[PATCH v2] ata: libahci_platform: Do not set mask_port_map when not needed

[PATCH v2] ata: libahci_platform: Do not set mask_port_map when not needed

[Damien Le Moal]

Commit 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive
port numbers") modified ahci_platform_get_resources() to allow
identifying the ports of a controller that are defined as child nodes of
the controller node in order to support non-consecutive port numbers (as
defined by the platform device tree).

However, this commit also erroneously sets bit 0 of
hpriv->mask_port_map when the platform devices tree does not define port
child nodes, to match the fact that the temporary default number of
ports used in that case is 1 (which is also consistent with the fact
that only index 0 of hpriv->phys[] is initialized with the call to
ahci_platform_get_phy(). But doing so causes ahci_platform_init_host()
to initialize and probe only the first port, even if this function
determines that the controller has in fact multiple ports using the
capability register of the controller (through a call to
ahci_nr_ports()). This can be seen with the ahci_mvebu driver (Armada
385 SoC) with the second port declared as "dummy":

ahci-mvebu f10a8000.sata: masking port_map 0x3 -> 0x1
ahci-mvebu f10a8000.sata: AHCI vers 0001.0000, 32 command slots, 6 Gbps, platform mode
ahci-mvebu f10a8000.sata: 1/2 ports implemented (port mask 0x1)
ahci-mvebu f10a8000.sata: flags: 64bit ncq sntf led only pmp fbs pio slum part sxs
scsi host0: ahci-mvebu
scsi host1: ahci-mvebu
ata1: SATA max UDMA/133 mmio [mem 0xf10a8000-0xf10a9fff] port 0x100 irq 40 lpm-pol 0
ata2: DUMMY

Fix this issue by removing setting bit 0 of hpriv->mask_port_map when
the platform device tree does not define port child nodes.

Reported-by: Klaus Kudielka <klaus.kudielka@gmail.com>
Fixes: 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive port numbers")
Tested-by: Klaus Kudielka <klaus.kudielka@gmail.com>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
---
Changes from v1:
 - Improved commit message (fixed typos, added example and removed
   cc-stable tag).

 drivers/ata/libahci_platform.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/ata/libahci_platform.c b/drivers/ata/libahci_platform.c
index 53b2c7719dc5..91d44302eac9 100644
--- a/drivers/ata/libahci_platform.c
+++ b/drivers/ata/libahci_platform.c
@@ -651,8 +651,6 @@ struct ahci_host_priv *ahci_platform_get_resources(struct platform_device *pdev,
 		 * If no sub-node was found, keep this for device tree
 		 * compatibility
 		 */
-		hpriv->mask_port_map |= BIT(0);
-
 		rc = ahci_platform_get_phy(hpriv, 0, dev, dev->of_node);
 		if (rc)
 			goto err_out;
-- 
2.48.1

Re: [PATCH v2] ata: libahci_platform: Do not set mask_port_map when not needed

[Josua Mayer]

