Re: [PATCH -next] acpi/hmat: fix memory leaks in hmat_init()

[Qian Cai]

On Tue, 2019-04-09 at 10:25 +0200, Rafael J. Wysocki wrote:
> On Sat, Apr 6, 2019 at 8:18 PM Qian Cai <cai@xxxxxx> wrote:
> > 
> > The commit 665ac7e92757 ("acpi/hmat: Register processor domain to its
> > memory") introduced some memory leaks below due to it fails to release
> > the heap memory in an error path, and then the stack __initdata memory
> > which reference them get freed during boot renders those heap memory as
> > leaks.
> > 
> > unreferenced object 0xc8ff8008349e9400 (size 128):
> > Â comm "swapper/0", pid 1, jiffies 4294709236 (age 48121.476s)
> > Â hex dump (first 32 bytes):
> > ÂÂÂÂ00 d0 9e 34 08 80 ff 84 d8 00 43 11 00 10 ff ffÂÂ...4......C.....
> > ÂÂÂÂ00 00 00 00 ff ff ff ff 00 00 00 00 00 00 00 00ÂÂ................
> > Â backtrace:
> > ÂÂÂÂ[<00000000869d4503>] __kmalloc+0x568/0x600
> > ÂÂÂÂ[<0000000070fd6afb>] alloc_memory_target+0x50/0xd8
> > ÂÂÂÂ[<00000000efa2081e>] srat_parse_mem_affinity+0x58/0x5c
> > ÂÂÂÂ[<000000008bfaef74>] acpi_parse_entries_array+0x1c8/0x2c0
> > ÂÂÂÂ[<0000000022804877>] acpi_table_parse_entries_array+0x11c/0x138
> > ÂÂÂÂ[<00000000ffe9cd34>] acpi_table_parse_entries+0x7c/0xac
> > ÂÂÂÂ[<00000000a7023afd>] hmat_init+0x90/0x174
> > ÂÂÂÂ[<00000000694a86c1>] do_one_initcall+0x2d8/0x5f8
> > ÂÂÂÂ[<0000000024889da9>] do_initcall_level+0x37c/0x3fc
> > ÂÂÂÂ[<000000009be02908>] do_basic_setup+0x38/0x50
> > ÂÂÂÂ[<0000000037b3ac0a>] kernel_init_freeable+0x194/0x258
> > ÂÂÂÂ[<00000000f5741184>] kernel_init+0x18/0x334
> > ÂÂÂÂ[<000000007b30f423>] ret_from_fork+0x10/0x18
> > ÂÂÂÂ[<000000006c7147a8>] 0xffffffffffffffff
> > 
> > Signed-off-by: Qian Cai <cai@xxxxxx>
> 
> Well, the catch is good, but the additional label is sort of excessive IMO.
> 
> > ---
> > Âdrivers/acpi/hmat/hmat.c | 5 +++--
> > Â1 file changed, 3 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/acpi/hmat/hmat.c b/drivers/acpi/hmat/hmat.c
> > index b7824a0309f7..c9b8abcf012c 100644
> > --- a/drivers/acpi/hmat/hmat.c
> > +++ b/drivers/acpi/hmat/hmat.c
> > @@ -637,7 +637,7 @@ static __init int hmat_init(void)
> > 
> > ÂÂÂÂÂÂÂÂstatus = acpi_get_table(ACPI_SIG_HMAT, 0, &tbl);
> > ÂÂÂÂÂÂÂÂif (ACPI_FAILURE(status))
> 
> I would just add a hmat_free_structures() call here or rework the
> _put_table() logic which is not really straightforward.

I am not sure how adding a hmat_free_structures() call there would be better. If
the name has bee changed to something else in the future, it ends up needing to
change 2 places. It also not save the adding line-count either. I don't see how
to rework the acpi_put_table() logic to be better than adding a label here
either.

> 
> > -ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂreturn 0;
> > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂgoto out_free;
> > 
> > ÂÂÂÂÂÂÂÂhmat_revision = tbl->revision;
> > ÂÂÂÂÂÂÂÂswitch (hmat_revision) {
> > @@ -659,8 +659,9 @@ static __init int hmat_init(void)
> > ÂÂÂÂÂÂÂÂ}
> > ÂÂÂÂÂÂÂÂhmat_register_targets();
> > Âout_put:
> > -ÂÂÂÂÂÂÂhmat_free_structures();
> > ÂÂÂÂÂÂÂÂacpi_put_table(tbl);
> > +out_free:
> > +ÂÂÂÂÂÂÂhmat_free_structures();
> > ÂÂÂÂÂÂÂÂreturn 0;
> > Â}
> > Âsubsys_initcall(hmat_init);
> > --