[PATCH 0/2] drm: tyr: use lifetime-bound IoMem

[PATCH 0/2] drm: tyr: use lifetime-bound IoMem

[Danilo Krummrich]

Adopt the driver core lifetime infrastructure for tyr.

Separate the driver type from the driver data and use lifetime-bound
IoMem directly in probe instead of wrapping it in Devres and Arc,
simplifying register access.

Changes in v2:
  - Don't return Result from GpuInfo::new()
  - Use "drm/tyr:" commit message prefix

Danilo Krummrich (2):
  drm/tyr: separate driver type from driver data
  drm/tyr: use IoMem directly instead of Devres

 drivers/gpu/drm/tyr/driver.rs | 29 ++++++++++++-----------------
 drivers/gpu/drm/tyr/gpu.rs    | 17 +++++------------
 drivers/gpu/drm/tyr/tyr.rs    |  4 ++--
 3 files changed, 19 insertions(+), 31 deletions(-)


base-commit: a3e50e7279996cd987001fd8a3db36e72665f8f7
-- 
2.54.0

[PATCH v2 1/2] drm/tyr: separate driver type from driver data

[Danilo Krummrich]

Introduce TyrPlatformDriver as a unit struct for the platform::Driver
trait implementation and keep TyrPlatformDriverData for the private
driver data.

Reviewed-by: Gary Guo <gary@garyguo.net>
Tested-by: Deborah Brouwer <deborah.brouwer@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Eliot Courtney <ecourtney@nvidia.com>
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
---
 drivers/gpu/drm/tyr/driver.rs | 10 ++++++----
 drivers/gpu/drm/tyr/tyr.rs    |  4 ++--
 2 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/tyr/driver.rs b/drivers/gpu/drm/tyr/driver.rs
index 98732afc096f..6276e9743c32 100644
--- a/drivers/gpu/drm/tyr/driver.rs
+++ b/drivers/gpu/drm/tyr/driver.rs
@@ -51,6 +51,8 @@
 /// Convenience type alias for the DRM device type for this driver.
 pub(crate) type TyrDrmDevice = drm::Device<TyrDrmDriver>;
 
+pub(crate) struct TyrPlatformDriver;
+
 #[pin_data(PinnedDrop)]
 pub(crate) struct TyrPlatformDriverData {
     _device: ARef<TyrDrmDevice>,
@@ -93,22 +95,22 @@ fn issue_soft_reset(dev: &Device<Bound>, iomem: &Devres<IoMem>) -> Result {
 kernel::of_device_table!(
     OF_TABLE,
     MODULE_OF_TABLE,
-    <TyrPlatformDriverData as platform::Driver>::IdInfo,
+    <TyrPlatformDriver as platform::Driver>::IdInfo,
     [
         (of::DeviceId::new(c"rockchip,rk3588-mali"), ()),
         (of::DeviceId::new(c"arm,mali-valhall-csf"), ())
     ]
 );
 
-impl platform::Driver for TyrPlatformDriverData {
+impl platform::Driver for TyrPlatformDriver {
     type IdInfo = ();
-    type Data<'bound> = Self;
+    type Data<'bound> = TyrPlatformDriverData;
     const OF_ID_TABLE: Option<of::IdTable<Self::IdInfo>> = Some(&OF_TABLE);
 
     fn probe<'bound>(
         pdev: &'bound platform::Device<Core<'_>>,
         _info: Option<&'bound Self::IdInfo>,
-    ) -> impl PinInit<Self, Error> + 'bound {
+    ) -> impl PinInit<Self::Data<'bound>, Error> + 'bound {
         let core_clk = Clk::get(pdev.as_ref(), Some(c"core"))?;
         let stacks_clk = OptionalClk::get(pdev.as_ref(), Some(c"stacks"))?;
         let coregroup_clk = OptionalClk::get(pdev.as_ref(), Some(c"coregroup"))?;
diff --git a/drivers/gpu/drm/tyr/tyr.rs b/drivers/gpu/drm/tyr/tyr.rs
index 9432ddd6b5b8..95cda7b0962f 100644
--- a/drivers/gpu/drm/tyr/tyr.rs
+++ b/drivers/gpu/drm/tyr/tyr.rs
@@ -5,7 +5,7 @@
 //! The name "Tyr" is inspired by Norse mythology, reflecting Arm's tradition of
 //! naming their GPUs after Nordic mythological figures and places.
 
