Re: [PATCH 5/9] vfio/pci: Provide a user-facing name for BAR mappings
From: Matt Evans
Date: Thu May 07 2026 - 13:18:18 EST
Sent too soon, :|
On 07/05/2026 17:56, Matt Evans wrote:
Hi Alex,
On 01/05/2026 23:44, Alex Williamson wrote:
On Thu, 16 Apr 2026 06:17:48 -0700
Matt Evans <mattev@xxxxxxxx> wrote:
Since converting BAR mmap()s to using DMABUFs, we lose the original
device path in /proc/<pid>/maps, lsof, etc. Generate a debug-oriented
synthetic 'filename' based on the cdev, plus BDF, plus resource index.
This applies only to BAR mappings via the VFIO device fd, as
explicitly-exported DMABUFs are named by userspace via the
DMA_BUF_SET_NAME ioctl.
Signed-off-by: Matt Evans <mattev@xxxxxxxx>
---
drivers/vfio/pci/vfio_pci_dmabuf.c | 27 +++++++++++++++++++++++++--
1 file changed, 25 insertions(+), 2 deletions(-)
diff --git a/drivers/vfio/pci/vfio_pci_dmabuf.c b/drivers/vfio/pci/ vfio_pci_dmabuf.c
index a12432825e5e..04c7733fe712 100644
--- a/drivers/vfio/pci/vfio_pci_dmabuf.c
+++ b/drivers/vfio/pci/vfio_pci_dmabuf.c
@@ -4,6 +4,7 @@
#include <linux/dma-buf-mapping.h>
#include <linux/pci-p2pdma.h>
#include <linux/dma-resv.h>
+#include <uapi/linux/dma-buf.h>
#include "vfio_pci_priv.h"
@@ -467,6 +468,7 @@ int vfio_pci_core_mmap_prep_dmabuf(struct vfio_pci_core_device *vdev,
{
struct vfio_pci_dma_buf *priv;
const unsigned int nr_ranges = 1;
+ char *bufname;
int ret;
priv = kzalloc_obj(*priv);
@@ -479,6 +481,20 @@ int vfio_pci_core_mmap_prep_dmabuf(struct vfio_pci_core_device *vdev,
goto err_free_priv;
}
+ bufname = kzalloc(DMA_BUF_NAME_LEN, GFP_KERNEL);
+ if (!bufname) {
+ ret = -ENOMEM;
+ goto err_free_phys;
+ }
+
+ /*
+ * Maximum size of the friendly debug name is
+ * vfio1234567890:ffff:ff:3f.7-9 = 30, which fits within
+ * DMA_BUF_NAME_LEN.
+ */
+ snprintf(bufname, DMA_BUF_NAME_LEN, "%s:%s/%x",
+ dev_name(&vdev->vdev.device), pci_name(vdev->pdev), res_index);
Comment suggests 9 is the max res_index that can be printed, but mmap
only directly supports standard BARs 0-5. Comment also uses a '-'
while the code uses a '/'. Thanks,
Right you are. Fixed, but, since...
https://lore.kernel.org/kvm/52162da4-e1cc-4f90-a95a-218d6089cd71@xxxxxxxx/
...I'm keeping the resource index encoded in the vm_pgoffs and as that's
in /proc/<pid>/maps it doesn't need to be in the name. I.e., an example
mapping of BAR 2 looks like:
ffffa9330000-ffffad300000 rw-s 20000030000 00:0b 12 / dmabuf:vfio0:0000:00:03.0
BUT, the name's visible via paths other than just /proc/<pid>/maps, e.g.
/sys/kernel/debug/dma_buf/bufinfo or /proc/<pid>/map_files which don't have the vm_offs, and so back to plan A. Just made the comment consistent.
Matt