| Message ID | 1467593044-10600-1-git-send-email-oohall@gmail.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
On 04/07/16 10:44, Oliver O'Halloran wrote: > This patch adds the kernel command line parameter "no_tb_segs" which > forces the kernel to use 256MB rather than 1TB segments. Forcing the use > of 256MB segments makes it considerably easier to test code that depends > on an SLB miss occurring. > > Suggested-by: Michael Neuling <mikey@neuling.org> > Suggested-by: Michael Ellerman <mpe@ellerman.id.au> > Signed-off-by: Oliver O'Halloran <oohall@gmail.com> > --- > arch/powerpc/mm/hash_utils_64.c | 15 +++++++++++++++ > 1 file changed, 15 insertions(+) > > diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c > index 5b22ba0b58bc..6da1a9d18e15 100644 > --- a/arch/powerpc/mm/hash_utils_64.c > +++ b/arch/powerpc/mm/hash_utils_64.c > @@ -321,6 +321,15 @@ int htab_remove_mapping(unsigned long vstart, unsigned long vend, > return ret; > } > > +static bool no_tb_segs = false; > + > +static int __init parse_no_tb_segs(char *p) > +{ > + no_tb_segs = true; > + return 0; > +} > +early_param("no_tb_segs", parse_no_tb_segs); Please update Documentation/kernel-paramaters.txt as well and document the use case and the fact that this for debugging. > + > static int __init htab_dt_scan_seg_sizes(unsigned long node, > const char *uname, int depth, > void *data) > @@ -339,6 +348,12 @@ static int __init htab_dt_scan_seg_sizes(unsigned long node, > for (; size >= 4; size -= 4, ++prop) { > if (be32_to_cpu(prop[0]) == 40) { > DBG("1T segment support detected\n"); > + > + if (no_tb_segs) { > + DBG("Forcing 256MB segments\n"); > + break; > + } > + > cur_cpu_spec->mmu_features |= MMU_FTR_1T_SEGMENT; > return 1; > } > Otherwise Acked-by: Balbir Singh <bsingharora@gmail.com> Balbir
On Mon, 2016-04-07 at 00:44:04 UTC, Oliver O'Halloran wrote: > This patch adds the kernel command line parameter "no_tb_segs" which > forces the kernel to use 256MB rather than 1TB segments. Forcing the use > of 256MB segments makes it considerably easier to test code that depends > on an SLB miss occurring. That's a fairly awful name :) Can you call it "disable_1T_segments" ? It should also be mentioned in Documentation/kernel-parameters.txt, with emphasis that it's a debug option and powerpc only. cheers
On Mon, 2016-07-04 at 16:09 +1000, Michael Ellerman wrote: > On Mon, 2016-04-07 at 00:44:04 UTC, Oliver O'Halloran wrote: > > > > This patch adds the kernel command line parameter "no_tb_segs" which > > forces the kernel to use 256MB rather than 1TB segments. Forcing the > > use > > of 256MB segments makes it considerably easier to test code that > > depends > > on an SLB miss occurring. > That's a fairly awful name :) > > Can you call it "disable_1T_segments" ? > > > It should also be mentioned in Documentation/kernel-parameters.txt, with > emphasis that it's a debug option and powerpc only. To that end, should we add "powerpc_" at the start of the option? Mikey
On Tue, Jul 5, 2016 at 10:24 AM, Michael Neuling <mikey@neuling.org> wrote: > On Mon, 2016-07-04 at 16:09 +1000, Michael Ellerman wrote: >> On Mon, 2016-04-07 at 00:44:04 UTC, Oliver O'Halloran wrote: >> > >> > This patch adds the kernel command line parameter "no_tb_segs" which >> > forces the kernel to use 256MB rather than 1TB segments. Forcing the >> > use >> > of 256MB segments makes it considerably easier to test code that >> > depends >> > on an SLB miss occurring. >> That's a fairly awful name :) >> >> Can you call it "disable_1T_segments" ? >> >> >> It should also be mentioned in Documentation/kernel-parameters.txt, with >> emphasis that it's a debug option and powerpc only. > > To that end, should we add "powerpc_" at the start of the option? I don't think it's necessary. There are annotations in kernel-parameters.txt that indicate what options are arch specific and it looks like none of the existing architecture specific options are prefixed. > > Mikey
diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c index 5b22ba0b58bc..6da1a9d18e15 100644 --- a/arch/powerpc/mm/hash_utils_64.c +++ b/arch/powerpc/mm/hash_utils_64.c @@ -321,6 +321,15 @@ int htab_remove_mapping(unsigned long vstart, unsigned long vend, return ret; } +static bool no_tb_segs = false; + +static int __init parse_no_tb_segs(char *p) +{ + no_tb_segs = true; + return 0; +} +early_param("no_tb_segs", parse_no_tb_segs); + static int __init htab_dt_scan_seg_sizes(unsigned long node, const char *uname, int depth, void *data) @@ -339,6 +348,12 @@ static int __init htab_dt_scan_seg_sizes(unsigned long node, for (; size >= 4; size -= 4, ++prop) { if (be32_to_cpu(prop[0]) == 40) { DBG("1T segment support detected\n"); + + if (no_tb_segs) { + DBG("Forcing 256MB segments\n"); + break; + } + cur_cpu_spec->mmu_features |= MMU_FTR_1T_SEGMENT; return 1; }
This patch adds the kernel command line parameter "no_tb_segs" which forces the kernel to use 256MB rather than 1TB segments. Forcing the use of 256MB segments makes it considerably easier to test code that depends on an SLB miss occurring. Suggested-by: Michael Neuling <mikey@neuling.org> Suggested-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Oliver O'Halloran <oohall@gmail.com> --- arch/powerpc/mm/hash_utils_64.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+)