From: Simon Horman <horms@verge.net.au>
To: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Cc: linux-mmc@vger.kernel.org, linux-sh@vger.kernel.org,
Chris Ball <cjb@laptop.org>, Magnus Damm <magnus.damm@gmail.com>,
Paul Mundt <lethal@linux-sh.org>
Subject: Re: [PATCH 3/4] mmc: sdhi: Make use of per-source irq handlers
Date: Wed, 17 Aug 2011 21:57:31 +0900 [thread overview]
Message-ID: <20110817125728.GD13086@verge.net.au> (raw)
In-Reply-To: <Pine.LNX.4.64.1108171416010.18317@axis700.grange>
On Wed, Aug 17, 2011 at 02:22:56PM +0200, Guennadi Liakhovetski wrote:
> On Wed, 17 Aug 2011, Simon Horman wrote:
>
> > On Wed, Aug 17, 2011 at 08:41:47PM +0900, Simon Horman wrote:
> > > On Wed, Aug 17, 2011 at 01:24:58PM +0200, Guennadi Liakhovetski wrote:
> > > > On Wed, 17 Aug 2011, Simon Horman wrote:
> > > >
> > > > > Make use of per-source irq handles if the
> > > > > platform (data) has multiple irq sources.
> > > > >
> > > > > Also, as suggested by Guennadi Liakhovetski,
> > > > > add and use defines the index or irqs in platform data.
> > > > >
> > > > > Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
> > > > > Cc: Magnus Damm <magnus.damm@gmail.com>
> > > > > Signed-off-by: Simon Horman <horms@verge.net.au>
> > > > >
> > > > > ---
> > > > >
> > > > > v5
> > > > > * As suggested by Guennadi Liakhovetski:
> > > > > - Allow only SH_MOBILE_SDHI_IRQ_SDCARD and
> > > > > SH_MOBILE_SDHI_IRQ_SDIO to be specified in platform data
> > > >
> > > > This means, CARD_DETECT is optional in a multi-irq configuration, agree.
> > >
> > > [snip]
> > >
> > > > > + ret = irq = platform_get_irq(pdev, SH_MOBILE_SDHI_IRQ_SDCARD);
> > > > > + if (irq >= 0) {
> > > > > + multi_irq = true;
> > > > > + ret = request_irq(irq, tmio_mmc_sdcard_irq, 0,
> > > > > + dev_name(&pdev->dev), host);
> > > > > + if (ret)
> > > > > + goto eirq_sdcard;
> > > > > }
> > > >
> > > > This means SDCARD is optional in both single- and multi-irq
> > > > configurations.
> > > >
> > > > > +
> > > > > + ret = irq = platform_get_irq(pdev, SH_MOBILE_SDHI_IRQ_CARD_DETECT);
> > > > > + if (irq < 0)
> > > > > + goto eirq_card_detect;
> > > >
> > > > This means, CARD_DETECT is compulsory in all configurations.
> > > >
> > > > One of us must be speaking klingonish today.
> > >
> > > Sorry, I messed that up a bit.
> > >
> > > I will make CARD_DETECT optional in multi-irq configurations.
> > > Likewise SDIO will be optional in multi-irq configurations.
> > >
> > > But SDCARD will always be required.
> >
> > Oops, that scheme won't work because in single-irq configuration
> > the only IRQ will be #0, but SDCARD != 0.
> >
> > So I have it in mind to allow the following combinations.
> >
> > i) SDIO, SDCARD, CARD_DETECT
> > ii) SDIO, SDCARD
> > iii) SDCARD, CARD_DETECT
> > iv) SDCARD only (sets multi_irq = true locally, but only one irq!)
> > v) CARD_DETECT only (uses tmio_mmc_irq(), traditional single-irq setup)
>
> I would make it simple:
>
> (1) 1 IRQ: only resource #0 (CARD_DETECT, use tmio_mmc_irq())
> (2) 2 or 3 IRQs: compulsory SDCARD and any further IRQs: use respective
> specialised ISRs.
>
> > Do you think iii and iv are necessary?
>
> Accordingly: iii - yes, iv - no.
Ok, how about this?
>From 6514563c5637e65a397d0bf88cda95168a69dea1 Mon Sep 17 00:00:00 2001
From: Simon Horman <horms@verge.net.au>
Date: Tue, 16 Aug 2011 10:42:35 +0900
Subject: [PATCH] mmc: sdhi: Make use of per-source irq handlers
Make use of per-source irq handles if the
platform (data) has multiple irq sources.
Also, as suggested by Guennadi Liakhovetski,
add and use defines the index or irqs in platform data.
Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Cc: Magnus Damm <magnus.damm@gmail.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
v6
* As discussed with Guennadi Liakhovetski:
- The SDCARD/SDIO change in v5 was not implemented as described.
