[PATCH 1/5] spi: spi-mem: Add support for panic mem writes

[Paul Cercueil]

Hi Miquèl,

Le mardi 13 octobre 2026 à 10:38 +0200, Miquel Raynal a écrit :
> On 13/10/2026 at 10:25:07 +02, Paul Cercueil <paul.cercueil at bootlin.com>
> wrote:
> 
> > Hi Mark,
> > 
> > Le jeudi 08 octobre 2026 à 10:32 +0200, Mark Brown a écrit :
> > > On Mon, Oct 05, 2026 at 10:11:36AM +0200, Paul Cercueil wrote:
> > > > From: Tanmay Jagdale <tanmay at marvell.com>
> > > > 
> > > > Add a 'panic_write' flag to the controller's capabilities struct.
> > > > When set, the SPI controller indicates that it will be able to send
> > > > commands to a connected NOR or NAND memory in the context of a kernel
> > > > panic, which means without sleeping and without relying on IRQs.
> > > 
> > > > --- a/drivers/spi/spi-mem.c
> > > > +++ b/drivers/spi/spi-mem.c
> > > > @@ -398,6 +398,31 @@ int spi_mem_exec_op(struct spi_mem *mem, const
> > > > struct
> > > > spi_mem_op *op)
> > > >  	u8 *tmpbuf;
> > > >  	int ret;
> > > >  
> > > > +	if (op->panic_write) {
> > > > +		if (!spi_mem_controller_is_capable(ctlr, panic_write))
> > > > +			return -EOPNOTSUPP;
> > > > +		/*
> > > > +		 * Only support panic writes if we'll go through
> > > > .exec_op().
> > > > +		 * The regular SPI path (spi_sync(), etc.) explicitly
> > > > mentions
> > > > +		 * that it can only be used from a context that may
> > > > sleep,
> > > > which
> > > > +		 * is not the case on a kernel panic.
> > > > +		 * If this changes in the future, this could be
> > > > reworked,
> > > > but
> > > > +		 * for now this will have to do.
> > > > +		 */
> > > 
> > > This also needs to ensure that the queue is idle - we'll still execute
> > > spi_mem_access_start() which flushes the queue, and we've got runtime PM
> > > operations there which might be a bit challenging in a panic situation.
> > 
> > Does it though? The kernel is panicking so the queue won't be processed
> > anymore,
> > it should be fine to just ignore it.
> > 
> > ACK for the PM - I think in the V2 I'll just bail if the controller has
> > auto-
> > runtime-PM on.
> 
> Maybe pm_runtime_get_if_active()? If already active it returns 1 and
> you can proceed.

Yes, good idea.

Cheers,
-Paul