SoCs out their that have a larger cachline size for caches > beyond > L1. In many cases they maintain coherency for all DMA devices > attached > and so this isn't a problem, but there are exceptions that do not. > > So, this is a rather large patch set and just covers those drivers > that are in the last kernel release and in drivers/iio. > > Many of these drivers are somewhat old so I haven't specifically > cc'd anyone so will be relying on those kind enough to sanity check > patches on drivers that are beyond their own. > > Given there is ongoing discussion around reducing the alignment > requirements where possible, I've adopted the existing IIO_ALIGN > define througout. That way we have a single point to update if > that becomes relevant in future. > Nice to see this in... Since we are here, I guess in a couple of patches where we have: struct { ... u8 rx[] __aligned(); u8 tx[] __aligned(); }; we could make it such that only the first member is aligned. But bah, I get that this is already a huge patchset to diverge from it. Anyhow, Acked-by: Nuno Sá <nuno.sa@analog.com>
On Wed, 4 May 2022 13:00:58 +0000 "Sa, Nuno" <Nuno.Sa@analog.com> wrote: > Hi Jonathan, > > > From: Jonathan Cameron <jic23@kernel.org> > > Sent: Tuesday, May 3, 2022 10:58 AM > > To: linux-iio@vger.kernel.org > > Cc: Akinobu Mita <akinobu.mita@gmail.com>; Alexandru Lazar > > <alazar@startmail.com>; Tachici, Alexandru > > <Alexandru.Tachici@analog.com>; Miclaus, Antoniu > > <Antoniu.Miclaus@analog.com>; Charles-Antoine Couret <charles- > > antoine.couret@essensium.com>; Tanislav, Cosmin > > <Cosmin.Tanislav@analog.com>; Cristian Pop > > <cristian.pop@analog.com>; David Lechner <david@lechnology.com>; > > Ivan Mikhaylov <i.mikhaylov@yadro.com>; Jacopo Mondi > > <jacopo+renesas@jmondi.org>; Jean-Baptiste Maneyrol > > <jmaneyrol@invensense.com>; Lars-Peter Clausen > > <lars@metafoo.de>; Marcelo Schmitt <marcelo.schmitt1@gmail.com>; > > Mårten Lindahl <martenli@axis.com>; Matt Ranostay > > <mranostay@gmail.com>; Hennerich, Michael > > <Michael.Hennerich@analog.com>; Michael Welling > > <mwelling@ieee.org>; Mugilraj Dhavachelvan > > <dmugil2000@gmail.com>; Navin Sankar Velliangiri > > <navin@linumiz.com>; Sa, Nuno <Nuno.Sa@analog.com>; Paul > > Cercueil <paul@crapouillou.net>; Phil Reid > > <preid@electromag.com.au>; Puranjay Mohan > > <puranjay12@gmail.com>; Ricardo Ribalda <ribalda@kernel.org>; > > Robert Jones <rjones@gateworks.com>; Rui Miguel Silva > > <rui.silva@linaro.org>; Sean Nyekjaer <sean.nyekjaer@prevas.dk>; > > Tomas Melin <tomas.melin@vaisala.com>; Tomislav Denis > > <tomislav.denis@avl.com>; Uwe Kleine-König <u.kleine- > > koenig@pengutronix.de>; Jonathan Cameron > > <Jonathan.Cameron@huawei.com>; catalin.marinas@arm.com > > Subject: [PATCH 00/92] IIO: Fix alignment of buffers for DMA > > > > [External] > > > > From: Jonathan Cameron <Jonathan.Cameron@huawei.com> > > > > Many years ago, IIO started using ____cacheline_aligned to ensure > > that buffers that might be used for DMA were not sharing a cacheline > > with other data that might lead to DMA safety issues. > > > > As it turns out, that was fine at the time, but not based on the > > correct alignment requirement (though I believe it was a conservative > > choice at the time). Note that on many architectures this was > > introducing > > unecessary padding. We should have been using > > ARCH_KMALLOC_MINALIGN > > as other subsystems such as crypto have done for a long time. > > > > Patch 1 discription contains more information but in short, there are > > ARM64 SoCs out their that have a larger cachline size for caches > > beyond > > L1. In many cases they maintain coherency for all DMA devices > > attached > > and so this isn't a problem, but there are exceptions that