And the logic wasn't fully described in any case. This
version (hope to!) implement the following logic:
1) 1 IRQ: only resource #0 (CARD_DETECT, use tmio_mmc_irq())
2) 2 or 3 IRQs: compulsory SDCARD and any further IRQs: use respective
specialised ISRs.
v5
* As suggested by Guennadi Liakhovetski:
- Allow only SH_MOBILE_SDHI_IRQ_SDCARD and
SH_MOBILE_SDHI_IRQ_SDIO to be specified in platform data
v4
* As suggested by Guennadi Liakhovetski:
- Correct inverted values of SH_MOBILE_SDHI_IRQ_SDCARD and
SH_MOBILE_SDHI_IRQ_CARD_DETECT
v3
* Update for changes to "mmc: tmio: Provide separate interrupt handlers"
* As suggested by Guennadi Liakhovetski:
- Merge in patch "mmc: sdhi: Add defines for platform irq indexes"
- Use an enum instead of defines for irq indexes
v2
* Update for changes to "mmc: tmio: Provide separate interrupt handlers"
* Make use of defines provided by
"mmc: sdhi: Add defines for platform irq indexes"
* As suggested by Guennadi Liakhovetski:
- Don't use a loop to initialise irq handlers, the unrolled version
is easier on the eyes (and exactly the same number of lines of code!)
---
drivers/mmc/host/sh_mobile_sdhi.c | 65 +++++++++++++++++++++++------------
include/linux/mmc/sh_mobile_sdhi.h | 7 ++++
2 files changed, 50 insertions(+), 22 deletions(-)
diff --git a/drivers/mmc/host/sh_mobile_sdhi.c b/drivers/mmc/host/sh_mobile_sdhi.c
index 774f643..2fa7bbc 100644
--- a/drivers/mmc/host/sh_mobile_sdhi.c
+++ b/drivers/mmc/host/sh_mobile_sdhi.c
@@ -96,7 +96,8 @@ static int __devinit sh_mobile_sdhi_probe(struct platform_device *pdev)
struct sh_mobile_sdhi_info *p = pdev->dev.platform_data;
struct tmio_mmc_host *host;
char clk_name[8];
- int i, irq, ret;
+ int irq, ret, irq_count = 0;
+ irqreturn_t (*f)(int irq, void *devid);
priv = kzalloc(sizeof(struct sh_mobile_sdhi), GFP_KERNEL);
if (priv == NULL) {
@@ -153,27 +154,39 @@ static int __devinit sh_mobile_sdhi_probe(struct platform_device *pdev)
if (ret < 0)
goto eprobe;
- for (i = 0; i < 3; i++) {
- irq = platform_get_irq(pdev, i);
- if (irq < 0) {
- if (i) {
- continue;
- } else {
- ret = irq;
- goto eirq;
- }
- }
- ret = request_irq(irq, tmio_mmc_irq, 0,
+ /* Allow a single IRQ resource #0 (CARD_DETECT) which will
+ * use tmio_mmc_irq() or;
+ * Allow 2 or 3 IRQ resources in which case SDCARD is required
+ * and specialised ISRs are used.
+ */
+ irq = platform_get_irq(pdev, SH_MOBILE_SDHI_IRQ_SDCARD);
+ if (irq >= 0) {
+ irq_count++;
+ ret = request_irq(irq, tmio_mmc_sdcard_irq, 0,
dev_name(&pdev->dev), host);
- if (ret) {
- while (i--) {
- irq = platform_get_irq(pdev, i);
- if (irq >= 0)
- free_irq(irq, host);
- }
- goto eirq;
- }
+ if (ret)
+ goto eirq_sdcard;
}
+
+ ret = irq = platform_get_irq(pdev, SH_MOBILE_SDHI_IRQ_SDIO);
+ if (irq_count && irq >= 0) {
+ irq_count++;
+ ret = request_irq(irq, tmio_mmc_sdio_irq, 0,
+ dev_name(&pdev->dev), host);
+ if (ret)
+ goto eirq_sdio;
+ } else if (!irq_count)
+ goto eirq_sdio;
+
+ ret = irq = platform_get_irq(pdev, SH_MOBILE_SDHI_IRQ_CARD_DETECT);
+ if (irq >= 0) {
+ f = irq_count ? tmio_mmc_card_detect_irq : tmio_mmc_irq;
+ ret = request_irq(irq, f, 0, dev_name(&pdev->dev), host);
+ if (ret)
+ goto eirq_card_detect;
+ } else if (irq_count < 2)
+ goto eirq_card_detect;
+
dev_info(&pdev->dev, "%s base at 0x%08lx clock rate %u MHz\n",
