On Sunday 17 April 2016 21:55:39 Stephen Boyd wrote:
On 04/17, Arnd Bergmann wrote:
On Saturday 16 April 2016 17:06:15 Olof's autobuilder wrote:
4 drivers/clk/clk-vt8500.c:428:14: warning: 'best_mul' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:429:12: warning: 'best_div1' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:430:12: warning: 'best_div2' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:507:14: warning: 'best_mul' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:508:12: warning: 'best_div1' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:509:12: warning: 'best_div2' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:558:14: warning: 'best_mul' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:559:12: warning: 'best_div1' may be used uninitialized in this function [-Wmaybe-uninitialized] 4 drivers/clk/clk-vt8500.c:560:12: warning: 'best_div2' may be used uninitialized in this function [-Wmaybe-uninitialized]
My patch is in linux-next, but didn't make it into mainline so far.
Stephen, could you send this as a bugfix for the next -rc?
I'm not sure which patch you're talking about. I thought we already merged commit 7001ec560a82 (clk: vt8500: don't return possibly uninitialized data, 2016-02-01). It's in v4.6-rc1.
From what I recall the warning went away with your patch on your compiler but doesn't go away on older versions of GCC. If there's another patch I'm not aware.
Sorry, your memory is better than mine, I completely forgot that we had this discussion before and that I asked for it to be done exactly like you did.
Everything is fine, except that the autobuilders don't use the same gcc version that I have.
Arnd