do not. > > > > So, this is a rather large patch set and just covers those drivers > > that are in the last kernel release and in drivers/iio. > > > > Many of these drivers are somewhat old so I haven't specifically > > cc'd anyone so will be relying on those kind enough to sanity check > > patches on drivers that are beyond their own. > > > > Given there is ongoing discussion around reducing the alignment > > requirements where possible, I've adopted the existing IIO_ALIGN > > define througout. That way we have a single point to update if > > that becomes relevant in future. > > > > Nice to see this in... Since we are here, I guess in a couple of patches where we have: > > struct { > ... > u8 rx[] __aligned(); > u8 tx[] __aligned(); > }; > > we could make it such that only the first member is aligned. But bah, I get that this > is already a huge patchset to diverge from it. Anyhow, Yup. I thought about tidying those up but decided it would just make things harder to explain so could be done as a follow up. I should have said so in the cover letter though! I particularly like the case where (a long time ago) I added a question on whether there was a need for separate alignment and left that question in the final code. Oops. > > Acked-by: Nuno Sá <nuno.sa@analog.com> Thanks for looking through all these. We haven't yet concluded on the questions of naming and possible additional define for __aligned(XXX) but if we do make such a change it should be purely mechanical hence I'll keep all the tags people have given and not bother them again :) Thanks, Jonathan
From: Jonathan Cameron <Jonathan.Cameron@huawei.com> Many years ago, IIO started using ____cacheline_aligned to ensure that buffers that might be used for DMA were not sharing a cacheline with other data that might lead to DMA safety issues. As it turns out, that was fine at the time, but not based on the correct alignment requirement (though I believe it was a conservative choice at the time). Note that on many architectures this was introducing unecessary padding. We should have been using ARCH_KMALLOC_MINALIGN as other subsystems such as crypto have done for a long time. Patch 1 discription contains more information but in short, there are ARM64 SoCs out their that have a larger cachline size for caches beyond L1. In many cases they maintain coherency for all DMA devices attached and so this isn't a problem, but there are exceptions that do not. So, this is a rather large patch set and just covers those drivers that are in the last kernel release and in drivers/iio. Many of these drivers are somewhat old so I haven't specifically cc'd anyone so will be relying on those kind enough to sanity check patches on drivers that are beyond their own. Given there is ongoing discussion around reducing the alignment requirements where possible, I've adopted the existing IIO_ALIGN define througout. That way we have a single point to update if that becomes relevant in future. Cc: catalin.marinas@arm.com Jonathan Cameron (92): iio: core: Fix IIO_ALIGN as it was not sufficiently large on some platforms. iio: accel: adxl313: Fix alignment for DMA safety iio: accel: adxl355: Fix alignment for DMA safety iio: accel: adxl367: Fix alignment for DMA safety iio: accel: bma220: Fix alignment for DMA safety iio: accel: bmi088: Fix alignment for DMA safety iio: accel: sca3000: Fix alignment for DMA safety iio: accel: sca3000: Fix alignment for DMA safety iio: adc: ad7266: Fix alignment for DMA safety iio: adc: ad7280a: Fix alignment for DMA safety iio: adc: ad7292: Fix alignment for DMA safety iio: adc: ad7298: Fix alignment for DMA safety iio: adc: ad7476: Fix alignment for DMA safety iio: adc: ad7606: Fix alignment for DMA safety iio: adc: ad7766: Fix alignment for DMA