Re: [PATCH] mtd: rawnand: fsmc: Keep CE enabled fix mb() drain

[Boris Brezillon <bbrezillon@kernel.org>]

On Wed,  9 Jan 2019 21:55:30 +0100
Linus Walleij <linus.walleij@linaro.org> wrote:

> Hammering the chip enable on and off between every command
> crashes the Nomadik NHK15 with this message:
> 
> Scanning device for bad blocks
> Unhandled fault: external abort on non-linefetch (0x008) at 0xcc95e000
> pgd = (ptrval)
> [cc95e000] *pgd=0b808811, *pte=40000653, *ppte=40000552
> Internal error: : 8 [#1] PREEMPT ARM
> Modules linked in:
> CPU: 0 PID: 1 Comm: swapper Not tainted 4.20.0-rc2+ #72
> Hardware name: Nomadik STn8815
> PC is at fsmc_exec_op+0x194/0x204
> (...)
> 
> This patch keeps the CE (chip enable, the only chip select)
> signal from the FSMC block enabled from the first command
> after probe() or resume() until the driver either suspend()
> or remove(). Create a state variable to track this.

I just read the Spear600 reference manual, and I'm not sure the
BANK_ENABLE bit controls the CE line. My understanding is that it just
marks the bank as active and CE line is asserted when you actually
access the AHB mem bank range (probably after making sure the FSMC bus
is idle).

If I'm correct, I'd recommend dropping fsmc_ce_ctrl() and marking the
bank enabled at probe time.

> 
> Runtime PM can possibly be used on top of this approach if
> we want to disable the CE line when no flash read/write is
> going on for a prolonged time, but for now it stabilizes
> the platform by simply keeping CE asserted as it used to
> be.
> 
> While we're at it also fix the mb() memory barrier drain
> order (patch folded in from Boris Brezillion).

Let's move that to a separate patch (but I'm not even sure it will
survive if we go for the "enable bank at probe time" approach).