-use crate::driver::TyrPlatformDriverData;
+use crate::driver::TyrPlatformDriver;
 
 mod driver;
 mod file;
@@ -14,7 +14,7 @@
 mod regs;
 
 kernel::module_platform_driver! {
-    type: TyrPlatformDriverData,
+    type: TyrPlatformDriver,
     name: "tyr",
     authors: ["The Tyr driver authors"],
     description: "Arm Mali Tyr DRM driver",
-- 
2.54.0

Re: [PATCH v2 1/2] drm/tyr: separate driver type from driver data

[Alexandre Courbot]

On Fri May 29, 2026 at 9:00 AM JST, Danilo Krummrich wrote:
> Introduce TyrPlatformDriver as a unit struct for the platform::Driver
> trait implementation and keep TyrPlatformDriverData for the private
> driver data.
>
> Reviewed-by: Gary Guo <gary@garyguo.net>
> Tested-by: Deborah Brouwer <deborah.brouwer@collabora.com>
> Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
> Reviewed-by: Eliot Courtney <ecourtney@nvidia.com>
> Signed-off-by: Danilo Krummrich <dakr@kernel.org>

Reviewed-by: Alexandre Courbot <acourbot@nvidia.com>

[PATCH v2 2/2] drm/tyr: use IoMem directly instead of Devres

[Danilo Krummrich]

Now that IoMem is lifetime-parameterized, use it directly in probe
rather than wrapping it in Devres and Arc. The I/O memory mapping is
only used during probe and not stored in driver data, so device-managed
revocation is unnecessary.

This removes the Devres access(dev) pattern from issue_soft_reset(),
GpuInfo::new(), and l2_power_on(), simplifying register access.

Reviewed-by: Eliot Courtney <ecourtney@nvidia.com>
Reviewed-by: Alexandre Courbot <acourbot@nvidia.com>
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
---
 drivers/gpu/drm/tyr/driver.rs | 19 ++++++-------------
 drivers/gpu/drm/tyr/gpu.rs    | 17 +++++------------
 2 files changed, 11 insertions(+), 25 deletions(-)

diff --git a/drivers/gpu/drm/tyr/driver.rs b/drivers/gpu/drm/tyr/driver.rs
index 6276e9743c32..227ea2adccea 100644
--- a/drivers/gpu/drm/tyr/driver.rs
+++ b/drivers/gpu/drm/tyr/driver.rs
@@ -6,11 +6,9 @@
         OptionalClk, //
     },
     device::{
-        Bound,
         Core,
         Device, //
     },
-    devres::Devres,
     dma::{
         Device as DmaDevice,
         DmaMask, //
@@ -30,7 +28,6 @@
     sizes::SZ_2M,
     sync::{
         aref::ARef,
-        Arc,
         Mutex, //
     },
     time, //
@@ -44,7 +41,7 @@
     regs::gpu_control::*, //
 };
 
-pub(crate) type IoMem = kernel::io::mem::IoMem<'static, SZ_2M>;
+pub(crate) type IoMem<'a> = kernel::io::mem::IoMem<'a, SZ_2M>;
 
 pub(crate) struct TyrDrmDriver;
 
@@ -74,15 +71,11 @@ pub(crate) struct TyrDrmDeviceData {
     pub(crate) gpu_info: GpuInfo,
 }
 
-fn issue_soft_reset(dev: &Device<Bound>, iomem: &Devres<IoMem>) -> Result {
-    let io = (*iomem).access(dev)?;
-    io.write_reg(GPU_COMMAND::reset(ResetMode::SoftReset));
+fn issue_soft_reset(dev: &Device, iomem: &IoMem<'_>) -> Result {
+    iomem.write_reg(GPU_COMMAND::reset(ResetMode::SoftReset));
 
