| Message ID | 20170918101709.30421-3-aik@ozlabs.ru |
|---|---|
| State | New |
| Headers | show |
| Series | memory: Reduce memory use | expand |
On 09/18/2017 05:16 AM, Alexey Kardashevskiy wrote: Subject line: s/Explicitely/Explicitly/ > This is not so mechanical change in order to move to shared FlatViews > so make it a separate patch. Hard to understand; it's unclear whether this is the mechanical addition of a parameter before the next patch does shared FlatViews, or whether the next patch is mechanical by virtue of this patch doing the non-mechanical changes. Maybe: Mechanical change to add a parameter to make it easier for the next patch to share FlatViews. > The first argument of > address_space_do_translate() will become a FlatView, however since > address_space_get_iotlb_entry() still wants AS, hence this change. > > Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> > --- > exec.c | 11 +++++++---- > 1 file changed, 7 insertions(+), 4 deletions(-) >
On 19/09/17 00:28, Eric Blake wrote: > On 09/18/2017 05:16 AM, Alexey Kardashevskiy wrote: > > Subject line: s/Explicitely/Explicitly/ > >> This is not so mechanical change in order to move to shared FlatViews >> so make it a separate patch. > > Hard to understand; it's unclear whether this is the mechanical addition > of a parameter before the next patch does shared FlatViews, or whether > the next patch is mechanical by virtue of this patch doing the > non-mechanical changes. Maybe: > > Mechanical change to add a parameter to make it easier for the next > patch to share FlatViews. I was told in this list that "mechanical" is when simple 's/lalala/mememe/' does the job, when it is bigger than this - then it is "this should not cause any behavioural change" :) I'll add such note where applicable. tbh only 1/13, 11/13 and 13/13 do cause a change in behaviour, others do not. > >> The first argument of >> address_space_do_translate() will become a FlatView, however since >> address_space_get_iotlb_entry() still wants AS, hence this change. >> >> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> >> --- >> exec.c | 11 +++++++---- >> 1 file changed, 7 insertions(+), 4 deletions(-) >>
diff --git a/exec.c b/exec.c index d20c34ca83..bd94248390 100644 --- a/exec.c +++ b/exec.c @@ -477,7 +477,8 @@ static MemoryRegionSection address_space_do_translate(AddressSpace *as, hwaddr *xlat, hwaddr *plen, bool is_write, - bool is_mmio) + bool is_mmio, + AddressSpace **target_as) { IOMMUTLBEntry iotlb; MemoryRegionSection *section; @@ -504,6 +505,7 @@ static MemoryRegionSection address_space_do_translate(AddressSpace *as, } as = iotlb.target_as; + *target_as = iotlb.target_as; } *xlat = addr; @@ -526,7 +528,7 @@ IOMMUTLBEntry address_space_get_iotlb_entry(AddressSpace *as, hwaddr addr, /* This can never be MMIO. */ section = address_space_do_translate(as, addr, &xlat, &plen, - is_write, false); + is_write, false, &as); /* Illegal translation */ if (section.mr == &io_mem_unassigned) { @@ -549,7 +551,7 @@ IOMMUTLBEntry address_space_get_iotlb_entry(AddressSpace *as, hwaddr addr, plen -= 1; return (IOMMUTLBEntry) { - .target_as = section.address_space, + .target_as = as, .iova = addr & ~plen, .translated_addr = xlat & ~plen, .addr_mask = plen, @@ -570,7 +572,8 @@ MemoryRegion *address_space_translate(AddressSpace *as, hwaddr addr, MemoryRegionSection section; /* This can be MMIO, so setup MMIO bit. */ - section = address_space_do_translate(as, addr, xlat, plen, is_write, true); + section = address_space_do_translate(as, addr, xlat, plen, is_write, true, + &as); mr = section.mr; if (xen_enabled() && memory_access_is_direct(mr, is_write)) {
This is not so mechanical change in order to move to shared FlatViews so make it a separate patch. The first argument of address_space_do_translate() will become a FlatView, however since address_space_get_iotlb_entry() still wants AS, hence this change. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> --- exec.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-)