Am 08.02.25 um 00:29 schrieb Damien Le Moal:
> Commit 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive
> port numbers") modified ahci_platform_get_resources() to allow
> identifying the ports of a controller that are defined as child nodes of
> the controller node in order to support non-consecutive port numbers (as
> defined by the platform device tree).
>
> However, this commit also erroneously sets bit 0 of
> hpriv->mask_port_map when the platform devices tree does not define port
> child nodes, to match the fact that the temporary default number of
> ports used in that case is 1 (which is also consistent with the fact
> that only index 0 of hpriv->phys[] is initialized with the call to
> ahci_platform_get_phy(). But doing so causes ahci_platform_init_host()
> to initialize and probe only the first port, even if this function
> determines that the controller has in fact multiple ports using the
> capability register of the controller (through a call to
> ahci_nr_ports()). This can be seen with the ahci_mvebu driver (Armada
> 385 SoC) with the second port declared as "dummy":
>
> ahci-mvebu f10a8000.sata: masking port_map 0x3 -> 0x1
> ahci-mvebu f10a8000.sata: AHCI vers 0001.0000, 32 command slots, 6 Gbps, platform mode
> ahci-mvebu f10a8000.sata: 1/2 ports implemented (port mask 0x1)
> ahci-mvebu f10a8000.sata: flags: 64bit ncq sntf led only pmp fbs pio slum part sxs
> scsi host0: ahci-mvebu
> scsi host1: ahci-mvebu
> ata1: SATA max UDMA/133 mmio [mem 0xf10a8000-0xf10a9fff] port 0x100 irq 40 lpm-pol 0
> ata2: DUMMY
>
> Fix this issue by removing setting bit 0 of hpriv->mask_port_map when
> the platform device tree does not define port child nodes.
>
> Reported-by: Klaus Kudielka <klaus.kudielka@gmail.com>
> Fixes: 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive port numbers")
> Tested-by: Klaus Kudielka <klaus.kudielka@gmail.com>
> Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
> ---
> Changes from v1:
>  - Improved commit message (fixed typos, added example and removed
>    cc-stable tag).
>
>  drivers/ata/libahci_platform.c | 2 --
>  1 file changed, 2 deletions(-)
>
> diff --git a/drivers/ata/libahci_platform.c b/drivers/ata/libahci_platform.c
> index 53b2c7719dc5..91d44302eac9 100644
> --- a/drivers/ata/libahci_platform.c
> +++ b/drivers/ata/libahci_platform.c
> @@ -651,8 +651,6 @@ struct ahci_host_priv *ahci_platform_get_resources(struct platform_device *pdev,
>  		 * If no sub-node was found, keep this for device tree
>  		 * compatibility
>  		 */
> -		hpriv->mask_port_map |= BIT(0);
> -
>  		rc = ahci_platform_get_phy(hpriv, 0, dev, dev->of_node);
>  		if (rc)
>  			goto err_out;

After testing on Armada 388 Helios-4 board with and without port subnodes,
I can confirm that before my patch and without port subnodes
both controller ports were fully operational.

Therefore I believe restoring the original behaviour here is correct.

Acked-by: Josua Mayer <josua@solid-run.com>

Re: [PATCH v2] ata: libahci_platform: Do not set mask_port_map when not needed

[Niklas Cassel]

On Sat, 08 Feb 2025 08:29:15 +0900, Damien Le Moal wrote:
> Commit 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive
> port numbers") modified ahci_platform_get_resources() to allow
> identifying the ports of a controller that are defined as child nodes of
> the controller node in order to support non-consecutive port numbers (as
> defined by the platform device tree).
> 
> However, this commit also erroneously sets bit 0 of
> hpriv->mask_port_map when the platform devices tree does not define port
> child nodes, to match the fact that the temporary default number of
> ports used in that case is 1 (which is also consistent with the fact
> that only index 0 of hpriv->phys[] is initialized with the call to
> ahci_platform_get_phy(). But doing so causes ahci_platform_init_host()
> to initialize and probe only the first port, even if this function
> determines that the controller has in fact multiple ports using the
> capability register of the controller (through a call to
> ahci_nr_ports()). This can be seen with the ahci_mvebu driver (Armada
> 385 SoC) with the second port declared as "dummy":
> 
> [...]

Applied to libata/linux.git (for-6.14), thanks!

[1/1] ata: libahci_platform: Do not set mask_port_map when not needed
      https://git.kernel.org/libata/linux/c/2c202e6c

Kind regards,
Niklas

Re: [PATCH v2] ata: libahci_platform: Do not set mask_port_map when not needed

[Marek Szyprowski]

