Re: [PATCH v4 12/23] ata: libahci: Extend port-cmd flags set with port capabilities

[Serge Semin]

On Sat, Jun 18, 2022 at 03:52:28PM +0900, Damien Le Moal wrote:
> On 6/18/22 05:31, Serge Semin wrote:
> > On Thu, Jun 16, 2022 at 09:28:18AM +0900, Damien Le Moal wrote:
> >> On 2022/06/16 5:58, Serge Semin wrote:
> >>> On Tue, Jun 14, 2022 at 05:32:41PM +0900, Damien Le Moal wrote:
> >>>> On 6/10/22 17:17, Serge Semin wrote:
> >>>>> Currently not all of the Port-specific capabilities listed in the
> >>>>
> >>>> s/listed/are listed
> >>>>
> >>>>> PORT_CMD-enumeration. Let's extend that set with the Cold Presence
> >>>>> Detection and Mechanical Presence Switch attached to the Port flags [1] so
> >>>>> to closeup the set of the platform-specific port-capabilities flags.  Note
> >>>>> these flags are supposed to be set by the platform firmware if there is
> >>>>> one. Alternatively as we are about to do they can be set by means of the
> >>>>> OF properties.
> >>>>>
> >>>>> While at it replace PORT_IRQ_DEV_ILCK with PORT_IRQ_DMPS and fix the
> >>>>> comment there. In accordance with [2] that IRQ flag is supposed to
> >>>>> indicate the state of the signal coming from the Mechanical Presence
> >>>>> Switch.
> >>>>>
> >>>>> [1] Serial ATA AHCI 1.3.1 Specification, p.27
> >>>>> [2] Serial ATA AHCI 1.3.1 Specification, p.24, p.88
> >>>>>
> >>>>> Signed-off-by: Serge Semin <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx>
> >>>>> Reviewed-by: Hannes Reinecke <hare@xxxxxxx>
> >>>>>
> >>>>> ---
> >>>>>
> >>>>> Changelog v4:
> >>>>> - Fix the DMPS macros name in the patch log. (@Sergei Shtylyov)
> >>>>> ---
> >>>>>  drivers/ata/ahci.h | 7 ++++++-
> >>>>>  1 file changed, 6 insertions(+), 1 deletion(-)
> >>>>>
> >>>>> diff --git a/drivers/ata/ahci.h b/drivers/ata/ahci.h
> >>>>> index 7d834deefeb9..f501531bd1b3 100644
> >>>>> --- a/drivers/ata/ahci.h
> >>>>> +++ b/drivers/ata/ahci.h
> >>>>> @@ -138,7 +138,7 @@ enum {
> >>>>>  	PORT_IRQ_BAD_PMP	= (1 << 23), /* incorrect port multiplier */
> >>>>>  
> >>>>>  	PORT_IRQ_PHYRDY		= (1 << 22), /* PhyRdy changed */
> >>>>> -	PORT_IRQ_DEV_ILCK	= (1 << 7), /* device interlock */
> >>>>> +	PORT_IRQ_DMPS		= (1 << 7), /* mechanical presence status */
> >>>>>  	PORT_IRQ_CONNECT	= (1 << 6), /* port connect change status */
> >>>>>  	PORT_IRQ_SG_DONE	= (1 << 5), /* descriptor processed */
> >>>>>  	PORT_IRQ_UNK_FIS	= (1 << 4), /* unknown FIS rx'd */
> >>>>> @@ -166,6 +166,8 @@ enum {
> >>>>>  	PORT_CMD_ATAPI		= (1 << 24), /* Device is ATAPI */
> >>>>>  	PORT_CMD_FBSCP		= (1 << 22), /* FBS Capable Port */
> >>>>>  	PORT_CMD_ESP		= (1 << 21), /* External Sata Port */
> >>>>> +	PORT_CMD_CPD		= (1 << 20), /* Cold Presence Detection */
> >>>>> +	PORT_CMD_MPSP		= (1 << 19), /* Mechanical Presence Switch */
> >>>>>  	PORT_CMD_HPCP		= (1 << 18), /* HotPlug Capable Port */
> >>>>>  	PORT_CMD_PMP		= (1 << 17), /* PMP attached */
> >>>>>  	PORT_CMD_LIST_ON	= (1 << 15), /* cmd list DMA engine running */
> >>>>> @@ -181,6 +183,9 @@ enum {
> >>>>>  	PORT_CMD_ICC_PARTIAL	= (0x2 << 28), /* Put i/f in partial state */
> >>>>>  	PORT_CMD_ICC_SLUMBER	= (0x6 << 28), /* Put i/f in slumber state */
> >>>>>  
> >>>>> +	PORT_CMD_CAP		= PORT_CMD_HPCP | PORT_CMD_MPSP |
> >>>>> +				  PORT_CMD_CPD | PORT_CMD_ESP | PORT_CMD_FBSCP,
> >>>>
> >>>
> >>>> What is this one for ? A comment above it would be nice.
> >>>
> >>> Isn't it obviously inferrable from the definition and the item name?
> >>
> > 
> >> I am guessing from the name. Am I guessing OK ? A comment would still be nice.
> >> Why just these bits ? There are more cap/support indicator bits in that port cmd
> >> bitfield. So why this particular set of bits ? What do they mean all together ?
> > 
> > Normally the variable/constant name should be self-content (as the
> > kernel coding style doc states and what the common sense suggests). So
> > the reader could correctly guess its purpose/content/value. In this
> > case PORT_CMD_CAP - means PORT CMD capabilities mask. All of the
> > possible flags have been set in that mask. There are no more
> > capabilities in the PORT CMD register left undeclared. That's why the
> > name is selected the way it is and why I haven't added any comment in
> > here (what the kernel coding style says about the over-commenting the
> > code).
> 

> Yes, I understood from the name what it is. What I do NOT understand is
> why all the feature bits are not there. Why this subset only ? A comment
> about that would be nice so that the reason for it is not lost.

Well, because it's indeed "PORT_CMD capabilities mask", and not features,
not setups, not settings, not status flags, etc. As I said all the port
Capabilities have been listed in that mask:
PORT_CMD_FBSCP	BIT(22) - FIS-based Switching Capable Port
PORT_CMD_ESP	BIT(21) - External SATA Port
PORT_CMD_CPD	BIT(20) - Cold Presence Detect
PORT_CMD_MPSP	BIT(19) - Mechanical Presence Switch Attached to Port
PORT_CMD_HPCP	BIT(18) - Hot Plug Capable Port
I've or'ed-them-up in a single mask => PORT_CMD_CAP in order to work
with them independently from the rest of the PORT_CMD CSR fields.

Unlike the generic controller CAP/CAP2 registers, which consists of the
device capabilities only, PORT_CMD contains various R/W settings (PM, LED
driver, etc), RO status flags (CMD-list running, FIS recv running, etc)
and amongst other the RO/Wo !port-specific capabilities!. The later ones
indicate the platform-specific device features. Since the register
contains flags with the intermixed nature, I need to have a mask to at
least get the capabilities and preserve them between the device
resets. That's why the PORT_CMD_CAP has been introduced in the
framework of this patch. Its name was chosen with a reference to the
CAP registers, see:
HOST_CAP, HOST_CAP2, and finally my PORT_CMD_CAP.

> 
> > 
> >>
> >> Sure I can go and read the specs to figure it out. But again, a comment would
> >> avoid readers of the code to have to decrypt all that.
> > 
> > If you still insist on having an additional comment. I can add
> > something like "/* PORT_CMD capabilities mask */". Are you ok with it?
> 

> That does not help on its own. The macro name says that already. I would
> like a note about why only these features are selected.

Please see the explanation above. I don't see what else to say about
that mask, because in short what I said above really means "PORT_CMD
capabilities mask". So should you have some more clever text, which
would be more suitable here, please tell me and I'll add it to the
patch.

Regarding what you said earlier. In order to fully understand the
AHCI driver a hacker would always need to read the specs. There is
just no way to do that effectively enough without the controller
manual at hands. And the PORT_CMD capabilities isn't the most
complicated part of the device.

-Sergey

> 
> > 
> > -Sergey
> > 
> >>
> >>>
> >>> -Sergey
> >>>
> >>>>
> >>>>> +
> >>>>>  	/* PORT_FBS bits */
> >>>>>  	PORT_FBS_DWE_OFFSET	= 16, /* FBS device with error offset */
> >>>>>  	PORT_FBS_ADO_OFFSET	= 12, /* FBS active dev optimization offset */
> >>>>
> >>>>
> >>>> -- 
> >>>> Damien Le Moal
> >>>> Western Digital Research
> >>
> >>
> >> -- 
> >> Damien Le Moal
> >> Western Digital Research
> 
> 
> -- 
> Damien Le Moal
> Western Digital Research