Arch Linux User Repository

On upgrading from kernel 6.13 (kernel-6.13.8-200.fc41.x86_64) Fedora-41 to kernel 6.14 (kernel-6.14.0-63.fc42.x86_64) Fedora-42, the dkms compilation broke. I had used the dkms code from here (aur-arch) to get my rts5227 working. I also needed to apply the previously mentioned patch. I am not a kernel expert. I don't really know what the sections of code do, nor the implication of the below code changes. But with some understanding of C language, and sifting through linux official git repo commit logs, I made the below changes to push forward through the compilation.

I hope a moderator/developer can notify me to make corrections if it turns out to be misleading/noise and possibly provide a more reliable patch.

Posting here to just to stir activity.

--- a/rtsx.c.orig
+++ b/rtsx.c
@@ -76,7 +76,7 @@ static const char* host_info(struct Scsi_Host *host)
        return "SCSI emulation for RTS5227";
 }

-static int slave_alloc (struct scsi_device *sdev)
+static int sdev_init (struct scsi_device *sdev)
 {
        /*
         * Set the INQUIRY transfer length to 36.  We don't use any of
@@ -87,7 +87,7 @@ static int slave_alloc (struct scsi_device *sdev)
        return 0;
 }

-static int slave_configure(struct scsi_device *sdev)
+static int sdev_configure(struct scsi_device *sdev, struct queue_limits *ql)
 {
        /* Set the SCSI level to at least 2.  We'll leave it at 3 if that's
         * what is originally reported.  We need this to avoid confusing
@@ -245,8 +245,8 @@ static struct scsi_host_template rtsx_host_template = {

        .this_id =                      -1,

-       .slave_alloc =                  slave_alloc,
-       .slave_configure =              slave_configure,
+       .sdev_init =                    sdev_init,
+       .sdev_configure =               sdev_configure,


        .sg_tablesize =                 SG_ALL,

Hello!

At commit https://github.com/torvalds/linux/commit/113d1dd9c8ea2186d56a641a787e2588673c9c32#diff-545daf8a36330dbe9e9bca384f13ff7dc878284c411be77102044393eab7e310 (Seek for blk_queue_dma_alignment in code) this function was removed from all drivers, including mainline rts5208/rtsx.c file.

That said, I applied this patch:

--- /var/lib/dkms/rts5227/1.08/source/rtsx.c.orig       2025-01-08 00:24:41.244169160 -0300
+++ /var/lib/dkms/rts5227/1.08/source/rtsx.c    2025-01-08 00:24:48.717199865 -0300
@@ -89,16 +89,6 @@

 static int slave_configure(struct scsi_device *sdev)
 {
-       /* Scatter-gather buffers (all but the last) must have a length
-        * divisible by the bulk maxpacket size.  Otherwise a data packet
-        * would end up being short, causing a premature end to the data
-        * transfer.  Since high-speed bulk pipes have a maxpacket size
-        * of 512, we'll use that as the scsi device queue's DMA alignment
-        * mask.  Guaranteeing proper alignment of the first buffer will
-        * have the desired effect because, except at the beginning and
-        * the end, scatter-gather buffers follow page boundaries. */
-       blk_queue_dma_alignment(sdev->request_queue, (512 - 1));
-
        /* Set the SCSI level to at least 2.  We'll leave it at 3 if that's
         * what is originally reported.  We need this to avoid confusing
         * the SCSI layer with devices that report 0 or 1, but need 10-byte

And module compiled. Since my hardware is the rtsx usb, this module still not serving me, but may help other people

Hi, I've got a compilation error with a 6.10.6 kernel :

/var/lib/dkms/rts5227/1.08/build/rtsx.c: In function 'slave_configure':
/var/lib/dkms/rts5227/1.08/build/rtsx.c:100:9: error: implicit declaration of function 'blk_queue_dma_alignment'; did you mean 'queue_dma_alignment'? [-Wimplicit-function-declaration]
  100 |         blk_queue_dma_alignment(sdev->request_queue, (512 - 1));
      |         ^~~~~~~~~~~~~~~~~~~~~~~
      |         queue_dma_alignment
make[2]: *** [scripts/Makefile.build:244: /var/lib/dkms/rts5227/1.08/build/rtsx.o] Error 1

Downgrading to a 6.9.10 kernel is OK.

Dear Robert, could you update the package to point at this repo: https://github.com/strahl/rts5227-dkms It appears to work fine on recent kernels. If you have no time for this I can maintain this package, ping me on github alexpyattaev.

I patched the driver in https://github.com/strahl/rts5227-dkms for 5.16 here: https://github.com/ewisuri/rts5227-dkms

i've tried replacing https://github.com/Zibri/Realtek-rts5229-linux-driver with https://github.com/strahl/rts5227-dkms in source, and 5229 with 5227 in $sourceDir/blacklist-rts5229.conf, but it still does not compile

hello! since the first kernel 5.16 I'm getting this error and sd card stops being mounted after cold/hard boot.

==> dkms install --no-depmod rts5227/1.07 -k 5.16.1-arch1-1 Deprecated feature: REMAKE_INITRD Error! Bad return status for module build on kernel: 5.16.1-arch1-1 (x86_64) Consult /var/lib/dkms/rts5227/1.07/build/make.log for more information. ==> WARNING: `dkms install --no-depmod rts5227/1.07 -k 5.16.1-arch1-1' exited 10

@Shara lspci output is not a valid way to debug these kind of issues."Unassigned class" string also appears on working devices.

@LittleHuba this issue was fixed and can't reproduce. Please reinstall the package and everything should work flawlessly

Build issue with newer kernels:

/var/lib/dkms/rts5227/1.07/build/rtsx.c: In Funktion »rtsx_probe«:
/var/lib/dkms/rts5227/1.07/build/rtsx.c:940:20: Fehler: Implizite Deklaration der Funktion »ioremap_nocache«; meinten Sie »ioremap_cache«? [-Werror=implicit-function-declaration]
  940 |  dev->remap_addr = ioremap_nocache(dev->addr, pci_resource_len(pci,0));
      |                    ^~~~~~~~~~~~~~~
      |                    ioremap_cache