Hi Damien,

On 08.02.2025 00:29, Damien Le Moal wrote:
> Commit 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive
> port numbers") modified ahci_platform_get_resources() to allow
> identifying the ports of a controller that are defined as child nodes of
> the controller node in order to support non-consecutive port numbers (as
> defined by the platform device tree).
>
> However, this commit also erroneously sets bit 0 of
> hpriv->mask_port_map when the platform devices tree does not define port
> child nodes, to match the fact that the temporary default number of
> ports used in that case is 1 (which is also consistent with the fact
> that only index 0 of hpriv->phys[] is initialized with the call to
> ahci_platform_get_phy(). But doing so causes ahci_platform_init_host()
> to initialize and probe only the first port, even if this function
> determines that the controller has in fact multiple ports using the
> capability register of the controller (through a call to
> ahci_nr_ports()). This can be seen with the ahci_mvebu driver (Armada
> 385 SoC) with the second port declared as "dummy":
>
> ahci-mvebu f10a8000.sata: masking port_map 0x3 -> 0x1
> ahci-mvebu f10a8000.sata: AHCI vers 0001.0000, 32 command slots, 6 Gbps, platform mode
> ahci-mvebu f10a8000.sata: 1/2 ports implemented (port mask 0x1)
> ahci-mvebu f10a8000.sata: flags: 64bit ncq sntf led only pmp fbs pio slum part sxs
> scsi host0: ahci-mvebu
> scsi host1: ahci-mvebu
> ata1: SATA max UDMA/133 mmio [mem 0xf10a8000-0xf10a9fff] port 0x100 irq 40 lpm-pol 0
> ata2: DUMMY
>
> Fix this issue by removing setting bit 0 of hpriv->mask_port_map when
> the platform device tree does not define port child nodes.
>
> Reported-by: Klaus Kudielka <klaus.kudielka@gmail.com>
> Fixes: 8c87215dd3a2 ("ata: libahci_platform: support non-consecutive port numbers")
> Tested-by: Klaus Kudielka <klaus.kudielka@gmail.com>
> Signed-off-by: Damien Le Moal <dlemoal@kernel.org>

This patch landed yesterday in linux-next as commit 2c202e6c4f4d ("ata: 
libahci_platform: Do not set mask_port_map when not needed"). 
Unfortunately it breaks SATA operation on legacy ARM Exynos5250 Arndale 
board (arch/arm/boot/dts/samsung/exynos5250-arndale.dts). I wonder if 
this patch is really correct as if I understand code correctly, it 
disables support for any legacy of-based, single port.

Here is the log from that board:

[    0.937868] libata version 3.00 loaded.
[    2.322611] ahci-dwc 122f0000.sata: supply ahci not found, using 
dummy regulator
[    2.329389] ahci-dwc 122f0000.sata: supply phy not found, using dummy 
regulator
[    2.336745] ahci-dwc 122f0000.sata: supply target not found, using 
dummy regulator
[    2.344254] ahci-dwc 122f0000.sata: forcing port_map 0x0 -> 0x1
[    2.349517] ahci-dwc 122f0000.sata: AHCI vers 0001.0300, 32 command 
slots, 6 Gbps, platform mode
[    2.358128] ahci-dwc 122f0000.sata: 1/1 ports implemented (port mask 
0x1)
[    2.364921] ahci-dwc 122f0000.sata: flags: ncq sntf pm led clo only 
pmp pio slum part ccc apst
[    2.385620] ata1: SATA max UDMA/133 mmio [mem 0x122f0000-0x122f01fe] 
port 0x100 irq 101 lpm-pol 0
[    2.716379] ata1: SATA link down (SStatus 0 SControl 300)

with the $subject patch reverted:

