| Message ID | alpine.DEB.2.00.1302120358130.14983@utopia.booyaka.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
* Paul Walmsley <paul@pwsan.com> [130211 20:02]: > > Commit c1d1cd597fc77af3086470f8627d77f52f7f8b6c ("ARM: OMAP2+: > omap_device: remove obsolete pm_lats and early_device code") missed a > few omap_device_build() calls that aren't included as part of the default > OMAP2+ Kconfig, omap2plus_defconfig. > > Ideally, all devices that are present on the SoC should be created by > default, and only the corresponding device driver should be configured > or deconfigured in Kconfig. This allows drivers to be built as > modules and loaded later, even if they weren't part of the original > kernel build. Unfortunately, we're not quite there yet. > > Thanks to Tony Lindgren for reporting this, found during his > randconfig tests. Thanks, I'll apply this into omap-for-v3.9/pm-wfi as the breaking commit is already in arm-soc/for-next. Tony
On Tue, 12 Feb 2013, Tony Lindgren wrote: > Thanks, I'll apply this into omap-for-v3.9/pm-wfi as the breaking > commit is already in arm-soc/for-next. OK thanks. Looks like I'll have one more fix for one of the commits in arm-soc/for-next; working on it now. - Paul
diff --git a/arch/arm/mach-omap2/am35xx-emac.c b/arch/arm/mach-omap2/am35xx-emac.c index a00d391..25b79a2 100644 --- a/arch/arm/mach-omap2/am35xx-emac.c +++ b/arch/arm/mach-omap2/am35xx-emac.c @@ -62,8 +62,7 @@ static int __init omap_davinci_emac_dev_init(struct omap_hwmod *oh, { struct platform_device *pdev; - pdev = omap_device_build(oh->class->name, 0, oh, pdata, pdata_len, - false); + pdev = omap_device_build(oh->class->name, 0, oh, pdata, pdata_len); if (IS_ERR(pdev)) { WARN(1, "Can't build omap_device for %s:%s.\n", oh->class->name, oh->name); diff --git a/arch/arm/mach-omap2/devices.c b/arch/arm/mach-omap2/devices.c index d8a0cc3..6ecc89a 100644 --- a/arch/arm/mach-omap2/devices.c +++ b/arch/arm/mach-omap2/devices.c @@ -382,7 +382,7 @@ static void __init omap_init_hdmi_audio(void) return; } - pdev = omap_device_build("omap-hdmi-audio-dai", -1, oh, NULL, 0, 0); + pdev = omap_device_build("omap-hdmi-audio-dai", -1, oh, NULL, 0); WARN(IS_ERR(pdev), "Can't build omap_device for omap-hdmi-audio-dai.\n"); diff --git a/arch/arm/mach-omap2/sr_device.c b/arch/arm/mach-omap2/sr_device.c index bb829e0..d7bc33f 100644 --- a/arch/arm/mach-omap2/sr_device.c +++ b/arch/arm/mach-omap2/sr_device.c @@ -152,7 +152,7 @@ static int __init sr_dev_init(struct omap_hwmod *oh, void *user) sr_data->enable_on_init = sr_enable_on_init; - pdev = omap_device_build(name, i, oh, sr_data, sizeof(*sr_data), 0); + pdev = omap_device_build(name, i, oh, sr_data, sizeof(*sr_data)); if (IS_ERR(pdev)) pr_warning("%s: Could not build omap_device for %s: %s.\n\n", __func__, name, oh->name);
Commit c1d1cd597fc77af3086470f8627d77f52f7f8b6c ("ARM: OMAP2+: omap_device: remove obsolete pm_lats and early_device code") missed a few omap_device_build() calls that aren't included as part of the default OMAP2+ Kconfig, omap2plus_defconfig. Ideally, all devices that are present on the SoC should be created by default, and only the corresponding device driver should be configured or deconfigured in Kconfig. This allows drivers to be built as modules and loaded later, even if they weren't part of the original kernel build. Unfortunately, we're not quite there yet. Thanks to Tony Lindgren for reporting this, found during his randconfig tests. Signed-off-by: Paul Walmsley <paul@pwsan.com> Cc: Tony Lindgren <tony@atomide.com> --- arch/arm/mach-omap2/am35xx-emac.c | 3 +-- arch/arm/mach-omap2/devices.c | 2 +- arch/arm/mach-omap2/sr_device.c | 2 +- 3 files changed, 3 insertions(+), 4 deletions(-)