Hello,
After Linux enables an OMAP device's clocks, it must wait for the
device to leave the 'idle' state, or register accesses may result in
an interconnect error and subsequent system crash. Code to do this is
currently in the OMAP clock code, but is overgeneralized, and will not
work correctly for several exception cases - one on 2430, and four on
OMAP3.
This series fixes these bugs by modularizing omap2_clk_wait_ready(),
allowing custom CM_IDLEST register and bit shift data to be supplied
on a per-clock basis, if necessary. (This was originally fixed in the
linux-omap kernel during 2.6.29 in a slightly different manner.)
In the medium term, the module wait code will be moved to omap_hwmod or a
similar OMAP device layer.
Tested on BeagleBoard rev C2 (OMAP3530 ES3); compile-tested with the
N800 defconfig.
This second version fills in the clock.c documentation and removes one
of the custom find_idlest functions (since DSS and USBHOST use the same
target IDLEST shift).
- Paul
---
size:
text data bss dec hex filename
3405015 191408 103632 3700055 387557 vmlinux.beagle.orig
3405475 192144 103632 3701251 387a03 vmlinux.beagle
Paul Walmsley (3):
OMAP3 clock: correct module IDLEST bits: SSI; DSS; USBHOST; HSOTGUSB
OMAP2 clock: 2430 I2CHS uses non-standard CM_IDLEST register
OMAP2/3 clock: split, rename omap2_wait_clock_ready()
arch/arm/mach-omap2/clock.c | 156 ++++++++++++++++---------------
arch/arm/mach-omap2/clock.h | 6 +
arch/arm/mach-omap2/clock24xx.c | 37 +++++++
arch/arm/mach-omap2/clock24xx.h | 4 -
arch/arm/mach-omap2/clock34xx.c | 118 ++++++++++++++++++++++-
arch/arm/mach-omap2/clock34xx.h | 85 +++++++++++++++--
arch/arm/mach-omap2/prcm.c | 43 +++++++++
arch/arm/plat-omap/include/mach/clock.h | 2
arch/arm/plat-omap/include/mach/prcm.h | 1
9 files changed, 353 insertions(+), 99 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
