| Message ID | 20200227204023.22125-5-grimm@linux.ibm.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | Ultravisor support in skiboot | expand |
| Context | Check | Description |
|---|---|---|
| snowpatch_ozlabs/apply_patch | success | Successfully applied on branch master (82aed17a5468aff6b600ee1694a10a60f942c018) |
| snowpatch_ozlabs/snowpatch_job_snowpatch-skiboot | success | Test snowpatch/job/snowpatch-skiboot on branch master |
| snowpatch_ozlabs/snowpatch_job_snowpatch-skiboot-dco | success | Signed-off-by present |
On 28/02/2020 07:40, Ryan Grimm wrote: > We can jump into any location in the uv image for testing. and how can a user set this "uv-entry"? I am guessing cronus (which most of us do not have). I'd rather have this merged into 2/16 or dropped (for debugging imho it is lot easier to fix the line, compile and copy the uv lid over to the bmc). > > Signed-off-by: Ryan Grimm <grimm@linux.ibm.com> > --- > hw/ultravisor.c | 7 ++++++- > 1 file changed, 6 insertions(+), 1 deletion(-) > > diff --git a/hw/ultravisor.c b/hw/ultravisor.c > index 7f263e3e..8dea456a 100644 > --- a/hw/ultravisor.c > +++ b/hw/ultravisor.c > @@ -14,8 +14,13 @@ static uint64_t uv_base_addr; > > static void cpu_start_ultravisor(void *fdt) > { > + uint64_t uv_entry = 0; > + > + if (dt_find_property(uv_fw_node, "uv-entry")) > + uv_entry = dt_prop_get_u64(uv_fw_node, "uv-entry"); > + > prlog(PR_DEBUG, "UV: Starting on CPU 0x%04x\n", this_cpu()->pir); > - start_uv(uv_base_addr, fdt); > + start_uv(uv_base_addr + uv_entry, fdt); > } > > int start_ultravisor(void *fdt) >
On Wed, 2020-03-04 at 17:44 +1100, Alexey Kardashevskiy wrote: > > On 28/02/2020 07:40, Ryan Grimm wrote: > > We can jump into any location in the uv image for testing. > > > and how can a user set this "uv-entry"? I am guessing cronus (which > most > of us do not have). I'd rather have this merged into 2/16 or dropped > (for debugging imho it is lot easier to fix the line, compile and > copy > the uv lid over to the bmc). > You can use Mambo too. My intent was to be able to actually test something in Mambo without having the ultravisor tree. We can use skiboot.lid and set uv-entry to disable_pef and it boots, so, I found that useful. But, yeah, squashed to earlier patch. -Ryan > > > > > Signed-off-by: Ryan Grimm <grimm@linux.ibm.com> > > --- > > hw/ultravisor.c | 7 ++++++- > > 1 file changed, 6 insertions(+), 1 deletion(-) > > > > diff --git a/hw/ultravisor.c b/hw/ultravisor.c > > index 7f263e3e..8dea456a 100644 > > --- a/hw/ultravisor.c > > +++ b/hw/ultravisor.c > > @@ -14,8 +14,13 @@ static uint64_t uv_base_addr; > > > > static void cpu_start_ultravisor(void *fdt) > > { > > + uint64_t uv_entry = 0; > > + > > + if (dt_find_property(uv_fw_node, "uv-entry")) > > + uv_entry = dt_prop_get_u64(uv_fw_node, "uv-entry"); > > + > > prlog(PR_DEBUG, "UV: Starting on CPU 0x%04x\n", this_cpu()- > > >pir); > > - start_uv(uv_base_addr, fdt); > > + start_uv(uv_base_addr + uv_entry, fdt); > > } > > > > int start_ultravisor(void *fdt) > > > >
diff --git a/hw/ultravisor.c b/hw/ultravisor.c index 7f263e3e..8dea456a 100644 --- a/hw/ultravisor.c +++ b/hw/ultravisor.c @@ -14,8 +14,13 @@ static uint64_t uv_base_addr; static void cpu_start_ultravisor(void *fdt) { + uint64_t uv_entry = 0; + + if (dt_find_property(uv_fw_node, "uv-entry")) + uv_entry = dt_prop_get_u64(uv_fw_node, "uv-entry"); + prlog(PR_DEBUG, "UV: Starting on CPU 0x%04x\n", this_cpu()->pir); - start_uv(uv_base_addr, fdt); + start_uv(uv_base_addr + uv_entry, fdt); } int start_ultravisor(void *fdt)
We can jump into any location in the uv image for testing. Signed-off-by: Ryan Grimm <grimm@linux.ibm.com> --- hw/ultravisor.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)