Re: [PATCH v7 9/9] PCI/TSM: Report active IDE streams

[<dan.j.williams@intel.com>]

Jonathan Cameron wrote:
> On Thu, 23 Oct 2025 19:04:18 -0700
> Dan Williams <dan.j.williams@intel.com> wrote:
> 
> > Given that the platform TSM owns IDE Stream ID allocation, report the
> > active streams via the TSM class device. Establish a symlink from the
> > class device to the PCI endpoint device consuming the stream, named by
> > the Stream ID.
> > 
> > Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> > Reviewed-by: Jonathan Cameron <jonathan.cameron@huawei.com>
> Couple of trivial things noticed whilst refreshing my memory.
> 
> > Signed-off-by: Dan Williams <dan.j.williams@intel.com>
> > ---
> 
> > diff --git a/drivers/virt/coco/tsm-core.c b/drivers/virt/coco/tsm-core.c
> > index 4499803cf20d..c0dae531b64f 100644
> > --- a/drivers/virt/coco/tsm-core.c
> > +++ b/drivers/virt/coco/tsm-core.c
> > @@ -2,14 +2,17 @@
> >  /* Copyright(c) 2024 Intel Corporation. All rights reserved. */
> >  
> >  #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
> > +#define dev_fmt(fmt) KBUILD_MODNAME ": " fmt
> 
> Why is this dev_fmt() in this patch (which doesn't seem to introduce
> anything that would use it)?

Probably a debug leftover. The dev_err() in tsm_register_pci_or_reset()
would use it, but that print is sufficiently descriptive.

> >  
> >  #include <linux/tsm.h>
> >  #include <linux/idr.h>
> > +#include <linux/pci.h>
> >  #include <linux/rwsem.h>
> >  #include <linux/device.h>
> >  #include <linux/module.h>
> >  #include <linux/cleanup.h>
> >  #include <linux/pci-tsm.h>
> > +#include <linux/pci-ide.h>
> >  
> >  static struct class *tsm_class;
> >  static DECLARE_RWSEM(tsm_rwsem);
> > @@ -106,6 +109,32 @@ void tsm_unregister(struct tsm_dev *tsm_dev)
> >  }
> >  EXPORT_SYMBOL_GPL(tsm_unregister);
> >  
> > +/* must be invoked between tsm_register / tsm_unregister */
> > +int tsm_ide_stream_register(struct pci_ide *ide)
> > +{
> > +	struct pci_dev *pdev = ide->pdev;
> > +	struct pci_tsm *tsm = pdev->tsm;
> > +	struct tsm_dev *tsm_dev = tsm->tsm_dev;
> > +	int rc;
> > +
> > +	rc = sysfs_create_link(&tsm_dev->dev.kobj, &pdev->dev.kobj, ide->name);
> > +	if (rc)
> > +		return rc;
> > +
> > +	ide->tsm_dev = tsm_dev;
> > +	return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(tsm_ide_stream_register);
> > +
> > +void tsm_ide_stream_unregister(struct pci_ide *ide)
> > +{
> > +	struct tsm_dev *tsm_dev = ide->tsm_dev;
> > +
> > +	sysfs_remove_link(&tsm_dev->dev.kobj, ide->name);
> > +	ide->tsm_dev = NULL;
> 
> Trivial preference for reverse order of register.  That means
> setting this NULL before removing the link.

Ok.

 9:  c24f0c9f0b9c !  9:  35aaec30c413 PCI/TSM: Report active IDE streams
    @@ drivers/pci/ide.c: void pci_ide_stream_release(struct pci_ide *ide)
     
      ## drivers/virt/coco/tsm-core.c ##
     @@
    - /* Copyright(c) 2024 Intel Corporation. All rights reserved. */
    - 
    - #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
    -+#define dev_fmt(fmt) KBUILD_MODNAME ": " fmt
      
      #include <linux/tsm.h>
      #include <linux/idr.h>
    @@ drivers/virt/coco/tsm-core.c: void tsm_unregister(struct tsm_dev *tsm_dev)
     +{
     +	struct tsm_dev *tsm_dev = ide->tsm_dev;
     +
    -+	sysfs_remove_link(&tsm_dev->dev.kobj, ide->name);
     +	ide->tsm_dev = NULL;
    ++	sysfs_remove_link(&tsm_dev->dev.kobj, ide->name);
     +}
     +EXPORT_SYMBOL_GPL(tsm_ide_stream_unregister);
     +