Re: [PATCH v4] memblock: Add flags and nid info in memblock debugfs
From: Mike Rapoport
Date: Wed May 24 2023 - 04:38:58 EST
On Wed, May 24, 2023 at 09:42:40AM +0530, Anshuman Khandual wrote:
> On 5/23/23 21:35, Mike Rapoport wrote:
> > Hi,
> >
> > On Fri, May 19, 2023 at 06:53:21PM +0800, Yuwei Guan wrote:
> >> Currently, the memblock debugfs can display the count of memblock_type and
> >> the base and end of the reg. However, when memblock_mark_*() or
> >> memblock_set_node() is executed on some range, the information in the
> >> existing debugfs cannot make it clear why the address is not consecutive.
> >>
> >> For example,
> >> cat /sys/kernel/debug/memblock/memory
> >> 0: 0x0000000080000000..0x00000000901fffff
> >> 1: 0x0000000090200000..0x00000000905fffff
> >> 2: 0x0000000090600000..0x0000000092ffffff
> >> 3: 0x0000000093000000..0x00000000973fffff
> >> 4: 0x0000000097400000..0x00000000b71fffff
> >> 5: 0x00000000c0000000..0x00000000dfffffff
> >> 6: 0x00000000e2500000..0x00000000f87fffff
> >> 7: 0x00000000f8800000..0x00000000fa7fffff
> >> 8: 0x00000000fa800000..0x00000000fd3effff
> >> 9: 0x00000000fd3f0000..0x00000000fd3fefff
> >> 10: 0x00000000fd3ff000..0x00000000fd7fffff
> >> 11: 0x00000000fd800000..0x00000000fd901fff
> >> 12: 0x00000000fd902000..0x00000000fd909fff
> >> 13: 0x00000000fd90a000..0x00000000fd90bfff
> >> 14: 0x00000000fd90c000..0x00000000ffffffff
> >> 15: 0x0000000880000000..0x0000000affffffff
> >>
> >> So we can add flags and nid to this debugfs.
> >>
> >> For example,
> >> cat /sys/kernel/debug/memblock/memory
> >> 0: 0x0000000080000000..0x00000000901fffff 0 NONE
> >> 1: 0x0000000090200000..0x00000000905fffff 0 NOMAP
> >> 2: 0x0000000090600000..0x0000000092ffffff 0 NONE
> >> 3: 0x0000000093000000..0x00000000973fffff 0 NOMAP
> >> 4: 0x0000000097400000..0x00000000b71fffff 0 NONE
> >> 5: 0x00000000c0000000..0x00000000dfffffff 0 NONE
> >> 6: 0x00000000e2500000..0x00000000f87fffff 0 NONE
> >> 7: 0x00000000f8800000..0x00000000fa7fffff 0 NOMAP
> >> 8: 0x00000000fa800000..0x00000000fd3effff 0 NONE
> >> 9: 0x00000000fd3f0000..0x00000000fd3fefff 0 NOMAP
> >> 10: 0x00000000fd3ff000..0x00000000fd7fffff 0 NONE
> >> 11: 0x00000000fd800000..0x00000000fd901fff 0 NOMAP
> >> 12: 0x00000000fd902000..0x00000000fd909fff 0 NONE
> >> 13: 0x00000000fd90a000..0x00000000fd90bfff 0 NOMAP
> >> 14: 0x00000000fd90c000..0x00000000ffffffff 0 NONE
> >> 15: 0x0000000880000000..0x0000000affffffff 0 NONE
> >>
> >> Signed-off-by: Yuwei Guan <ssawgyw@xxxxxxxxx>
> >> ---
> >> v4:
> >> - show string value for each memblock flag
> >> ---
> >> mm/memblock.c | 24 ++++++++++++++++++++++--
> >> 1 file changed, 22 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/mm/memblock.c b/mm/memblock.c
> >> index 511d4783dcf1..10d0ddbeebc1 100644
> >> --- a/mm/memblock.c
> >> +++ b/mm/memblock.c
> >> @@ -2136,12 +2136,19 @@ void __init memblock_free_all(void)
> >> }
> >>
> >> #if defined(CONFIG_DEBUG_FS) && defined(CONFIG_ARCH_KEEP_MEMBLOCK)
> >> +static const char * const flagname[] = {
> >> + [ilog2(MEMBLOCK_HOTPLUG)] = "HOTPLUG",
> >> + [ilog2(MEMBLOCK_MIRROR)] = "MIRROR",
> >> + [ilog2(MEMBLOCK_NOMAP)] = "NOMAP",
> >> + [ilog2(MEMBLOCK_DRIVER_MANAGED)] = "DRV_MNG",
> >> +};
> >>
> >> static int memblock_debug_show(struct seq_file *m, void *private)
> >> {
> >> struct memblock_type *type = m->private;
> >> struct memblock_region *reg;
> >> - int i;
> >> + int i, j;
> >> + unsigned int count = ARRAY_SIZE(flagname);
> >> phys_addr_t end;
> >>
> >> for (i = 0; i < type->cnt; i++) {
> >> @@ -2149,7 +2156,20 @@ static int memblock_debug_show(struct seq_file *m, void *private)
> >> end = reg->base + reg->size - 1;
> >>
> >> seq_printf(m, "%4d: ", i);
> >> - seq_printf(m, "%pa..%pa\n", ®->base, &end);
> >> + seq_printf(m, "%pa..%pa ", ®->base, &end);
> >> + seq_printf(m, "%4d ", memblock_get_region_node(reg));
> >> + if (reg->flags) {
> >> + for (j = 0; j < count; j++) {
> >> + if (reg->flags & (1U << j)) {
> >> + seq_printf(m, "%s\n", flagname[j]);
> >> + break;
> >
> > The flags are mostly mutually exclusive because it's unlikely they are used
> > together, but not because there are some restrictions on possible flags
> > combinations. So generally it's possible to have multiple flags set on the
> > same region.
>
> But do we really need to account for that as of now, when no memblock region
> carries more than a single flag ?
Right now we don't care indeed, I'm a bit concerned of what will happen if
we'd have regions with multiple flags set.
> But in that case all flags here need to be printed with "|" in between ?
We could use space rather than "|" for simplicity, but thinking more about
it, I'm inclined to just take this version and we'll deal with multiple
flags if/when required.
> > Sorry, I missed that in v3.
> >
> >> + }
> >> + }
> >> + if (j == count)
> >> + seq_printf(m, "%s\n", "UNKNOWN");
> >> + } else {
> >> + seq_printf(m, "%s\n", "NONE");
> >> + }
> >> }
> >> return 0;
> >> }
> >> --
> >> 2.34.1
> >>
> >
--
Sincerely yours,
Mike.