From: Sam Ravnborg <sam@ravnborg.org>
To: Thomas Zimmermann <tzimmermann@suse.de>
Cc: john.p.donnelly@oracle.com, dri-devel@lists.freedesktop.org,
kraxel@redhat.com, airlied@redhat.com
Subject: Re: [PATCH 05/17] drm/mgag200: Clean up mga_set_start_address()
Date: Wed, 29 Apr 2020 20:20:23 +0200 [thread overview]
Message-ID: <20200429182023.GE31071@ravnborg.org> (raw)
In-Reply-To: <20200429143238.10115-6-tzimmermann@suse.de>
Hi Thomas,
On Wed, Apr 29, 2020 at 04:32:26PM +0200, Thomas Zimmermann wrote:
> All register names and fields are now named according to the
> MGA programming manuals. The function doesn't need the CRTC, so
> callers pass in the device structure directly. The logging now
> uses device-specific macros.
>
> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
> ---
> drivers/gpu/drm/mgag200/mgag200_drv.h | 5 ++
> drivers/gpu/drm/mgag200/mgag200_mode.c | 82 +++++++++++++++-----------
> 2 files changed, 53 insertions(+), 34 deletions(-)
>
> diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h b/drivers/gpu/drm/mgag200/mgag200_drv.h
> index 4403145e3593c..9b957d9fc7e04 100644
> --- a/drivers/gpu/drm/mgag200/mgag200_drv.h
> +++ b/drivers/gpu/drm/mgag200/mgag200_drv.h
> @@ -61,6 +61,11 @@
> WREG8(MGAREG_CRTC_DATA, v); \
> } while (0) \
>
> +#define RREG_ECRT(reg, v) \
> + do { \
> + WREG8(MGAREG_CRTCEXT_INDEX, reg); \
> + v = RREG8(MGAREG_CRTCEXT_DATA); \
> + } while (0) \
>
> #define WREG_ECRT(reg, v) \
> do { \
> diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c
> index 3d894b37a0812..b16a73c8617d6 100644
> --- a/drivers/gpu/drm/mgag200/mgag200_mode.c
> +++ b/drivers/gpu/drm/mgag200/mgag200_mode.c
> @@ -819,49 +819,53 @@ static void mga_g200wb_commit(struct drm_crtc *crtc)
> }
>
> /*
> - This is how the framebuffer base address is stored in g200 cards:
> - * Assume @offset is the gpu_addr variable of the framebuffer object
> - * Then addr is the number of _pixels_ (not bytes) from the start of
> - VRAM to the first pixel we want to display. (divided by 2 for 32bit
> - framebuffers)
> - * addr is stored in the CRTCEXT0, CRTCC and CRTCD registers
> - addr<20> -> CRTCEXT0<6>
> - addr<19-16> -> CRTCEXT0<3-0>
> - addr<15-8> -> CRTCC<7-0>
> - addr<7-0> -> CRTCD<7-0>
> - CRTCEXT0 has to be programmed last to trigger an update and make the
> - new addr variable take effect.
> + * This is how the framebuffer base address is stored in g200 cards:
> + * * Assume @offset is the gpu_addr variable of the framebuffer object
> + * * Then addr is the number of _pixels_ (not bytes) from the start of
> + * VRAM to the first pixel we want to display. (divided by 2 for 32bit
> + * framebuffers)
> + * * addr is stored in the CRTCEXT0, CRTCC and CRTCD registers
> + * addr<20> -> CRTCEXT0<6>
> + * addr<19-16> -> CRTCEXT0<3-0>
> + * addr<15-8> -> CRTCC<7-0>
> + * addr<7-0> -> CRTCD<7-0>
> + *
> + * CRTCEXT0 has to be programmed last to trigger an update and make the
> + * new addr variable take effect.