     poll::read_poll_timeout(
-        || {
-            let io = (*iomem).access(dev)?;
-            Ok(io.read(GPU_IRQ_RAWSTAT))
-        },
+        || Ok(iomem.read(GPU_IRQ_RAWSTAT)),
         |status| status.reset_completed(),
         time::Delta::from_millis(1),
         time::Delta::from_millis(100),
@@ -123,12 +116,12 @@ fn probe<'bound>(
         let sram_regulator = Regulator::<regulator::Enabled>::get(pdev.as_ref(), c"sram")?;
 
         let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
-        let iomem = Arc::new(request.iomap_sized::<SZ_2M>()?.into_devres()?, GFP_KERNEL)?;
+        let iomem = request.iomap_sized::<SZ_2M>()?;
 
         issue_soft_reset(pdev.as_ref(), &iomem)?;
         gpu::l2_power_on(pdev.as_ref(), &iomem)?;
 
-        let gpu_info = GpuInfo::new(pdev.as_ref(), &iomem)?;
+        let gpu_info = GpuInfo::new(&iomem);
         gpu_info.log(pdev.as_ref());
 
         let pa_bits = MMU_FEATURES::from_raw(gpu_info.mmu_features)
diff --git a/drivers/gpu/drm/tyr/gpu.rs b/drivers/gpu/drm/tyr/gpu.rs
index 652556026f50..592b8bb16eba 100644
--- a/drivers/gpu/drm/tyr/gpu.rs
+++ b/drivers/gpu/drm/tyr/gpu.rs
@@ -9,7 +9,6 @@
         Bound,
         Device, //
     },
-    devres::Devres,
     io::{
         poll,
         register::Array,
@@ -40,10 +39,8 @@
 pub(crate) struct GpuInfo(pub(crate) uapi::drm_panthor_gpu_info);
 
 impl GpuInfo {
-    pub(crate) fn new(dev: &Device<Bound>, iomem: &Devres<IoMem>) -> Result<Self> {
-        let io = (*iomem).access(dev)?;
-
-        Ok(Self(uapi::drm_panthor_gpu_info {
+    pub(crate) fn new(io: &IoMem<'_>) -> Self {
+        Self(uapi::drm_panthor_gpu_info {
             gpu_id: io.read(GPU_ID).into_raw(),
             gpu_rev: io.read(REVIDR).into_raw(),
             csf_id: io.read(CSF_ID).into_raw(),
@@ -81,7 +78,7 @@ pub(crate) fn new(dev: &Device<Bound>, iomem: &Devres<IoMem>) -> Result<Self> {
             pad: 0,
             //GPU_FEATURES register is not available; it was introduced in arch 11.x.
             gpu_features: 0,
-        }))
+        })
     }
 
     pub(crate) fn log(&self, dev: &Device<Bound>) {
@@ -163,15 +160,11 @@ struct GpuModels {
 }];
 
 /// Powers on the l2 block.
-pub(crate) fn l2_power_on(dev: &Device<Bound>, iomem: &Devres<IoMem>) -> Result {
-    let io = (*iomem).access(dev)?;
+pub(crate) fn l2_power_on(dev: &Device, io: &IoMem<'_>) -> Result {
     io.write_reg(L2_PWRON_LO::zeroed().with_const_request::<1>());
 
     poll::read_poll_timeout(
-        || {
-            let io = (*iomem).access(dev)?;
-            Ok(io.read(L2_READY_LO))
-        },
+        || Ok(io.read(L2_READY_LO)),
         |status| status.ready() == 1,
         Delta::from_millis(1),
         Delta::from_millis(100),
-- 
2.54.0

Re: [PATCH v2 2/2] drm/tyr: use IoMem directly instead of Devres

[Alice Ryhl]

On Fri, May 29, 2026 at 02:00:54AM +0200, Danilo Krummrich wrote:
> Now that IoMem is lifetime-parameterized, use it directly in probe
> rather than wrapping it in Devres and Arc. The I/O memory mapping is
> only used during probe and not stored in driver data, so device-managed
> revocation is unnecessary.
> 
> This removes the Devres access(dev) pattern from issue_soft_reset(),
> GpuInfo::new(), and l2_power_on(), simplifying register access.
> 
> Reviewed-by: Eliot Courtney <ecourtney@nvidia.com>
> Reviewed-by: Alexandre Courbot <acourbot@nvidia.com>
> Signed-off-by: Danilo Krummrich <dakr@kernel.org>

> -pub(crate) type IoMem = kernel::io::mem::IoMem<'static, SZ_2M>;
> +pub(crate) type IoMem<'a> = kernel::io::mem::IoMem<'a, SZ_2M>;

It'd make more sense for me to put 'b or 'bound here.

>          let sram_regulator = Regulator::<regulator::Enabled>::get(pdev.as_ref(), c"sram")?;
>  
>          let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
> -        let iomem = Arc::new(request.iomap_sized::<SZ_2M>()?.into_devres()?, GFP_KERNEL)?;
> +        let iomem = request.iomap_sized::<SZ_2M>()?;
>  
>          issue_soft_reset(pdev.as_ref(), &iomem)?;
>          gpu::l2_power_on(pdev.as_ref(), &iomem)?;
>  
> -        let gpu_info = GpuInfo::new(pdev.as_ref(), &iomem)?;
> +        let gpu_info = GpuInfo::new(&iomem);


While this change is fine, I notice that we don't actually keep the
iomem alive past the probe method. I assume we're going to need that,
which leads to the question of whether we can store the iomem in the
places we need it.

As far as I can tell, we can store it in TyrPlatformDriverData but not
in TyrDrmDeviceData, is that right?

I guess it does make sense because the io memory goes away if the
underlying platform device (the bus) goes away, even if the drm device
still exists due to open fds from userspace.

Alice

Re: [PATCH v2 2/2] drm/tyr: use IoMem directly instead of Devres

[Danilo Krummrich]

On Mon Jun 1, 2026 at 11:35 AM CEST, Alice Ryhl wrote:
> On Fri, May 29, 2026 at 02:00:54AM +0200, Danilo Krummrich wrote:
>> Now that IoMem is lifetime-parameterized, use it directly in probe
>> rather than wrapping it in Devres and Arc. The I/O memory mapping is
>> only used during probe and not stored in driver data, so device-managed
>> revocation is unnecessary.
>> 
>> This removes the Devres access(dev) pattern from issue_soft_reset(),
>> GpuInfo::new(), and l2_power_on(), simplifying register access.
>> 
>> Reviewed-by: Eliot Courtney <ecourtney@nvidia.com>
>> Reviewed-by: Alexandre Courbot <acourbot@nvidia.com>
>> Signed-off-by: Danilo Krummrich <dakr@kernel.org>
>
>> -pub(crate) type IoMem = kernel::io::mem::IoMem<'static, SZ_2M>;
>> +pub(crate) type IoMem<'a> = kernel::io::mem::IoMem<'a, SZ_2M>;
>
> It'd make more sense for me to put 'b or 'bound here.

You can name it 'b of course, but the semantics for 'bound has been agreed to
describe the lifetime of something actually being in place for the entire
duration of a device being bound to a driver, which isn't the case here.

>>          let sram_regulator = Regulator::<regulator::Enabled>::get(pdev.as_ref(), c"sram")?;
>>  
>>          let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
>> -        let iomem = Arc::new(request.iomap_sized::<SZ_2M>()?.into_devres()?, GFP_KERNEL)?;
>> +        let iomem = request.iomap_sized::<SZ_2M>()?;
>>  
>>          issue_soft_reset(pdev.as_ref(), &iomem)?;
>>          gpu::l2_power_on(pdev.as_ref(), &iomem)?;
>>  
>> -        let gpu_info = GpuInfo::new(pdev.as_ref(), &iomem)?;
>> +        let gpu_info = GpuInfo::new(&iomem);
>
>
> While this change is fine, I notice that we don't actually keep the
> iomem alive past the probe method. I assume we're going to need that,
> which leads to the question of whether we can store the iomem in the
> places we need it.
>
> As far as I can tell, we can store it in TyrPlatformDriverData but not
> in TyrDrmDeviceData, is that right?

TyrDrmDeviceData will eventually become the drm::Registration data, so it will
be a ForLt type as well and IoMem<'a> can directly be stored in there.

The DRM device private data being tied to the lifetime of the DRM device itself
doesn't seem to be useful anymore after IOCTLs won't potentially reach into the
driver after device unbind anymore.

In the meantime (or even after this change, although I wouldn't recommend that),
you can call into_devres() before storing it in TyrDrmDeviceData.