diff --git a/ci.sh b/ci.sh
index 323e752d..b325e376 100755
--- a/ci.sh
+++ b/ci.sh
@@ -63,7 +63,7 @@ cargo batch  \
     --- build --release --manifest-path embassy-rp/Cargo.toml --target thumbv6m-none-eabi --features nightly,intrinsics \
     --- build --release --manifest-path embassy-stm32/Cargo.toml --target thumbv7em-none-eabi --features nightly,stm32f410tb,defmt,exti,time-driver-any,unstable-traits \
     --- build --release --manifest-path embassy-stm32/Cargo.toml --target thumbv7em-none-eabi --features nightly,stm32f411ce,defmt,exti,time-driver-any,unstable-traits \
-    --- build --release --manifest-path embassy-stm32/Cargo.toml --target thumbv7em-none-eabi --features nightly,stm32f429zi,log,exti,time-driver-any,unstable-traits \
+    --- build --release --manifest-path embassy-stm32/Cargo.toml --target thumbv7em-none-eabi --features nightly,stm32f429zi,log,exti,time-driver-any,unstable-traits,embedded-sdmmc \
     --- build --release --manifest-path embassy-stm32/Cargo.toml --target thumbv7em-none-eabi --features nightly,stm32h755zi-cm7,defmt,exti,time-driver-any,unstable-traits \
     --- build --release --manifest-path embassy-stm32/Cargo.toml --target thumbv7em-none-eabi --features nightly,stm32h7b3ai,defmt,exti,time-driver-any,unstable-traits \
     --- build --release --manifest-path embassy-stm32/Cargo.toml --target thumbv7em-none-eabi --features nightly,stm32l476vg,defmt,exti,time-driver-any,unstable-traits \
diff --git a/embassy-stm32/Cargo.toml b/embassy-stm32/Cargo.toml
index d9c1f6dc..10e5e48c 100644
--- a/embassy-stm32/Cargo.toml
+++ b/embassy-stm32/Cargo.toml
@@ -58,7 +58,7 @@ cortex-m = "0.7.6"
 futures = { version = "0.3.17", default-features = false, features = ["async-await"] }
 rand_core = "0.6.3"
 sdio-host = "0.5.0"
-embedded-sdmmc = { git = "https://github.com/thalesfragoso/embedded-sdmmc-rs", branch = "async", optional = true }
+embedded-sdmmc = { git = "https://github.com/embassy-rs/embedded-sdmmc-rs", rev = "46d1b1c2ff13e31e282ec1e352421721694f126a", optional = true }
 critical-section = "1.1"
 atomic-polyfill = "1.0.1"
 stm32-metapac = { version = "0.1.0", path = "../stm32-metapac", features = ["rt"] }
@@ -77,7 +77,6 @@ stm32-metapac = { version = "0.1.0", path = "../stm32-metapac", default-features
 
 [features]
 defmt = ["dep:defmt", "bxcan/unstable-defmt", "embassy-sync/defmt", "embassy-executor/defmt", "embassy-embedded-hal/defmt", "embassy-hal-common/defmt", "embedded-io?/defmt", "embassy-usb-driver?/defmt", "embassy-net-driver/defmt"]
-sdmmc-rs = ["embedded-sdmmc"]
 memory-x = ["stm32-metapac/memory-x"]
 subghz = []
 exti = []
diff --git a/embassy-stm32/src/sdmmc/mod.rs b/embassy-stm32/src/sdmmc/mod.rs
index a52c65b9..3f07e0c0 100644
--- a/embassy-stm32/src/sdmmc/mod.rs
+++ b/embassy-stm32/src/sdmmc/mod.rs
@@ -1532,7 +1532,7 @@ foreach_peripheral!(
     };
 );
 
-#[cfg(feature = "sdmmc-rs")]
+#[cfg(feature = "embedded-sdmmc")]
 mod sdmmc_rs {
     use core::future::Future;
 
@@ -1540,7 +1540,7 @@ mod sdmmc_rs {
 
     use super::*;
 
-    impl<'d, T: Instance, P: Pins<T>> BlockDevice for Sdmmc<'d, T, P> {
+    impl<'d, T: Instance, Dma: SdmmcDma<T>> BlockDevice for Sdmmc<'d, T, Dma> {
         type Error = Error;
 
         type ReadFuture<'a> = impl Future<Output = Result<(), Self::Error>> + 'a
@@ -1558,19 +1558,14 @@ mod sdmmc_rs {
             _reason: &str,
         ) -> Self::ReadFuture<'a> {
             async move {
-                let card_capacity = self.card()?.card_type;
-                let inner = T::inner();
-                let state = T::state();
                 let mut address = start_block_idx.0;
 
                 for block in blocks.iter_mut() {
                     let block: &mut [u8; 512] = &mut block.contents;
 
                     // NOTE(unsafe) Block uses align(4)
-                    let buf = unsafe { &mut *(block as *mut [u8; 512] as *mut [u32; 128]) };
-                    inner
-                        .read_block(address, buf, card_capacity, state, self.config.data_transfer_timeout)
-                        .await?;
+                    let block = unsafe { &mut *(block as *mut _ as *mut DataBlock) };
+                    self.read_block(address, block).await?;
                     address += 1;
                 }
                 Ok(())
@@ -1579,19 +1574,14 @@ mod sdmmc_rs {
 
         fn write<'a>(&'a mut self, blocks: &'a [Block], start_block_idx: BlockIdx) -> Self::WriteFuture<'a> {
             async move {
-                let card = self.card.as_mut().ok_or(Error::NoCard)?;
-                let inner = T::inner();
-                let state = T::state();
                 let mut address = start_block_idx.0;
 
                 for block in blocks.iter() {
                     let block: &[u8; 512] = &block.contents;
 
                     // NOTE(unsafe) DataBlock uses align 4
-                    let buf = unsafe { &*(block as *const [u8; 512] as *const [u32; 128]) };
-                    inner
-                        .write_block(address, buf, card, state, self.config.data_transfer_timeout)
-                        .await?;
+                    let block = unsafe { &*(block as *const _ as *const DataBlock) };
+                    self.write_block(address, block).await?;
                     address += 1;
                 }
                 Ok(())