> */
> -static void mga_set_start_address(struct drm_crtc *crtc, unsigned offset)
> +static void mgag200_set_startadd(struct mga_device *mdev,
> + unsigned long offset)
> {
> - struct mga_device *mdev = crtc->dev->dev_private;
> - u32 addr;
> - int count;
> - u8 crtcext0;
> + struct drm_device *dev = mdev->dev;
> + uint32_t startadd;
> + uint8_t crtcc, crtcd, crtcext0;
>
> - while (RREG8(0x1fda) & 0x08);
> - while (!(RREG8(0x1fda) & 0x08));
> + startadd = offset / 8;
>
> - count = RREG8(MGAREG_VCOUNT) + 2;
> - while (RREG8(MGAREG_VCOUNT) < count);
> -
> - WREG8(MGAREG_CRTCEXT_INDEX, 0);
> - crtcext0 = RREG8(MGAREG_CRTCEXT_DATA);
> - crtcext0 &= 0xB0;
> - addr = offset / 8;
> - /* Can't store addresses any higher than that...
> - but we also don't have more than 16MB of memory, so it should be fine. */
> - WARN_ON(addr > 0x1fffff);
> - crtcext0 |= (!!(addr & (1<<20)))<<6;
> - WREG_CRT(0x0d, (u8)(addr & 0xff));
> - WREG_CRT(0x0c, (u8)(addr >> 8) & 0xff);
> - WREG_ECRT(0x0, ((u8)(addr >> 16) & 0xf) | crtcext0);
> + /*
> + * Can't store addresses any higher than that, but we also
> + * don't have more than 16MB of memory, so it should be fine.
> + */
> + drm_WARN_ON(dev, startadd > 0x1fffff);
> +
> + RREG_ECRT(0x00, crtcext0);
> +
> + crtcc = (startadd >> 8) & 0xff;
> + crtcd = startadd & 0xff;
> + crtcext0 &= 0xb0;
> + crtcext0 |= ((startadd >> 14) & BIT(6)) |
It is not so obvious that the value of bit 20 is stored in bit 6 here.
Maybe:
crtcext0 |= ((startadd & BIT(20) >> 14) |
I would find the above easier to parse.
> + ((startadd >> 16) & 0x0f);
> +
> + WREG_CRT(0x0c, crtcc);
> + WREG_CRT(0x0d, crtcd);
> + WREG_ECRT(0x00, crtcext0);
> }
>
> static int mga_crtc_do_set_base(struct drm_crtc *crtc,
> struct drm_framebuffer *fb,
> int x, int y, int atomic)
> {
> + struct mga_device *mdev = crtc->dev->dev_private;
Could you use a crtc_to_mdev() macro here.
So we avoid adding new users of dev_private?
> struct drm_gem_vram_object *gbo;
> int ret;
> s64 gpu_addr;
Make this unsigned long and..
> @@ -882,7 +886,7 @@ static int mga_crtc_do_set_base(struct drm_crtc *crtc,
> goto err_drm_gem_vram_unpin;
> }
>
> - mga_set_start_address(crtc, (u32)gpu_addr);
> + mgag200_set_startadd(mdev, (unsigned long)gpu_addr);
drop this cast.
>
> return 0;
>
> @@ -894,6 +898,16 @@ static int mga_crtc_do_set_base(struct drm_crtc *crtc,
> static int mga_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
> struct drm_framebuffer *old_fb)
> {
> + struct drm_device *dev = crtc->dev;
> + struct mga_device *mdev = dev->dev_private;
> + unsigned int count;
> +
> + while (RREG8(0x1fda) & 0x08) { }
> + while (!(RREG8(0x1fda) & 0x08)) { }
> +
> + count = RREG8(MGAREG_VCOUNT) + 2;
> + while (RREG8(MGAREG_VCOUNT) < count) { }
> +
> return mga_crtc_do_set_base(crtc, old_fb, x, y, 0);
> }
I do not really see why this code was lifter two functions up.
Before is was deep in mga_set_start_address(), now it is in
mga_crtc_mode_set_base().
Puzzeled?
