Message ID | a421afd8-d841-4d3c-bfe9-9dee707bb319@debian.org |
---|---|
State | New |
Headers | show |
Series | Disabling link power-management on Samsung SSD 870 QVO drives to make them work. | expand |
Hello Daniel, On Sat, Jan 18, 2025 at 06:52:30AM +0100, Daniel Baumann wrote: > Hi, > > since the kernel enabled LPM by default, Samsung 870 QVO SATA SSDs don't > work anymore. Disabling LPM for them solves it, please consider applying > the attached patch. > > I've added a new case for 870 QVO specifically as regular 870 don't have > the issue. This should have been in the commit message. Out of curiosity, did you test on regular 870, so you know they are not broken as well? > > [I'm not subscribed to the list, please CC on replies] > > Regards, > Daniel > From 55e2cb12faa48cf429ad4a299ea4e4a3a9ba25f5 Mon Sep 17 00:00:00 2001 > From: Daniel Baumann <daniel@debian.org> > Date: Sat, 18 Jan 2025 06:36:43 +0100 > Subject: [PATCH] Disabling link power-management on Samsung SSD 870 QVO drives > to make them work. The subject should be something like: ata: libata-core: Add ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives > > Signed-off-by: Daniel Baumann <daniel@debian.org> You probably also want to add: Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type") Cc: stable@vger.kernel.org > --- > drivers/ata/libata-core.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c > index c085dd81ebe7..63ec2f218431 100644 > --- a/drivers/ata/libata-core.c > +++ b/drivers/ata/libata-core.c > @@ -4143,6 +4143,10 @@ static const struct ata_dev_quirks_entry __ata_dev_quirks[] = { > { "Samsung SSD 860*", NULL, ATA_QUIRK_NO_NCQ_TRIM | > ATA_QUIRK_ZERO_AFTER_TRIM | > ATA_QUIRK_NO_NCQ_ON_ATI }, > + { "Samsung SSD 870 QVO*", NULL, ATA_QUIRK_NO_NCQ_TRIM | > + ATA_QUIRK_ZERO_AFTER_TRIM | > + ATA_QUIRK_NO_NCQ_ON_ATI | > + ATA_QUIRK_NOLPM }, > { "Samsung SSD 870*", NULL, ATA_QUIRK_NO_NCQ_TRIM | > ATA_QUIRK_ZERO_AFTER_TRIM | > ATA_QUIRK_NO_NCQ_ON_ATI }, > -- > 2.45.2 >
Hi Niklas, On 1/20/25 14:18, Niklas Cassel wrote: >> I've added a new case for 870 QVO specifically as regular 870 don't have >> the issue. > > This should have been in the commit message. thanks for the pointer, I have now adjusted the commit message. > Out of curiosity, did you test on regular 870, so you know they are not > broken as well? we don't have many of non-QVO variants, but so far none of them has needed it. I've added that to the commit message as well. > You probably also want to add: > Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type") > Cc: stable@vger.kernel.org added, thanks! Updated patch attached. Regards, Daniel
On 1/21/25 12:08 AM, Daniel Baumann wrote: > Hi Niklas, > > On 1/20/25 14:18, Niklas Cassel wrote: >>> I've added a new case for 870 QVO specifically as regular 870 don't have >>> the issue. >> >> This should have been in the commit message. > > thanks for the pointer, I have now adjusted the commit message. > >> Out of curiosity, did you test on regular 870, so you know they are not >> broken as well? > > we don't have many of non-QVO variants, but so far none of them has > needed it. I've added that to the commit message as well. > >> You probably also want to add: >> Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type") >> Cc: stable@vger.kernel.org > > added, thanks! > > Updated patch attached. For reference for future patches: please use "git send-email" to send patches. Also please make sure to version the patches (git format-patch -vX ...) and add changelog notes after the "---" at the end of the commit message and before the first line of the patch content (in your case, the line "drivers/ata/libata-core.c | 4 ++++") Thank you.
On 1/23/25 05:21, Damien Le Moal wrote:
> For reference for future patches:
sorry, I wasn't aware.
I'll make sure of it for next time, thank you for the help.
