On 02.07.19 16:12, Jan Kiszka wrote:
On 02.07.19 15:58, Andrej Utz wrote:
On some systems the config generator permissively maps huge chunks of
PCI Bus MMIO space. This area needs to be intercepted by the hypervisor,
as parts of ivshmem-net devices may be behind that area.
Just the make the boundary conditions clearer: This mapping only happens
when the PCI bus memory region has no device resources so far. If
ivshmem then adds a region later on, we are doomed. But if there is
device already, nothing was broken so far. Am I right?
Correct.
Hence, ignore such regions.
Signed-off-by: Andrej Utz <[email protected]>
---
pyjailhouse/sysfs_parser.py | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/pyjailhouse/sysfs_parser.py b/pyjailhouse/sysfs_parser.py
index f48f249f..4a06ccdc 100644
--- a/pyjailhouse/sysfs_parser.py
+++ b/pyjailhouse/sysfs_parser.py
@@ -105,6 +105,10 @@ def parse_iomem(pcidevices):
ret = []
for r in regions:
+ # filter empty PCI buses
"Filter PCI buses in order to avoid mapping empty ones that might
require interception when becoming non-empty." Or so.
Certainly an improvement.
+ if r.typestr.startswith('PCI Bus'):
+ continue
+
append_r = True
# filter the list for MSI-X pages
for d in pcidevices:
Looks good.
Jan
Andrej
--
You received this message because you are subscribed to the Google Groups
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/jailhouse-dev/caa1784e-d110-ffcc-6fd0-d0aadd71ea04%40st.oth-regensburg.de.
For more options, visit https://groups.google.com/d/optout.