[    0.938268] libata version 3.00 loaded.
[    2.283918] ahci-dwc 122f0000.sata: supply ahci not found, using 
dummy regulator
[    2.290714] ahci-dwc 122f0000.sata: supply phy not found, using dummy 
regulator
[    2.298064] ahci-dwc 122f0000.sata: supply target not found, using 
dummy regulator
[    2.312275] ahci-dwc 122f0000.sata: forcing port_map 0x0 -> 0x1
[    2.316847] ahci-dwc 122f0000.sata: masking port_map 0x1 -> 0x1
[    2.322791] ahci-dwc 122f0000.sata: AHCI vers 0001.0300, 32 command 
slots, 6 Gbps, platform mode
[    2.331464] ahci-dwc 122f0000.sata: 1/1 ports implemented (port mask 
0x1)
[    2.338258] ahci-dwc 122f0000.sata: flags: ncq sntf pm led clo only 
pmp pio slum part ccc apst
[    2.359896] ata1: SATA max UDMA/133 mmio [mem 0x122f0000-0x122f01fe] 
port 0x100 irq 101 lpm-pol 0
[    2.687374] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[    2.744673] ata1.00: HPA detected: current 117229295, native 117231408
[    2.757119] ata1.00: ATA-8: Corsair CSSD-F60GB2, 1.1, max UDMA/133
[    2.772062] ata1.00: 117229295 sectors, multi 1: LBA48 NCQ (depth 32)
[    2.834520] ata1.00: configured for UDMA/133


> ---
> Changes from v1:
>   - Improved commit message (fixed typos, added example and removed
>     cc-stable tag).
>
>   drivers/ata/libahci_platform.c | 2 --
>   1 file changed, 2 deletions(-)
>
> diff --git a/drivers/ata/libahci_platform.c b/drivers/ata/libahci_platform.c
> index 53b2c7719dc5..91d44302eac9 100644
> --- a/drivers/ata/libahci_platform.c
> +++ b/drivers/ata/libahci_platform.c
> @@ -651,8 +651,6 @@ struct ahci_host_priv *ahci_platform_get_resources(struct platform_device *pdev,
>   		 * If no sub-node was found, keep this for device tree
>   		 * compatibility
>   		 */
> -		hpriv->mask_port_map |= BIT(0);
> -
>   		rc = ahci_platform_get_phy(hpriv, 0, dev, dev->of_node);
>   		if (rc)
>   			goto err_out;

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland

Re: [PATCH v2] ata: libahci_platform: Do not set mask_port_map when not needed

[Niklas Cassel]

Hello Marek,

On Tue, Feb 25, 2025 at 10:34:27AM +0100, Marek Szyprowski wrote:
> with the $subject patch reverted:
> 
> [    0.938268] libata version 3.00 loaded.
> [    2.283918] ahci-dwc 122f0000.sata: supply ahci not found, using 
> dummy regulator
> [    2.290714] ahci-dwc 122f0000.sata: supply phy not found, using dummy 
> regulator
> [    2.298064] ahci-dwc 122f0000.sata: supply target not found, using 
> dummy regulator
> [    2.312275] ahci-dwc 122f0000.sata: forcing port_map 0x0 -> 0x1
> [    2.316847] ahci-dwc 122f0000.sata: masking port_map 0x1 -> 0x1
> [    2.322791] ahci-dwc 122f0000.sata: AHCI vers 0001.0300, 32 command 
> slots, 6 Gbps, platform mode
> [    2.331464] ahci-dwc 122f0000.sata: 1/1 ports implemented (port mask 
> 0x1)
> [    2.338258] ahci-dwc 122f0000.sata: flags: ncq sntf pm led clo only 
> pmp pio slum part ccc apst
> [    2.359896] ata1: SATA max UDMA/133 mmio [mem 0x122f0000-0x122f01fe] 
> port 0x100 irq 101 lpm-pol 0
> [    2.687374] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
> [    2.744673] ata1.00: HPA detected: current 117229295, native 117231408
> [    2.757119] ata1.00: ATA-8: Corsair CSSD-F60GB2, 1.1, max UDMA/133
> [    2.772062] ata1.00: 117229295 sectors, multi 1: LBA48 NCQ (depth 32)
> [    2.834520] ata1.00: configured for UDMA/133

Could you please test this patch:
https://lore.kernel.org/linux-ide/20250225141612.942170-2-cassel@kernel.org/T/#u
and see if it helps.


Kind regards,
Niklas