Sam
>
> --
> 2.26.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
next prev parent reply other threads:[~2020-04-29 18:20 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-29 14:32 [PATCH 00/17] drm/mgag200: Convert to atomic modesetting Thomas Zimmermann
2020-04-29 14:32 ` [PATCH 01/17] drm/mgag200: Remove HW cursor Thomas Zimmermann
2020-04-29 17:51 ` Sam Ravnborg
2020-04-30 7:03 ` Gerd Hoffmann
2020-04-30 8:10 ` Thomas Zimmermann
2020-04-30 9:19 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 02/17] drm/mgag200: Remove unused fields from struct mga_device Thomas Zimmermann
2020-04-29 17:49 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 03/17] drm/mgag200: Embed connector instance in " Thomas Zimmermann
2020-04-29 15:24 ` Ruhl, Michael J
2020-04-29 17:49 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 04/17] drm/mgag200: Use managed mode-config initialization Thomas Zimmermann
2020-04-29 17:55 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 05/17] drm/mgag200: Clean up mga_set_start_address() Thomas Zimmermann
2020-04-29 18:20 ` Sam Ravnborg [this message]
2020-04-30 8:23 ` Thomas Zimmermann
2020-05-11 12:41 ` Thomas Zimmermann
2020-04-29 14:32 ` [PATCH 06/17] drm/mgag200: Clean up mga_crtc_do_set_base() Thomas Zimmermann
2020-04-29 18:23 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 07/17] drm/mgag200: Move mode-setting code into separate helper function Thomas Zimmermann
2020-04-29 18:24 ` Sam Ravnborg
2020-04-30 8:27 ` Thomas Zimmermann
2020-04-29 14:32 ` [PATCH 08/17] drm/mgag200: Split MISC register update into PLL selection, SYNC and I/O Thomas Zimmermann
2020-05-03 15:34 ` Sam Ravnborg
2020-05-04 13:03 ` Thomas Zimmermann
2020-05-04 14:25 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 09/17] drm/mgag200: Update mode registers after plane registers Thomas Zimmermann
2020-05-03 15:34 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 10/17] drm/mgag200: Set pitch in a separate helper function Thomas Zimmermann
2020-05-03 15:42 ` Sam Ravnborg
2020-05-04 13:10 ` Thomas Zimmermann
2020-04-29 14:32 ` [PATCH 11/17] drm/mgag200: Set primary plane's format in " Thomas Zimmermann
2020-04-29 14:32 ` [PATCH 12/17] drm/mgag200: Move TAGFIFO reset into separate function Thomas Zimmermann
2020-05-03 16:25 ` Sam Ravnborg
2020-05-04 13:11 ` Thomas Zimmermann
2020-05-04 14:29 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 13/17] drm/mgag200: Move hiprilvl setting into separate functions Thomas Zimmermann
2020-05-03 17:23 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 14/17] drm/mgag200: Move register initialization into separate function Thomas Zimmermann
2020-05-03 17:25 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 15/17] drm/mgag200: Remove waiting from DPMS code Thomas Zimmermann
2020-05-04 12:10 ` Daniel Vetter
2020-05-04 12:40 ` Thomas Zimmermann
2020-04-29 14:32 ` [PATCH 16/17] drm/mgag200: Convert to simple KMS helper Thomas Zimmermann
2020-05-03 17:36 ` Sam Ravnborg
2020-04-29 14:32 ` [PATCH 17/17] drm/mgag200: Replace VRAM helpers with SHMEM helpers Thomas Zimmermann
2020-05-04 12:29 ` Emil Velikov
2020-05-04 12:45 ` Thomas Zimmermann
2020-04-30 0:11 ` [PATCH 00/17] drm/mgag200: Convert to atomic modesetting John Donnelly
2020-04-30 8:29 ` Thomas Zimmermann
2020-04-30 12:09 ` John Donnelly
2020-05-04 13:39 ` Thomas Zimmermann
2020-05-04 20:39 ` John Donnelly
2020-05-05 12:20 ` John Donnelly
2020-05-06 7:29 ` Thomas Zimmermann
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200429182023.GE31071@ravnborg.org \
--to=sam@ravnborg.org \
--cc=airlied@redhat.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=john.p.donnelly@oracle.com \
--cc=kraxel@redhat.com \
--cc=tzimmermann@suse.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox