Received-SPF: pass (google.com: domain of linux-kernel+bounces-58021-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Precedence: bulk
MIME-Version: 1.0
References: <vbpzr7uqpfemb3qa6xy2fxioct44l5vugg2wkywyolfpzqcmau@jgrrhmk2scaj>
 <7559d940-f191-4fe0-e147-17ffa6c1dfc4@gmail.com> <CA+PODjpOE=LGPi1G1ebvEwGeXAfpuZ+s_k4uMUwu3i6st9y--g@mail.gmail.com>
 <ZcSSzDI2qL3LSQvr@x1-carbon>
In-Reply-To: <ZcSSzDI2qL3LSQvr@x1-carbon>
From: Andrey Melnikov <temnota.am@gmail.com>
Date: Thu, 8 Feb 2024 15:04:33 +0300
Message-ID: <CA+PODjqTwko0+2cw227OpcLzP7DD3CouU26jQ=j-MirSHFEGEg@mail.gmail.com>
Subject: Re: [PATCH] ahci: asm1064: correct count of reported ports
To: Niklas Cassel <cassel@kernel.org>
Cc: Serge Semin <fancer.lancer@gmail.com>, Sergei Shtylyov <sergei.shtylyov@gmail.com>, 
	linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org, dlemoal@kernel.org, 
	hdegoede@redhat.com
Content-Type: text/plain; charset="UTF-8"

Hello.

> Hello Serge, Andrey,
>
> On Thu, Feb 08, 2024 at 10:27:11AM +0300, Andrey Melnikov wrote:
> > > On 2/7/24 12:58 PM, Andrey Jr. Melnikov wrote:
> > >
> > > > The ASM1064 SATA host controller always reports wrongly,
> > > > that it has 24 ports. But in reality, it only has four ports.
> > > >
> > > > before:
> > > > ahci 0000:04:00.0: SSS flag set, parallel bus scan disabled
> > > > ahci 0000:04:00.0: AHCI 0001.0301 32 slots 24 ports 6 Gbps 0xffff0f impl SATA mode
> > > > ahci 0000:04:00.0: flags: 64bit ncq sntf stag pm led only pio sxs deso sadm sds apst
> > > >
> > > > after:
> > > > ahci 0000:04:00.0: ASM1064 has only four ports
> > > > ahci 0000:04:00.0: forcing port_map 0xffff0f -> 0xf
> > > > ahci 0000:04:00.0: SSS flag set, parallel bus scan disabled
> > > > ahci 0000:04:00.0: AHCI 0001.0301 32 slots 24 ports 6 Gbps 0xf impl SATA mode
> > > > ahci 0000:04:00.0: flags: 64bit ncq sntf stag pm led only pio sxs deso sadm sds apst
> > > >
> > > >
> > > > Signed-off-by: Andrey Jr. Melnikov <temnota.am@gmail.com>
> > > >
> > > > diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c
> > > > index da2e74fce2d9..ec30d8330d16 100644
> > > > --- a/drivers/ata/ahci.c
> > > > +++ b/drivers/ata/ahci.c
> > > > @@ -671,9 +671,14 @@ MODULE_PARM_DESC(mobile_lpm_policy, "Default LPM policy for mobile chipsets");
> > > >  static void ahci_pci_save_initial_config(struct pci_dev *pdev,
> > > >                                        struct ahci_host_priv *hpriv)
> > > >  {
> > > > -     if (pdev->vendor == PCI_VENDOR_ID_ASMEDIA && pdev->device == 0x1166) {
> > > > -             dev_info(&pdev->dev, "ASM1166 has only six ports\n");
> > > > -             hpriv->saved_port_map = 0x3f;
> > > > +     if (pdev->vendor == PCI_VENDOR_ID_ASMEDIA) {
> > > > +             if (pdev->device == 0x1166) {
> > >
> > >    Maybe *switch* instead?
> >
> > Ok.
> > Can someone explain to me - which method I should use here - override
> > portmap via hpriv->saved_port_map or mask it via hpriv->mask_port_map
> > ?
>
> Serge, you seem to be the last person to touch this in
> commit 88589772e80c ("ata: libahci: Discard redundant force_port_map
> parameter"), where you removed force_port_map, so you are already
> familiar with this code.
>
> From my quick look, it looks like mask_port_map is used to disable one or
> more ports from the port_map.
>
> But do you know if there is a reason why platforms that need to do that
> can't just change saved_port map directly?

Well, after deep dive into git history - initially, overriding
port_map introduced by commits
d799e083a80b220f3681d7790f11e77d1704022b (jmb361 workaround) and
cd70c26617f4686355263be4533ce8030242740e (marvel 6145 support), both
override port mapping, but in different ways - first set to 1, second
- apply masks. Logically - port_map = 1 and port_map &= 1 - are the
same.
Later, in commit 394d6e535f15c6f2d3c7fe2e228ee595acf0648c this two
hacks moved around, and setting port_map splitted into two ways - set
it directly via force_port_map (jmb361 workaround) and apply mask via
mask_port_map (marvel 6145 hack).
And now, we have two knobs which do the same! After that, commit
566d1827df2ef0cbe921d3d6946ac3007b1a6938 add checking AHCI version
(yet another SATA glitches workaround, user report [1] shows 8 port
and bitmask 0xff - 32 ports)  and this break updating saved_port_map
for AHCI >= 1.3. Later, commit
2fd0f46cb1b82587c7ae4a616d69057fb9bd0af7 fix saving port_map again,
and now overriding port_map completely splitted. saved_port_map is
really saved (and written back to HOST_PORTS_IMPL register on
saved_port map()) for all AHCI versions, mask_port_map is applied and
not updated saved_port_map.

1. https://lists.openwall.net/linux-kernel/2016/01/13/171