Regards,
Daniel
On Mon, Jan 20, 2025 at 04:08:16PM +0100, Daniel Baumann wrote: > Hi Niklas, > > On 1/20/25 14:18, Niklas Cassel wrote: > >> I've added a new case for 870 QVO specifically as regular 870 don't have > >> the issue. > > > > This should have been in the commit message. > > thanks for the pointer, I have now adjusted the commit message. > > > Out of curiosity, did you test on regular 870, so you know they are not > > broken as well? > > we don't have many of non-QVO variants, but so far none of them has > needed it. I've added that to the commit message as well. > > > You probably also want to add: > > Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type") > > Cc: stable@vger.kernel.org > > added, thanks! > > Updated patch attached. > > Regards, > Daniel > From f2b44baf60d11882696a99445bf217d6ee409903 Mon Sep 17 00:00:00 2001 > From: Daniel Baumann <daniel@debian.org> > Date: Sat, 18 Jan 2025 06:36:43 +0100 > Subject: [PATCH] ata: libata-core: Add ATA_QUIRK_NOLPM for Samsung SSD 870 QVO > drives > > Disabling link power management on Samsung SSD 870 QVO drives > to make them work again after the switch of the default LPM > policy to low. > > Testing so far has shown that regular Samsung SSD 870 > (the non QVO variants) do not need it and work fine with > the default LPM policy. > > Cc: stable@vger.kernel.org > Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type") > Signed-off-by: Daniel Baumann <daniel@debian.org> > --- > drivers/ata/libata-core.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c > index c085dd81ebe7..63ec2f218431 100644 > --- a/drivers/ata/libata-core.c > +++ b/drivers/ata/libata-core.c > @@ -4143,6 +4143,10 @@ static const struct ata_dev_quirks_entry __ata_dev_quirks[] = { > { "Samsung SSD 860*", NULL, ATA_QUIRK_NO_NCQ_TRIM | > ATA_QUIRK_ZERO_AFTER_TRIM | > ATA_QUIRK_NO_NCQ_ON_ATI }, > + { "Samsung SSD 870 QVO*", NULL, ATA_QUIRK_NO_NCQ_TRIM | > + ATA_QUIRK_ZERO_AFTER_TRIM | > + ATA_QUIRK_NO_NCQ_ON_ATI | > + ATA_QUIRK_NOLPM }, > { "Samsung SSD 870*", NULL, ATA_QUIRK_NO_NCQ_TRIM | > ATA_QUIRK_ZERO_AFTER_TRIM | > ATA_QUIRK_NO_NCQ_ON_ATI }, > -- > 2.45.2 > Applied to libata/linux.git (for-6.14), thanks! [1/1] ata: libata-core: Add ATA_QUIRK_NOLPM for Samsung SSD 870 QVO drives https://git.kernel.org/libata/linux/c/cc77e2ce Kind regards, Niklas
From 55e2cb12faa48cf429ad4a299ea4e4a3a9ba25f5 Mon Sep 17 00:00:00 2001 From: Daniel Baumann <daniel@debian.org> Date: Sat, 18 Jan 2025 06:36:43 +0100 Subject: [PATCH] Disabling link power-management on Samsung SSD 870 QVO drives to make them work. Signed-off-by: Daniel Baumann <daniel@debian.org> --- drivers/ata/libata-core.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c index c085dd81ebe7..63ec2f218431 100644 --- a/drivers/ata/libata-core.c +++ b/drivers/ata/libata-core.c @@ -4143,6 +4143,10 @@ static const struct ata_dev_quirks_entry __ata_dev_quirks[] = { { "Samsung SSD 860*", NULL, ATA_QUIRK_NO_NCQ_TRIM | ATA_QUIRK_ZERO_AFTER_TRIM | ATA_QUIRK_NO_NCQ_ON_ATI }, + { "Samsung SSD 870 QVO*", NULL, ATA_QUIRK_NO_NCQ_TRIM | + ATA_QUIRK_ZERO_AFTER_TRIM | + ATA_QUIRK_NO_NCQ_ON_ATI | + ATA_QUIRK_NOLPM }, { "Samsung SSD 870*", NULL, ATA_QUIRK_NO_NCQ_TRIM | ATA_QUIRK_ZERO_AFTER_TRIM | ATA_QUIRK_NO_NCQ_ON_ATI }, -- 2.45.2