mmc_hostname(host->mmc), (unsigned long)
(platform_get_resource(pdev,IORESOURCE_MEM, 0)->start),
@@ -181,7 +194,15 @@ static int __devinit sh_mobile_sdhi_probe(struct platform_device *pdev)
return ret;
-eirq:
+eirq_card_detect:
+ irq = platform_get_irq(pdev, SH_MOBILE_SDHI_IRQ_SDIO);
+ if (irq >= 0)
+ free_irq(irq, host);
+eirq_sdio:
+ irq = platform_get_irq(pdev, SH_MOBILE_SDHI_IRQ_SDCARD);
+ if (irq >= 0)
+ free_irq(irq, host);
+eirq_sdcard:
tmio_mmc_host_remove(host);
eprobe:
clk_disable(priv->clk);
@@ -203,7 +224,7 @@ static int sh_mobile_sdhi_remove(struct platform_device *pdev)
tmio_mmc_host_remove(host);
- for (i = 0; i < 3; i++) {
+ for (i = 0; i < SH_MOBILE_SDHI_IRQ_MAX; i++) {
irq = platform_get_irq(pdev, i);
if (irq >= 0)
free_irq(irq, host);
diff --git a/include/linux/mmc/sh_mobile_sdhi.h b/include/linux/mmc/sh_mobile_sdhi.h
index bd50b36..80d3629 100644
--- a/include/linux/mmc/sh_mobile_sdhi.h
+++ b/include/linux/mmc/sh_mobile_sdhi.h
@@ -3,6 +3,13 @@
#include <linux/types.h>
+enum {
+ SH_MOBILE_SDHI_IRQ_CARD_DETECT = 0,
+ SH_MOBILE_SDHI_IRQ_SDCARD,
+ SH_MOBILE_SDHI_IRQ_SDIO,
+ SH_MOBILE_SDHI_IRQ_MAX
+};
+
struct platform_device;
struct tmio_mmc_data;
--
1.7.5.4
next prev parent reply other threads:[~2011-08-17 12:57 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-08-17 10:59 [PATCH 0/4 v5] mmc: tmio, sdhi: provide multiple irq handlers Simon Horman
2011-08-17 10:59 ` [PATCH 1/4] mmc: tmio: Cache interrupt masks Simon Horman
2011-08-17 10:59 ` [PATCH 2/4] mmc: tmio: Provide separate interrupt handlers Simon Horman
2011-08-17 10:59 ` [PATCH 3/4] mmc: sdhi: Make use of per-source irq handlers Simon Horman
2011-08-17 11:24 ` Guennadi Liakhovetski
2011-08-17 11:41 ` Simon Horman
2011-08-17 11:54 ` Guennadi Liakhovetski
2011-08-17 12:07 ` Simon Horman
2011-08-17 12:22 ` Guennadi Liakhovetski
2011-08-17 12:57 ` Simon Horman [this message]
2011-08-17 22:02 ` [PATCH 3/4 v7] " Guennadi Liakhovetski
2011-08-17 10:59 ` [PATCH 4/4] ARM: shmobile: ag5evm, ap4: Make use of irq index enum Simon Horman
-- strict thread matches above, loose matches on Subject: below --
2011-08-19 1:10 [PATCH 0/4 v6] mmc: tmio, sdhi: provide multiple irq handlers Simon Horman
2011-08-19 1:10 ` [PATCH 3/4] mmc: sdhi: Make use of per-source " Simon Horman
2011-08-17 0:50 [PATCH 0/4 v4] mmc: tmio, sdhi: provide multiple " Simon Horman
2011-08-17 0:50 ` [PATCH 3/4] mmc: sdhi: Make use of per-source " Simon Horman
2011-08-17 8:20 ` Guennadi Liakhovetski
2011-08-17 9:49 ` Simon Horman
2011-08-17 10:06 ` Guennadi Liakhovetski
2011-08-17 10:46 ` Simon Horman
2011-08-16 10:11 [PATCH 0/4 v3] mmc: tmio, sdhi: provide multiple " Simon Horman
2011-08-16 10:11 ` [PATCH 3/4] mmc: sdhi: Make use of per-source " Simon Horman
2011-08-16 11:11 ` Guennadi Liakhovetski
2011-08-16 11:51 ` Simon Horman
2011-08-16 12:40 ` Guennadi Liakhovetski
2011-08-16 13:45 ` Simon Horman
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=20110817125728.GD13086@verge.net.au \
--to=horms@verge.net.au \
--cc=cjb@laptop.org \
--cc=g.liakhovetski@gmx.de \
--cc=lethal@linux-sh.org \
--cc=linux-mmc@vger.kernel.org \
--cc=linux-sh@vger.kernel.org \
--cc=magnus.damm@gmail.com \
/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