safety iio: adc: ad7768-1: Fix alignment for DMA safety iio: adc: ad7887: Fix alignment for DMA safety iio: adc: ad7923: Fix alignment for DMA safety iio: adc: ad7949: Fix alignment for DMA safety iio: adc: hi8435: Fix alignment for DMA safety iio: adc: ltc2496: Fix alignment for DMA safety iio: adc: ltc2497: Fix alignment for DMA safety iio: adc: max1027: Fix alignment for DMA safety iio: adc: max11100: Fix alignment for DMA safety iio: adc: max1118: Fix alignment for DMA safety iio: adc: max1241: Fix alignment for DMA safety iio: adc: mcp320x: Fix alignment for DMA safety iio: adc: ti-adc0832: Fix alignment for DMA safety iio: adc: ti-adc084s021: Fix alignment for DMA safety iio: adc: ti-adc108s102: Fix alignment for DMA safety iio: adc: ti-adc12138: Fix alignment for DMA safety iio: adc: ti-adc128s052: Fix alignment for DMA safety iio: adc: ti-adc161s626: Fix alignment for DMA safety iio: adc: ti-ads124s08: Fix alignment for DMA safety iio: adc: ti-ads131e08: Fix alignment for DMA safety iio: adc: ti-ads7950: Fix alignment for DMA safety iio: adc: ti-ads8344: Fix alignment for DMA safety iio: adc: ti-ads8688: Fix alignment for DMA safety iio: adc: ti-tlc4541: Fix alignment for DMA safety iio: addac: ad74413r: Fix alignment for DMA safety iio: amplifiers: ad8366: Fix alignment for DMA safety iio: common: ssp: Fix alignment for DMA safety iio: dac: ad5064: Fix alignment for DMA safety iio: dac: ad5360: Fix alignment for DMA safety iio: dac: ad5421: Fix alignment for DMA safety iio: dac: ad5449: Fix alignment for DMA safety iio: dac: ad5504: Fix alignment for DMA safety iio: dac: ad5592r: Fix alignment for DMA safety iio: dac: ad5686: Fix alignment for DMA safety iio: dac: ad5755: Fix alignment for DMA safety iio: dac: ad5761: Fix alignment for DMA safety iio: dac: ad5764: Fix alignment for DMA safety iio: dac: ad5766: Fix alignment for DMA safety iio: dac: ad5770r: Fix alignment for DMA safety iio: dac: ad5791: Fix alignment for DMA saftey iio: dac: ad7293: Fix alignment for DMA safety iio: dac: ad7303: Fix alignment for DMA safety iio: dac: ad8801: Fix alignment for DMA safety iio: dac: ltc2688: Fix alignment for DMA safety iio: dac: mcp3922: Fix alignment for DMA safety iio: dac: ti-dac082s085: Fix alignment for DMA safety iio: dac: ti-dac5571: Fix alignment for DMA safety iio: dac: ti-dac7311: Fix alignment for DMA safety iio: dac: ti-dac7612: Fix alignment for DMA safety iio: frequency: ad9523: Fix alignment for DMA safety iio: frequency: adf4350: Fix alignment for DMA safety iio: frequency: adf4371: Fix alignment for DMA safety iio: frequency: admv1013: Fix alignment for DMA safety iio: frequency: admv1014: Fix alignment for DMA safety iio: frequency: admv4420: Fix alignment for DMA safety iio: frequency: adrf6780: Fix alignment for DMA safety iio: gyro: adis16080: Fix alignment for DMA safety iio: gyro: adis16130: Fix alignment for DMA safety iio: gyro: adxrs450: Fix alignment for DMA safety iio: gyro: fxas210002c: Fix alignment for DMA safety iio: imu: fxos8700: Fix alignment for DMA safety iio: imu_icm42600: Fix alignment for DMA safety iio: imu: icm42600: Fix alignment for DMA safety in buffer code. iio: imu: mpu6050: Fix alignment for DMA safety iio: potentiometer: ad5110: Fix alignment for DMA safety iio: potentiometer: ad5272: Fix alignment for DMA safety iio: potentiometer: max5481: Fix alignment for DMA safety iio: potentiometer: mcp41010: Fix alignment for DMA safety iio: potentiometer: mcp4131: Fix alignment for DMA safety iio: pressure: dlhl60d: Drop unnecessary alignment forcing. iio: proximity: as3935: Fix alignment for DMA safety iio: proximity: vcnl3020: Drop unnecessary alignment requirement for i2c device iio: resolver: ad2s1200: Fix alignment for DMA safety iio: resolver: