>>> On 28.08.15 at 16:15, wrote:
> Then I decided to see if I can expand that to also be part of the
> 'tmem_op', which looked legit (as it is the same size and offset and
> pahole wise it looks right).
>
> But sadly the compat layer is not happy with me:
>
>
> In file included from
> /home/ko
> >>> +struct xen_sysctl_tmem_op {
> >>> +uint32_t cmd; /* IN: XEN_SYSCTL_TMEM_OP_* . */
> >>> +int32_t pool_id;/* IN: 0 by default unless _SAVE_*, RESTORE_* .*/
> >>> +uint32_t cli_id;/* IN: client id, 0 for
> >>> XEN_SYSCTL_TMEM_QUERY_FREEABLE_MB
> >>> +
>>> On 27.08.15 at 21:20, wrote:
>> >>> --- a/xen/include/public/tmem.h
>> >>> +++ b/xen/include/public/tmem.h
>> >>> @@ -33,7 +33,7 @@
>> >>> #define TMEM_SPEC_VERSION 1
>> >>>
>> >>> /* Commands to HYPERVISOR_tmem_op() */
>> >>> -#define TMEM_CONTROL 0
>> >>> +#define
>>> On 27.08.15 at 20:43, wrote:
> On Thu, Aug 27, 2015 at 09:11:35AM -0600, Jan Beulich wrote:
>> >>> On 27.08.15 at 13:02, wrote:
>> > --- a/xen/include/public/sysctl.h
>> > +++ b/xen/include/public/sysctl.h
>> > @@ -710,6 +710,48 @@ struct xen_sysctl_psr_cat_op {
>> > typedef struct xen_sysct
> >>> +struct xen_sysctl_tmem_op {
> >>> +uint32_t cmd; /* IN: XEN_SYSCTL_TMEM_OP_* . */
> >>> +int32_t pool_id;/* IN: 0 by default unless _SAVE_*, RESTORE_* .*/
> >>> +uint32_t cli_id;/* IN: client id, 0 for
> >>> XEN_SYSCTL_TMEM_QUERY_FREEABLE_MB
> >>> +
On 27/08/15 19:43, Konrad Rzeszutek Wilk wrote:
>
>>> --- a/xen/include/public/sysctl.h
>>> +++ b/xen/include/public/sysctl.h
>>> @@ -710,6 +710,48 @@ struct xen_sysctl_psr_cat_op {
>>> typedef struct xen_sysctl_psr_cat_op xen_sysctl_psr_cat_op_t;
>>> DEFINE_XEN_GUEST_HANDLE(xen_sysctl_psr_cat_op
On 27/08/15 19:47, Konrad Rzeszutek Wilk wrote:
>
>>> --- a/xen/include/public/sysctl.h
>>> +++ b/xen/include/public/sysctl.h
>>> @@ -710,6 +710,48 @@ struct xen_sysctl_psr_cat_op {
>>> typedef struct xen_sysctl_psr_cat_op xen_sysctl_psr_cat_op_t;
>>> DEFINE_XEN_GUEST_HANDLE(xen_sysctl_psr_cat_op
On Thu, Aug 27, 2015 at 02:12:35PM +0100, Andrew Cooper wrote:
> On 27/08/15 12:02, Konrad Rzeszutek Wilk wrote:
> > --- a/tools/python/xen/lowlevel/xc/xc.c
> > +++ b/tools/python/xen/lowlevel/xc/xc.c
> > @@ -1808,25 +1808,25 @@ static PyObject *pyxc_tmem_control(XcObject *self,
> >
On Thu, Aug 27, 2015 at 09:11:35AM -0600, Jan Beulich wrote:
> >>> On 27.08.15 at 13:02, wrote:
> > @@ -2666,9 +2662,9 @@ long do_tmem_op(tmem_cli_op_t uops)
> > /* Acquire wirte lock for all command at first */
> > write_lock(&tmem_rwlock);
> >
> > -if ( op.cmd == TMEM_CONTROL )
>
>>> On 27.08.15 at 13:02, wrote:
> @@ -2666,9 +2662,9 @@ long do_tmem_op(tmem_cli_op_t uops)
> /* Acquire wirte lock for all command at first */
> write_lock(&tmem_rwlock);
>
> -if ( op.cmd == TMEM_CONTROL )
> +if ( op.cmd == TMEM_CONTROL_MOVED )
> {
> -rc = do_tme
>>> On 27.08.15 at 15:12, wrote:
> On 27/08/15 12:02, Konrad Rzeszutek Wilk wrote:
>> --- a/xen/include/xen/tmem.h
>> +++ b/xen/include/xen/tmem.h
>> @@ -9,6 +9,9 @@
>> #ifndef __XEN_TMEM_H__
>> #define __XEN_TMEM_H__
>>
>> +struct xen_sysctl_tmem_op;
>
> #include please.
Why? Forward declar
On 27/08/15 12:02, Konrad Rzeszutek Wilk wrote:
> --- a/tools/python/xen/lowlevel/xc/xc.c
> +++ b/tools/python/xen/lowlevel/xc/xc.c
> @@ -1808,25 +1808,25 @@ static PyObject *pyxc_tmem_control(XcObject *self,
> &pool_id, &subop, &cli_id, &arg1, &arg2, &buf) )
> ret
On Thu, Aug 27, 2015 at 07:02:00AM -0400, Konrad Rzeszutek Wilk wrote:
> The operations are to be used by an control domain to set parameters,
> list pools, clients, and to be used during migration.
>
> There is no need to have them in the tmem hypercall path.
>
> This patch moves code without ad
The operations are to be used by an control domain to set parameters,
list pools, clients, and to be used during migration.
There is no need to have them in the tmem hypercall path.
This patch moves code without adding fixes - and in fact in
some cases makes the parameters soo long that they hurt
14 matches
Mail list logo
