lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20180518112348.GA15574@e110455-lin.cambridge.arm.com>
Date:   Fri, 18 May 2018 12:23:48 +0100
From:   Liviu Dudau <Liviu.Dudau@....com>
To:     Brian Starkey <brian.starkey@....com>
Cc:     Mali DP Maintainers <malidp@...s.arm.com>,
        Gustavo Padovan <gustavo@...ovan.org>,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Sean Paul <seanpaul@...omium.org>,
        DRI-devel <dri-devel@...ts.freedesktop.org>,
        LKML <linux-kernel@...r.kernel.org>,
        David Airlie <airlied@...ux.ie>,
        Alexandru-Cosmin Gheorghe <alexandru-cosmin.gheorghe@....com>,
        Eric Anholt <eric@...olt.net>,
        Boris Brezillon <boris.brezillon@...tlin.com>,
        Maxime Ripard <maxime.ripard@...tlin.com>,
        Daniel Stone <daniels@...labora.com>
Subject: Re: [PATCH v7 3/5] drm/mali-dp: Add writeback support for DP500.

On Fri, May 18, 2018 at 09:56:20AM +0100, Brian Starkey wrote:
> Hi Liviu,
> 
> On Fri, May 18, 2018 at 09:24:21AM +0100, Liviu Dudau wrote:
> > Mali DP500 behaves differently from the rest of the Mali DP IP,
> > in that it does not have a one-shot mode and keeps writing the
> > content of the current frame to the provided memory area until
> > stopped. As a way of emulating the one-shot behaviour, we are
> > going to use the CVAL interrupt that is being raised at the
> > start of each frame, during prefetch phase, to act as End-of-Write
> > signal, but with a twist: we are going to disable the memory
> > write engine right after we're notified that it has been enabled,
> > using the knowledge that the bit controlling the enabling will
> > only be acted upon on the next vblank/prefetch.
> > 
> > CVAL interrupt will fire durint the next prefetch phase every time
> > the global CVAL bit gets set, so we need a state byte to track
> > the memory write enabling.
> > 
> > Signed-off-by: Liviu Dudau <liviu.dudau@....com>
> > ---
> > drivers/gpu/drm/arm/malidp_hw.c   | 77 +++++++++++++++++++++++++++++--
> > drivers/gpu/drm/arm/malidp_hw.h   |  5 +-
> > drivers/gpu/drm/arm/malidp_regs.h |  3 +-
> > 3 files changed, 80 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/arm/malidp_hw.c b/drivers/gpu/drm/arm/malidp_hw.c
> > index 455a83689d039..d9a7f19c9f219 100644
> > --- a/drivers/gpu/drm/arm/malidp_hw.c
> > +++ b/drivers/gpu/drm/arm/malidp_hw.c
> > @@ -22,6 +22,13 @@
> > #include "malidp_drv.h"
> > #include "malidp_hw.h"
> > 
> > +enum {
> > +	MW_NOT_ENABLED = 0,	/* SE writeback not enabled */
> > +	MW_ONESHOT,		/* SE in one-shot mode for writeback */
> > +	MW_START,		/* SE started writeback */
> > +	MW_STOP,		/* SE finished writeback */
> > +};
> > +
> > static const struct malidp_format_id malidp500_de_formats[] = {
> > 	/*    fourcc,   layers supporting the format,     internal id  */
> > 	{ DRM_FORMAT_ARGB2101010, DE_VIDEO1 | DE_GRAPHICS1 | DE_GRAPHICS2,  0 },
> > @@ -368,6 +375,50 @@ static long malidp500_se_calc_mclk(struct malidp_hw_device *hwdev,
> > 	return ret;
> > }
> > 
> > +static int malidp500_enable_memwrite(struct malidp_hw_device *hwdev,
> > +				     dma_addr_t *addrs, s32 *pitches,
> > +				     int num_planes, u16 w, u16 h, u32 fmt_id)
> > +{
> > +	u32 base = MALIDP500_SE_MEMWRITE_BASE;
> > +	u32 de_base = malidp_get_block_base(hwdev, MALIDP_DE_BLOCK);
> > +
> > +	/* enable the scaling engine block */
> > +	malidp_hw_setbits(hwdev, MALIDP_SCALE_ENGINE_EN, de_base + MALIDP_DE_DISPLAY_FUNC);
> > +
> > +	hwdev->mw_state = MW_START;
> > +
> > +	malidp_hw_write(hwdev, fmt_id, base + MALIDP_MW_FORMAT);
