On Tuesday, June 11, 2019 at 1:12:04 PM UTC-4, Nadav Har'El wrote:
>
>
> On Wed, Jun 5, 2019 at 7:33 AM Waldemar Kozaczuk <jwkoz...@gmail.com
> <javascript:>> wrote:
>
>> This patch improves memory utilization by
>> making OSv use area below where kernel is loaded.
>>
>> This normally saves only 1M + 640K (low memory)
>> howewer makes bigger difference when kernel_base in
>> Makefile is bumped higher. So if one changes kernel_base
>> to 10M then before this patch we would waste almost 10M of
>> memory.
>>
>> Fixes #1012
>>
>
> Thanks. I committed this, and tested that it actually allows me to run my
> favorite application (rogue ;-))
> with 1MB less.
>
> I'm curious if there isn't a problem with the physical address "0". Is it
> ok that we go all the way to 0, or
> does this code only starts somewhere else (I don't remember any details
> about this code... sorry...).
>
I had the same thought.
I have just looked at ```page_pool::free_initial_memory_range``` :
void free_initial_memory_range(void* addr, size_t size)
{
if (!size) {
return;
}
if (addr == nullptr) {
++addr;
--size;
}
auto a = reinterpret_cast<uintptr_t>(addr);
auto delta = align_up(a, page_size) - a;
if (delta > size) {
return;
}
addr += delta;
size -= delta;
size = align_down(size, page_size);
if (!size) {
return;
}
on_new_memory(size);
on_free(size);
auto pr = new (addr) page_range(size);
free_page_ranges.initial_add(pr);
}
And it seems to a handle addr 0 (the range 0-640KiB) but I think there may
be a bug in there - it bumps addr by 1, but the size should be decremented
by 8 not 1, am I wrong?
>
>> Signed-off-by: Waldemar Kozaczuk <jwkoz...@gmail.com <javascript:>>
>> ---
>> arch/x64/arch-setup.cc | 14 ++++++++++++--
>> 1 file changed, 12 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/x64/arch-setup.cc b/arch/x64/arch-setup.cc
>> index 9317dd25..62236486 100644
>> --- a/arch/x64/arch-setup.cc
>> +++ b/arch/x64/arch-setup.cc
>> @@ -130,7 +130,7 @@ void arch_setup_free_memory()
>> // freed.
>> for_each_e820_entry(e820_buffer, e820_size, [] (e820ent ent) {
>> // can't free anything below edata, it's core code.
>> - // FIXME: can free below 2MB.
>> + // can't free anything below kernel at this moment
>> if (ent.addr + ent.size <= edata) {
>> return;
>> }
>> @@ -159,6 +159,16 @@ void arch_setup_free_memory()
>> // now that we have some free memory, we can start mapping the rest
>> mmu::switch_to_runtime_page_tables();
>> for_each_e820_entry(e820_buffer, e820_size, [] (e820ent ent) {
>> + //
>> + // Free the memory below elf_start which we could not before
>> + if (ent.addr < (u64)elf_start) {
>> + if (ent.addr + ent.size >= (u64)elf_start) {
>> + ent = truncate_above(ent, (u64) elf_start);
>> + }
>> + mmu::free_initial_memory_range(ent.addr, ent.size);
>> + return;
>> + }
>> + //
>> // Ignore memory already freed above
>> if (ent.addr + ent.size <= initial_map) {
>> return;
>> @@ -167,7 +177,7 @@ void arch_setup_free_memory()
>> ent = truncate_below(ent, initial_map);
>> }
>> for (auto&& area : mmu::identity_mapped_areas) {
>> - auto base = reinterpret_cast<void*>(get_mem_area_base(area));
>> + auto base = reinterpret_cast<void*>(get_mem_area_base(area));
>> mmu::linear_map(base + ent.addr, ent.addr, ent.size, ~0);
>> }
>> mmu::free_initial_memory_range(ent.addr, ent.size);
>> --
>> 2.20.1
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "OSv Development" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to osv...@googlegroups.com <javascript:>.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/osv-dev/20190605043342.31586-1-jwkozaczuk%40gmail.com
>> .
>> For more options, visit https://groups.google.com/d/optout.
>>
>
--
You received this message because you are subscribed to the Google Groups "OSv
Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to osv-dev+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/osv-dev/a2f06ade-cec4-47a5-91ff-552000871dd5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
