mbox series

[0/6] ufs-exynos fixes for gs101

Message ID 20250226220414.343659-1-peter.griffin@linaro.org
Headers show
Series ufs-exynos fixes for gs101 | expand

Message

Peter Griffin Feb. 26, 2025, 10:04 p.m. UTC 
Hi folks,

Whilst investigating some stability issues with the upstream ufs-exynos
driver on gs101/Pixel 6 the following fixes have been authored.

Whilst some of the stability issues remain these patches do improve
certain aspects and make things more deterministic especially module
load/unload.

regards,

Peter

Peter Griffin (6):
  scsi: ufs: exynos: ensure pre_link() executes before
    exynos_ufs_phy_init()
  scsi: ufs: exynos: move ufs shareability value to drvdata
  scsi: ufs: exynos: ensure consistent phy reference counts
  scsi: ufs: exynos: Enable PRDT pre-fetching with UFSHCD_CAP_CRYPTO
  scsi: ufs: exynos: Move phy calls to .exit() callback
  scsi: ufs: exynos: put ufs device in reset on .exit() and .suspend()

 drivers/ufs/host/ufs-exynos.c | 62 +++++++++++++++++++++++++----------
 drivers/ufs/host/ufs-exynos.h |  2 ++
 2 files changed, 46 insertions(+), 18 deletions(-)

Comments

Bart Van Assche Feb. 28, 2025, 7:20 p.m. UTC | #1 
On 2/26/25 2:04 PM, Peter Griffin wrote:
> +static void exynos_ufs_exit(struct ufs_hba *hba)
> +{
> +	struct exynos_ufs *ufs = ufshcd_get_variant(hba);
> +
> +	phy_power_off(ufs->phy);
> +	phy_exit(ufs->phy);
> +}
> +
> +

For future patches, please follow the convention that is used elsewhere
in the kernel and separate functions with a single blank line instead of
two.


Thanks,

Bart.
Peter Griffin March 4, 2025, 11:10 a.m. UTC | #2 
Hi Bart,

Thanks for your review feedback!

On Fri, 28 Feb 2025 at 19:20, Bart Van Assche <bvanassche@acm.org> wrote:
>
> On 2/26/25 2:04 PM, Peter Griffin wrote:
> > +static void exynos_ufs_exit(struct ufs_hba *hba)
> > +{
> > +     struct exynos_ufs *ufs = ufshcd_get_variant(hba);
> > +
> > +     phy_power_off(ufs->phy);
> > +     phy_exit(ufs->phy);
> > +}
> > +
> > +
>
> For future patches, please follow the convention that is used elsewhere
> in the kernel and separate functions with a single blank line instead of
> two.

That was an oversight on my part, will fix.

Thanks,

Peter