> > +	switch (num_planes) {
> > +	case 2:
> > +		malidp_hw_write(hwdev, lower_32_bits(addrs[1]), base + MALIDP_MW_P2_PTR_LOW);
> > +		malidp_hw_write(hwdev, upper_32_bits(addrs[1]), base + MALIDP_MW_P2_PTR_HIGH);
> > +		malidp_hw_write(hwdev, pitches[1], base + MALIDP_MW_P2_STRIDE);
> > +		/* fall through */
> > +	case 1:
> > +		malidp_hw_write(hwdev, lower_32_bits(addrs[0]), base + MALIDP_MW_P1_PTR_LOW);
> > +		malidp_hw_write(hwdev, upper_32_bits(addrs[0]), base + MALIDP_MW_P1_PTR_HIGH);
> > +		malidp_hw_write(hwdev, pitches[0], base + MALIDP_MW_P1_STRIDE);
> > +		break;
> > +	default:
> > +		WARN(1, "Invalid number of planes");
> > +	}
> > +
> > +	malidp_hw_write(hwdev, MALIDP_DE_H_ACTIVE(w) | MALIDP_DE_V_ACTIVE(h),
> > +			MALIDP500_SE_MEMWRITE_OUT_SIZE);
> > +	malidp_hw_setbits(hwdev, MALIDP_SE_MEMWRITE_EN, MALIDP500_SE_CONTROL);
> > +
> > +	return 0;
> > +}
> > +
> > +static void malidp500_disable_memwrite(struct malidp_hw_device *hwdev)
> > +{
> > +	u32 base = malidp_get_block_base(hwdev, MALIDP_DE_BLOCK);
> > +	if (hwdev->mw_state == MW_START)
> > +		hwdev->mw_state = MW_STOP;
> > +	malidp_hw_clearbits(hwdev, MALIDP_SE_MEMWRITE_EN, MALIDP500_SE_CONTROL);
> > +	malidp_hw_clearbits(hwdev, MALIDP_SCALE_ENGINE_EN, base + MALIDP_DE_DISPLAY_FUNC);
> > +}
> > +
> > static int malidp550_query_hw(struct malidp_hw_device *hwdev)
> > {
> > 	u32 conf = malidp_hw_read(hwdev, MALIDP550_CONFIG_ID);
> > @@ -598,6 +649,8 @@ static int malidp550_enable_memwrite(struct malidp_hw_device *hwdev,
> > 	/* enable the scaling engine block */
> > 	malidp_hw_setbits(hwdev, MALIDP_SCALE_ENGINE_EN, de_base + MALIDP_DE_DISPLAY_FUNC);
> > 
> > +	hwdev->mw_state = MW_ONESHOT;
> > +
> > 	malidp_hw_write(hwdev, fmt_id, base + MALIDP_MW_FORMAT);
> > 	switch (num_planes) {
> > 	case 2:
> > @@ -676,8 +729,9 @@ const struct malidp_hw malidp_device[MALIDP_MAX_DEVICES] = {
> > 				.vsync_irq = MALIDP500_DE_IRQ_VSYNC,
> > 			},
> > 			.se_irq_map = {
> > -				.irq_mask = MALIDP500_SE_IRQ_CONF_MODE,
> > -				.vsync_irq = 0,
> > +				.irq_mask = MALIDP500_SE_IRQ_CONF_MODE |
> > +					    MALIDP500_SE_IRQ_CONF_VALID,
> > +				.vsync_irq = MALIDP500_SE_IRQ_CONF_VALID,
> > 			},
> > 			.dc_irq_map = {
> > 				.irq_mask = MALIDP500_DE_IRQ_CONF_VALID,
> > @@ -696,6 +750,8 @@ const struct malidp_hw malidp_device[MALIDP_MAX_DEVICES] = {
> > 		.rotmem_required = malidp500_rotmem_required,
> > 		.se_set_scaling_coeffs = malidp500_se_set_scaling_coeffs,
> > 		.se_calc_mclk = malidp500_se_calc_mclk,
> > +		.enable_memwrite = malidp500_enable_memwrite,
> > +		.disable_memwrite = malidp500_disable_memwrite,
> > 		.features = MALIDP_DEVICE_LV_HAS_3_STRIDES,
> > 	},
> > 	[MALIDP_550] = {
> > @@ -934,7 +990,21 @@ static irqreturn_t malidp_se_irq(int irq, void *arg)
> > 	mask = malidp_hw_read(hwdev, hw->map.se_base + MALIDP_REG_MASKIRQ);
> > 	status = malidp_hw_read(hwdev, hw->map.se_base + MALIDP_REG_STATUS);
> > 	status &= mask;
> > -	/* ToDo: status decoding and firing up of VSYNC and page flip events */
> > +
> > +	if (status & se->vsync_irq) {
> > +		switch (hwdev->mw_state) {
> > +		case MW_STOP:
> > +		case MW_ONESHOT:
> > +			/* disable writeback after stop or oneshot */
> > +			hwdev->mw_state = MW_NOT_ENABLED;
> > +			break;
> > +		case MW_START:
> > +			/* writeback started, need to emulate one-shot mode */
> > +			hw->disable_memwrite(hwdev);
> > +			hw->set_config_valid(hwdev);
> 
> Is this serialised with incoming atomic commits? We have to make sure
> we don't set CVAL while a new scene configuration is in-progress.