ad2s90: Fix alignment for DMA safety iio: temp: ltc2983: Fix alignment for DMA safety iio: temp: max31865: Fix alignment for DMA safety iio: temp: maxim_thermocouple: Fix alignment for DMA safety drivers/iio/accel/adxl313_core.c | 2 +- drivers/iio/accel/adxl355_core.c | 2 +- drivers/iio/accel/adxl367.c | 2 +- drivers/iio/accel/adxl367_spi.c | 8 +++++--- drivers/iio/accel/bma220_spi.c | 2 +- drivers/iio/accel/bmi088-accel-core.c | 2 +- drivers/iio/accel/sca3000.c | 4 ++-- drivers/iio/accel/sca3300.c | 2 +- drivers/iio/adc/ad7266.c | 6 +++--- drivers/iio/adc/ad7280a.c | 2 +- drivers/iio/adc/ad7292.c | 2 +- drivers/iio/adc/ad7298.c | 2 +- drivers/iio/adc/ad7476.c | 5 ++--- drivers/iio/adc/ad7606.h | 6 +++--- drivers/iio/adc/ad7766.c | 5 ++--- drivers/iio/adc/ad7768-1.c | 4 ++-- drivers/iio/adc/ad7887.c | 5 ++--- drivers/iio/adc/ad7923.c | 4 ++-- drivers/iio/adc/ad7949.c | 2 +- drivers/iio/adc/hi8435.c | 2 +- drivers/iio/adc/ltc2496.c | 4 ++-- drivers/iio/adc/ltc2497.c | 4 ++-- drivers/iio/adc/max1027.c | 2 +- drivers/iio/adc/max11100.c | 4 ++-- drivers/iio/adc/max1118.c | 2 +- drivers/iio/adc/max1241.c | 2 +- drivers/iio/adc/mcp320x.c | 2 +- drivers/iio/adc/ti-adc0832.c | 2 +- drivers/iio/adc/ti-adc084s021.c | 4 ++-- drivers/iio/adc/ti-adc108s102.c | 4 ++-- drivers/iio/adc/ti-adc12138.c | 2 +- drivers/iio/adc/ti-adc128s052.c | 2 +- drivers/iio/adc/ti-adc161s626.c | 2 +- drivers/iio/adc/ti-ads124s08.c | 2 +- drivers/iio/adc/ti-ads131e08.c | 2 +- drivers/iio/adc/ti-ads7950.c | 4 ++-- drivers/iio/adc/ti-ads8344.c | 2 +- drivers/iio/adc/ti-ads8688.c | 2 +- drivers/iio/adc/ti-tlc4541.c | 4 ++-- drivers/iio/addac/ad74413r.c | 4 ++-- drivers/iio/amplifiers/ad8366.c | 4 ++-- drivers/iio/common/ssp_sensors/ssp.h | 3 +-- drivers/iio/dac/ad5064.c | 4 ++-- drivers/iio/dac/ad5360.c | 4 ++-- drivers/iio/dac/ad5421.c | 4 ++-- drivers/iio/dac/ad5449.c | 4 ++-- drivers/iio/dac/ad5504.c | 2 +- drivers/iio/dac/ad5592r-base.h | 4 +++- drivers/iio/dac/ad5686.h | 6 ++++-- drivers/iio/dac/ad5755.c | 4 ++-- drivers/iio/dac/ad5761.c | 4 ++-- drivers/iio/dac/ad5764.c | 4 ++-- drivers/iio/dac/ad5766.c | 2 +- drivers/iio/dac/ad5770r.c | 2 +- drivers/iio/dac/ad5791.c | 2 +- drivers/iio/dac/ad7293.c | 2 +- drivers/iio/dac/ad7303.c | 4 ++-- drivers/iio/dac/ad8801.c | 2 +- drivers/iio/dac/ltc2688.c | 4 ++-- drivers/iio/dac/mcp4922.c | 2 +- drivers/iio/dac/ti-dac082s085.c | 2 +- drivers/iio/dac/ti-dac5571.c | 2 +- drivers/iio/dac/ti-dac7311.c | 2 +- drivers/iio/dac/ti-dac7612.c | 4 ++-- drivers/iio/frequency/ad9523.c | 6 +++--- drivers/iio/frequency/adf4350.c | 6 +++--- drivers/iio/frequency/adf4371.c | 2 +- drivers/iio/frequency/admv1013.c | 2 +- drivers/iio/frequency/admv1014.c | 2 +- drivers/iio/frequency/admv4420.c | 2 +- drivers/iio/frequency/adrf6780.c | 2 +- drivers/iio/gyro/adis16080.c | 2 +- drivers/iio/gyro/adis16130.c | 2 +- drivers/iio/gyro/adxrs450.c | 2 +- drivers/iio/gyro/fxas21002c_core.c | 6 +++--- drivers/iio/imu/fxos8700_core.c | 2 +- drivers/iio/imu/inv_icm42600/inv_icm42600.h | 2 +- drivers/iio/imu/inv_icm42600/inv_icm42600_buffer.h | 2 +- drivers/iio/imu/inv_mpu6050/inv_mpu_iio.h | 2 +- drivers/iio/potentiometer/ad5110.c | 6 +++--- drivers/iio/potentiometer/ad5272.c | 2 +- drivers/iio/potentiometer/max5481.c | 2 +- drivers/iio/potentiometer/mcp41010.c | 2 +- drivers/iio/potentiometer/mcp4131.c | 2 +- drivers/iio/pressure/dlhl60d.c | 2 +- drivers/iio/proximity/as3935.c | 2 +- drivers/iio/proximity/vcnl3020.c | 4 ++-- drivers/iio/resolver/ad2s1200.c | 2 +- drivers/iio/resolver/ad2s90.c | 2 +- drivers/iio/temperature/ltc2983.c | 6 +++--- drivers/iio/temperature/max31865.c | 2 +- drivers/iio/temperature/maxim_thermocouple.c | 2 +- include/linux/iio/iio.h | 10 ++++++++-- 93 files changed, 149 insertions(+), 141 deletions(-)