| Message ID | 20240227144844.23606-1-calebs@linux.vnet.ibm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | None | expand |
Hello Caleb, On 2/27/24 15:48, Caleb Schlossin wrote: > Pervasive topology(PIR) calculation for core, thread ID was > wrong for big cores (SMT8). Fixing for P10. > > Based on: <20240123195005.8965-1-calebs@linux.vnet.ibm.com> > Signed-off-by: Caleb Schlossin <calebs@linux.vnet.ibm.com> Since the initial patch [1] is not merged yet, you can simply send a v2 with the update. There is still some time before soft freeze [2]. The Subject of this patch [PATCH 2/2] seems to refer to a series. Is there a patch 1/2 ? Thanks, C. [1] https://lore.kernel.org/all/20240123195005.8965-1-calebs@linux.vnet.ibm.com/ [2] https://wiki.qemu.org/Planning/9.0 > --- > hw/ppc/pnv.c | 15 +++++++++++++-- > 1 file changed, 13 insertions(+), 2 deletions(-) > > diff --git a/hw/ppc/pnv.c b/hw/ppc/pnv.c > index 2f53883916..aa5aba60b4 100644 > --- a/hw/ppc/pnv.c > +++ b/hw/ppc/pnv.c > @@ -1068,12 +1068,23 @@ static uint32_t pnv_chip_pir_p9(PnvChip *chip, uint32_t core_id, > } > } > > +/* > + * 0:48 Reserved - Read as zeroes > + * 49:52 Node ID > + * 53:55 Chip ID > + * 56 Reserved - Read as zero > + * 57:59 Quad ID > + * 60 Core Chiplet Pair ID > + * 61:63 Thread/Core Chiplet ID t0-t2 > + * > + * We only care about the lower bits. uint32_t is fine for the moment. > + */ > static uint32_t pnv_chip_pir_p10(PnvChip *chip, uint32_t core_id, > uint32_t thread_id) > { > if (chip->nr_threads == 8) { > - return (chip->chip_id << 8) | ((thread_id & 1) << 2) | (core_id << 3) | > - (thread_id >> 1); > + return (chip->chip_id << 8) | ((core_id / 4) << 4) | > + ((core_id % 2) << 3) | thread_id; > } else { > return (chip->chip_id << 8) | (core_id << 2) | thread_id; > }
Cedric, I'll resend both patches, so this new patch shows up properly as [PATCH 2/2]. Thanks, Caleb On 2/27/24 10:15 AM, Cédric Le Goater wrote: > Hello Caleb, > > On 2/27/24 15:48, Caleb Schlossin wrote: >> Pervasive topology(PIR) calculation for core, thread ID was >> wrong for big cores (SMT8). Fixing for P10. >> >> Based on: <20240123195005.8965-1-calebs@linux.vnet.ibm.com> >> Signed-off-by: Caleb Schlossin <calebs@linux.vnet.ibm.com> > > Since the initial patch [1] is not merged yet, you can simply send a v2 > with the update. There is still some time before soft freeze [2]. > > The Subject of this patch [PATCH 2/2] seems to refer to a series. Is > there a patch 1/2 ? > > Thanks, > > C. > > > > [1] https://lore.kernel.org/all/20240123195005.8965-1-calebs@linux.vnet.ibm.com/ > [2] https://wiki.qemu.org/Planning/9.0 > >> --- >> hw/ppc/pnv.c | 15 +++++++++++++-- >> 1 file changed, 13 insertions(+), 2 deletions(-) >> >> diff --git a/hw/ppc/pnv.c b/hw/ppc/pnv.c >> index 2f53883916..aa5aba60b4 100644 >> --- a/hw/ppc/pnv.c >> +++ b/hw/ppc/pnv.c >> @@ -1068,12 +1068,23 @@ static uint32_t pnv_chip_pir_p9(PnvChip *chip, uint32_t core_id, >> } >> } >> +/* >> + * 0:48 Reserved - Read as zeroes >> + * 49:52 Node ID >> + * 53:55 Chip ID >> + * 56 Reserved - Read as zero >> + * 57:59 Quad ID >> + * 60 Core Chiplet Pair ID >> + * 61:63 Thread/Core Chiplet ID t0-t2 >> + * >> + * We only care about the lower bits. uint32_t is fine for the moment. >> + */ >> static uint32_t pnv_chip_pir_p10(PnvChip *chip, uint32_t core_id, >> uint32_t thread_id) >> { >> if (chip->nr_threads == 8) { >> - return (chip->chip_id << 8) | ((thread_id & 1) << 2) | (core_id << 3) | >> - (thread_id >> 1); >> + return (chip->chip_id << 8) | ((core_id / 4) << 4) | >> + ((core_id % 2) << 3) | thread_id; >> } else { >> return (chip->chip_id << 8) | (core_id << 2) | thread_id; >> } >
diff --git a/hw/ppc/pnv.c b/hw/ppc/pnv.c index 2f53883916..aa5aba60b4 100644 --- a/hw/ppc/pnv.c +++ b/hw/ppc/pnv.c @@ -1068,12 +1068,23 @@ static uint32_t pnv_chip_pir_p9(PnvChip *chip, uint32_t core_id, } } +/* + * 0:48 Reserved - Read as zeroes + * 49:52 Node ID + * 53:55 Chip ID + * 56 Reserved - Read as zero + * 57:59 Quad ID + * 60 Core Chiplet Pair ID + * 61:63 Thread/Core Chiplet ID t0-t2 + * + * We only care about the lower bits. uint32_t is fine for the moment. + */ static uint32_t pnv_chip_pir_p10(PnvChip *chip, uint32_t core_id, uint32_t thread_id) { if (chip->nr_threads == 8) { - return (chip->chip_id << 8) | ((thread_id & 1) << 2) | (core_id << 3) | - (thread_id >> 1); + return (chip->chip_id << 8) | ((core_id / 4) << 4) | + ((core_id % 2) << 3) | thread_id; } else { return (chip->chip_id << 8) | (core_id << 2) | thread_id; }
Pervasive topology(PIR) calculation for core, thread ID was wrong for big cores (SMT8). Fixing for P10. Based on: <20240123195005.8965-1-calebs@linux.vnet.ibm.com> Signed-off-by: Caleb Schlossin <calebs@linux.vnet.ibm.com> --- hw/ppc/pnv.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-)