We are not serialised with the incoming atomic commit, and we probably
should. I'll have a think on how to sort this one out.

> 
> I also think the current state tracking won't work properly for two
> subsequent frames using writeback. In enable_memwrite() you change the
> mw_state, but the currently ongoing job is relying on it to correctly
> signal. We probably need to either block incoming commits until the
> writeback is finished, or we need to enhance the state tracking to
> manage multiple commits.

I will go back to the drawing boards and come up with something more
solid.

Best regards,
Liviu

> 
> Thanks,
> -Brian
> 
> > +			break;
> > +		}
> > +	}
> > 
> > 	malidp_hw_clear_irq(hwdev, MALIDP_SE_BLOCK, status);
> > 
> > @@ -964,6 +1034,7 @@ int malidp_se_irq_init(struct drm_device *drm, int irq)
> > 		return ret;
> > 	}
> > 
> > +	hwdev->mw_state = MW_NOT_ENABLED;
> > 	malidp_hw_enable_irq(hwdev, MALIDP_SE_BLOCK,
> > 			     hwdev->hw->map.se_irq_map.irq_mask);
> > 
> > diff --git a/drivers/gpu/drm/arm/malidp_hw.h b/drivers/gpu/drm/arm/malidp_hw.h
> > index a242e97cf6428..c479738b81af5 100644
> > --- a/drivers/gpu/drm/arm/malidp_hw.h
> > +++ b/drivers/gpu/drm/arm/malidp_hw.h
> > @@ -178,7 +178,7 @@ struct malidp_hw {
> > 	long (*se_calc_mclk)(struct malidp_hw_device *hwdev,
> > 			     struct malidp_se_config *se_config,
> > 			     struct videomode *vm);
> > -	/**
> > +	/*
> > 	 * Enable writing to memory the content of the next frame
> > 	 * @param hwdev - malidp_hw_device structure containing the HW description
> > 	 * @param addrs - array of addresses for each plane
> > @@ -232,6 +232,9 @@ struct malidp_hw_device {
> > 	/* track the device PM state */
> > 	bool pm_suspended;
> > 
> > +	/* track the SE memory writeback state */
> > +	u8 mw_state;
> > +
> > 	/* size of memory used for rotating layers, up to two banks available */
> > 	u32 rotation_memory[2];
> > };
> > diff --git a/drivers/gpu/drm/arm/malidp_regs.h b/drivers/gpu/drm/arm/malidp_regs.h
> > index e2b2c496225e3..93b198f3af864 100644
> > --- a/drivers/gpu/drm/arm/malidp_regs.h
> > +++ b/drivers/gpu/drm/arm/malidp_regs.h
> > @@ -198,7 +198,8 @@
> > #define MALIDP500_DE_LG2_PTR_BASE	0x0031c
> > #define MALIDP500_SE_BASE		0x00c00
> > #define MALIDP500_SE_CONTROL		0x00c0c
> > -#define MALIDP500_SE_PTR_BASE		0x00e0c
> > +#define MALIDP500_SE_MEMWRITE_OUT_SIZE	0x00c2c
> > +#define MALIDP500_SE_MEMWRITE_BASE	0x00e00
> > #define MALIDP500_DC_IRQ_BASE		0x00f00
> > #define MALIDP500_CONFIG_VALID		0x00f00
> > #define MALIDP500_CONFIG_ID		0x00fd4
> > -- 
> > 2.17.0
> > 

-- 
====================
| I would like to |
| fix the world,  |
| but they're not |
| giving me the   |
 \ source code!  /
  ---------------
    ¯\_(ツ)_/¯

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