summaryrefslogtreecommitdiffstats
path: root/arch/arm/mach-gemini
diff options
context:
space:
mode:
authorRajendra Nayak <rnayak@ti.com>2013-03-31 20:22:23 -0600
committerPaul Walmsley <paul@pwsan.com>2013-03-31 20:43:06 -0600
commit469d633d20c774ecd34ac615c838193e1e150c62 (patch)
tree3a7688f595a54562e1099fcede018fe5df6fa6d3 /arch/arm/mach-gemini
parentbd70f6eb3e78b7c9c6bec2cf5be6423dd2b37f6f (diff)
downloadblackbird-op-linux-469d633d20c774ecd34ac615c838193e1e150c62.tar.gz
blackbird-op-linux-469d633d20c774ecd34ac615c838193e1e150c62.zip
ARM: OMAP: dpll: enable bypass clock only when attempting dpll bypass
omap3_noncore_dpll_set_rate() attempts an enable of bypass clk as well as ref clk for every .set_rate attempt on a noncore DPLL, regardless of whether the .set_rate results in the DPLL being locked or put in bypass. Early at boot, while some of these DPLLs are programmed and locked (using .set_rate for the DPLL), this causes an ordering issue. For instance, on OMAP5, the USB DPLL derives its bypass clk from ABE DPLL. If a .set_rate of USB DPLL which programmes the M,N and locks it is called before the one for ABE, the enable of USB bypass clk (derived from ABE DPLL) then attempts to lock the ABE DPLL and fails as the M,N values for ABE are yet to be programmed. To get rid of this ordering needs, enable bypass clk for a DPLL as part of its .set_rate only when its being put in bypass, and only enable the ref clk when its locked. Reported-by: Roger Quadros <rogerq@ti.com> Signed-off-by: Rajendra Nayak <rnayak@ti.com> Signed-off-by: Paul Walmsley <paul@pwsan.com>
Diffstat (limited to 'arch/arm/mach-gemini')
0 files changed, 0 insertions, 0 deletions
OpenPOWER on IntegriCloud