Re: [Xen-devel] [PATCH v2 1/2] libxl: use xen-blkback for 'vbd' disk types by default

[Marek Marczykowski-Górecki]

On Mon, Jul 31, 2017 at 05:01:08PM +0100, Wei Liu wrote:
> On Mon, Jul 31, 2017 at 04:56:04PM +0100, Wei Liu wrote:
> > On Fri, Jul 28, 2017 at 06:42:13PM +0200, Marek Marczykowski-Górecki wrote:
> > > This will allow later to make HVM domain without qemu in dom0 (in
> > > addition to the one in stubdomain).
> > > 
> > > Signed-off-by: Marek Marczykowski-Górecki 
> > > 
> > > 
> > > ---
> > > This is extracted from v1 of "libxl: do not start dom0 qemu for
> > > stubdomain when not needed".
> > > 
> > > Signed-off-by: Marek Marczykowski-Górecki 
> > > 
> > > ---
> > >  tools/libxl/libxl_disk.c | 4 +++-
> > >  1 file changed, 3 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/tools/libxl/libxl_disk.c b/tools/libxl/libxl_disk.c
> > > index 63de75c..7842d9b 100644
> > > --- a/tools/libxl/libxl_disk.c
> > > +++ b/tools/libxl/libxl_disk.c
> > > @@ -56,10 +56,12 @@ static void disk_eject_xswatch_callback(libxl__egc 
> > > *egc, libxl__ev_xswatch *w,
> > >  "/local/domain/%d/backend/%" TOSTRING(BACKEND_STRING_SIZE)
> > > "[a-z]/%*d/%*d",
> > > >backend_domid, backend_type);
> > > -if (!strcmp(backend_type, "tap") || !strcmp(backend_type, "vbd")) {
> > > +if (!strcmp(backend_type, "tap")) {
> > >  disk->backend = LIBXL_DISK_BACKEND_TAP;
> > >  } else if (!strcmp(backend_type, "qdisk")) {
> > >  disk->backend = LIBXL_DISK_BACKEND_QDISK;
> > > +} else if (!strcmp(backend_type, "vbd")) {
> > > +disk->backend = LIBXL_DISK_BACKEND_PHY;
> > 
> > Wait, it only occurred to me until now this patch is changing
> > disk_eject_xswatch_callback.
> > 
> > Is this a bug fix? How is it possible for the backend_type to be "vbd"
> > when there isn't such thing in libxl_types.idl?
> 
> Oh, I'm an idiot. That's read from xenstore path. I think this patch is
> correct. But I still tend to think this is a bug fix. How do you
> discover this problem? Has disk eject ever worked for you?

Hmm, I think I've failed with forward-porting this patch. In current Xen
it isn't possible to create cdrom with backend_type!=qdisk...

So, this patch alone isn't sufficient. Please ignore it.

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?


signature.asc
Description: PGP signature
___
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

Re: [Xen-devel] [PATCH v2 1/2] libxl: use xen-blkback for 'vbd' disk types by default

[Wei Liu]

On Mon, Jul 31, 2017 at 04:56:04PM +0100, Wei Liu wrote:
> On Fri, Jul 28, 2017 at 06:42:13PM +0200, Marek Marczykowski-Górecki wrote:
> > This will allow later to make HVM domain without qemu in dom0 (in
> > addition to the one in stubdomain).
> > 
> > Signed-off-by: Marek Marczykowski-Górecki 
> > 
> > ---
> > This is extracted from v1 of "libxl: do not start dom0 qemu for
> > stubdomain when not needed".
> > 
> > Signed-off-by: Marek Marczykowski-Górecki 
> > ---
> >  tools/libxl/libxl_disk.c | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> > 
> > diff --git a/tools/libxl/libxl_disk.c b/tools/libxl/libxl_disk.c
> > index 63de75c..7842d9b 100644
> > --- a/tools/libxl/libxl_disk.c
> > +++ b/tools/libxl/libxl_disk.c
> > @@ -56,10 +56,12 @@ static void disk_eject_xswatch_callback(libxl__egc 
> > *egc, libxl__ev_xswatch *w,
> >  "/local/domain/%d/backend/%" TOSTRING(BACKEND_STRING_SIZE)
> > "[a-z]/%*d/%*d",
> > >backend_domid, backend_type);
> > -if (!strcmp(backend_type, "tap") || !strcmp(backend_type, "vbd")) {
> > +if (!strcmp(backend_type, "tap")) {
> >  disk->backend = LIBXL_DISK_BACKEND_TAP;
> >  } else if (!strcmp(backend_type, "qdisk")) {
> >  disk->backend = LIBXL_DISK_BACKEND_QDISK;
> > +} else if (!strcmp(backend_type, "vbd")) {
> > +disk->backend = LIBXL_DISK_BACKEND_PHY;
> 
> Wait, it only occurred to me until now this patch is changing
> disk_eject_xswatch_callback.
> 
> Is this a bug fix? How is it possible for the backend_type to be "vbd"
> when there isn't such thing in libxl_types.idl?

Oh, I'm an idiot. That's read from xenstore path. I think this patch is
correct. But I still tend to think this is a bug fix. How do you
discover this problem? Has disk eject ever worked for you?

___
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

Re: [Xen-devel] [PATCH v2 1/2] libxl: use xen-blkback for 'vbd' disk types by default

[Wei Liu]

On Fri, Jul 28, 2017 at 06:42:13PM +0200, Marek Marczykowski-Górecki wrote:
> This will allow later to make HVM domain without qemu in dom0 (in
> addition to the one in stubdomain).
> 
> Signed-off-by: Marek Marczykowski-Górecki 
> 
> ---
> This is extracted from v1 of "libxl: do not start dom0 qemu for
> stubdomain when not needed".
> 
> Signed-off-by: Marek Marczykowski-Górecki 
> ---
>  tools/libxl/libxl_disk.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/libxl/libxl_disk.c b/tools/libxl/libxl_disk.c
> index 63de75c..7842d9b 100644
> --- a/tools/libxl/libxl_disk.c
> +++ b/tools/libxl/libxl_disk.c
> @@ -56,10 +56,12 @@ static void disk_eject_xswatch_callback(libxl__egc *egc, 
> libxl__ev_xswatch *w,
>  "/local/domain/%d/backend/%" TOSTRING(BACKEND_STRING_SIZE)
> "[a-z]/%*d/%*d",
> >backend_domid, backend_type);
> -if (!strcmp(backend_type, "tap") || !strcmp(backend_type, "vbd")) {
> +if (!strcmp(backend_type, "tap")) {
>  disk->backend = LIBXL_DISK_BACKEND_TAP;
>  } else if (!strcmp(backend_type, "qdisk")) {
>  disk->backend = LIBXL_DISK_BACKEND_QDISK;
> +} else if (!strcmp(backend_type, "vbd")) {
> +disk->backend = LIBXL_DISK_BACKEND_PHY;

Wait, it only occurred to me until now this patch is changing
disk_eject_xswatch_callback.

Is this a bug fix? How is it possible for the backend_type to be "vbd"
when there isn't such thing in libxl_types.idl?

___
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

[Xen-devel] [PATCH v2 1/2] libxl: use xen-blkback for 'vbd' disk types by default

[Marek Marczykowski-Górecki]

This will allow later to make HVM domain without qemu in dom0 (in
addition to the one in stubdomain).

Signed-off-by: Marek Marczykowski-Górecki 

---
This is extracted from v1 of "libxl: do not start dom0 qemu for
stubdomain when not needed".

Signed-off-by: Marek Marczykowski-Górecki 
---
 tools/libxl/libxl_disk.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/tools/libxl/libxl_disk.c b/tools/libxl/libxl_disk.c
index 63de75c..7842d9b 100644
--- a/tools/libxl/libxl_disk.c
+++ b/tools/libxl/libxl_disk.c
@@ -56,10 +56,12 @@ static void disk_eject_xswatch_callback(libxl__egc *egc, 
libxl__ev_xswatch *w,
 "/local/domain/%d/backend/%" TOSTRING(BACKEND_STRING_SIZE)
"[a-z]/%*d/%*d",
>backend_domid, backend_type);
-if (!strcmp(backend_type, "tap") || !strcmp(backend_type, "vbd")) {
+if (!strcmp(backend_type, "tap")) {
 disk->backend = LIBXL_DISK_BACKEND_TAP;
 } else if (!strcmp(backend_type, "qdisk")) {
 disk->backend = LIBXL_DISK_BACKEND_QDISK;
+} else if (!strcmp(backend_type, "vbd")) {
+disk->backend = LIBXL_DISK_BACKEND_PHY;
 } else {
 disk->backend = LIBXL_DISK_BACKEND_UNKNOWN;
 }
-- 
2.7.5


___
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel