| Message ID | 1463158112-28625-4-git-send-email-alexander.deucher@amd.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi all, On 13 May 2016 at 17:48, Alex Deucher <alexdeucher@gmail.com> wrote: > From: Sonny Jiang <sonny.jiang@amd.com> > > Signed-off-by: Sonny Jiang <sonny.jiang@amd.com> > Reviewed-by: Alex Deucher <alexander.deucher@amd.com> > Signed-off-by: Alex Deucher <alexander.deucher@amd.com> > --- > tests/amdgpu/cs_tests.c | 48 ++++++++++++++++++++++++++++++++++++++++++------ > 1 file changed, 42 insertions(+), 6 deletions(-) > > diff --git a/tests/amdgpu/cs_tests.c b/tests/amdgpu/cs_tests.c > index c6930c0..a01ee48 100644 > --- a/tests/amdgpu/cs_tests.c > +++ b/tests/amdgpu/cs_tests.c > @@ -43,6 +43,8 @@ static amdgpu_device_handle device_handle; > static uint32_t major_version; > static uint32_t minor_version; > static uint32_t family_id; > +static uint32_t chip_rev; > +static uint32_t chip_id; > > static amdgpu_context_handle context_handle; > static amdgpu_bo_handle ib_handle; > @@ -78,6 +80,9 @@ int suite_cs_tests_init(void) > return CUE_SINIT_FAILED; > > family_id = device_handle->info.family_id; > + /* VI asic POLARIS10/11 have specific external_rev_id */ > + chip_rev = device_handle->info.chip_rev; > + chip_id = device_handle->info.chip_external_rev; > > r = amdgpu_cs_ctx_create(device_handle, &context_handle); > if (r) > @@ -200,8 +205,17 @@ static void amdgpu_cs_uvd_create(void) > CU_ASSERT_EQUAL(r, 0); > > memcpy(msg, uvd_create_msg, sizeof(uvd_create_msg)); > - if (family_id >= AMDGPU_FAMILY_VI) > + if (family_id >= AMDGPU_FAMILY_VI) { > ((uint8_t*)msg)[0x10] = 7; > + /* chip polaris 10/11 */ > + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) { > + /* dpb size */ > + ((uint8_t*)msg)[0x28] = 0x00; > + ((uint8_t*)msg)[0x29] = 0x94; > + ((uint8_t*)msg)[0x2A] = 0x6B; > + ((uint8_t*)msg)[0x2B] = 0x00; I realise that many of the UVD stuff is 'top secret', although one should really try and give symbolic names for magic numbers. With them it's be easier and less error prone when/if the above value changes. > + } > + } > > r = amdgpu_bo_cpu_unmap(buf_handle); > CU_ASSERT_EQUAL(r, 0); > @@ -230,8 +244,8 @@ static void amdgpu_cs_uvd_create(void) > > static void amdgpu_cs_uvd_decode(void) > { > - const unsigned dpb_size = 15923584, dt_size = 737280; > - uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, dt_addr, it_addr; > + const unsigned dpb_size = 15923584, ctx_size = 5287680, dt_size = 737280; > + uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, ctx_addr, dt_addr, it_addr; > struct amdgpu_bo_alloc_request req = {0}; > amdgpu_bo_handle buf_handle; > amdgpu_va_handle va_handle; > @@ -269,8 +283,21 @@ static void amdgpu_cs_uvd_decode(void) > memcpy(ptr, uvd_decode_msg, sizeof(uvd_create_msg)); > if (family_id >= AMDGPU_FAMILY_VI) { > ptr[0x10] = 7; > - ptr[0x98] = 0xb0; > - ptr[0x99] = 0x1; > + ptr[0x98] = 0x00; > + ptr[0x99] = 0x02; > + /* chip polaris10/11 */ > + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) { > + /*dpb size */ > + ptr[0x24] = 0x00; > + ptr[0x25] = 0x94; > + ptr[0x26] = 0x6B; > + ptr[0x27] = 0x00; Based on the const dpb_size a few lines above... this value is incorrect. So either the comment is off, or one/both of the values ? > + /*ctx size */ > + ptr[0x2C] = 0x00; > + ptr[0x2D] = 0xAF; > + ptr[0x2E] = 0x50; > + ptr[0x2F] = 0x00; > + } While this one does match ctx_size above, one should really set a macro for these magic values and use them throughout. Also considering that there's three almost identical places where this happens perhaps it's better to have a common helper ? Regards, Emil
Am 14.05.2016 um 16:19 schrieb Emil Velikov: > Hi all, > > On 13 May 2016 at 17:48, Alex Deucher <alexdeucher@gmail.com> wrote: >> From: Sonny Jiang <sonny.jiang@amd.com> >> >> Signed-off-by: Sonny Jiang <sonny.jiang@amd.com> >> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> >> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> >> --- >> tests/amdgpu/cs_tests.c | 48 ++++++++++++++++++++++++++++++++++++++++++------ >> 1 file changed, 42 insertions(+), 6 deletions(-) >> >> diff --git a/tests/amdgpu/cs_tests.c b/tests/amdgpu/cs_tests.c >> index c6930c0..a01ee48 100644 >> --- a/tests/amdgpu/cs_tests.c >> +++ b/tests/amdgpu/cs_tests.c >> @@ -43,6 +43,8 @@ static amdgpu_device_handle device_handle; >> static uint32_t major_version; >> static uint32_t minor_version; >> static uint32_t family_id; >> +static uint32_t chip_rev; >> +static uint32_t chip_id; >> >> static amdgpu_context_handle context_handle; >> static amdgpu_bo_handle ib_handle; >> @@ -78,6 +80,9 @@ int suite_cs_tests_init(void) >> return CUE_SINIT_FAILED; >> >> family_id = device_handle->info.family_id; >> + /* VI asic POLARIS10/11 have specific external_rev_id */ >> + chip_rev = device_handle->info.chip_rev; >> + chip_id = device_handle->info.chip_external_rev; >> >> r = amdgpu_cs_ctx_create(device_handle, &context_handle); >> if (r) >> @@ -200,8 +205,17 @@ static void amdgpu_cs_uvd_create(void) >> CU_ASSERT_EQUAL(r, 0); >> >> memcpy(msg, uvd_create_msg, sizeof(uvd_create_msg)); >> - if (family_id >= AMDGPU_FAMILY_VI) >> + if (family_id >= AMDGPU_FAMILY_VI) { >> ((uint8_t*)msg)[0x10] = 7; >> + /* chip polaris 10/11 */ >> + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) { >> + /* dpb size */ >> + ((uint8_t*)msg)[0x28] = 0x00; >> + ((uint8_t*)msg)[0x29] = 0x94; >> + ((uint8_t*)msg)[0x2A] = 0x6B; >> + ((uint8_t*)msg)[0x2B] = 0x00; > I realise that many of the UVD stuff is 'top secret', although one > should really try and give symbolic names for magic numbers. With them > it's be easier and less error prone when/if the above value changes. Actually we have exposed mostly everything in the UVD headers in mesa and those binary values here are just captured example streams. Saying so I would also prefer that we don't patch the messages on the fly as necessary, but rather have a full copy for each chipset family they differ. On the other hand it's just the unit tests. Regards, Christian. > >> + } >> + } >> >> r = amdgpu_bo_cpu_unmap(buf_handle); >> CU_ASSERT_EQUAL(r, 0); >> @@ -230,8 +244,8 @@ static void amdgpu_cs_uvd_create(void) >> >> static void amdgpu_cs_uvd_decode(void) >> { >> - const unsigned dpb_size = 15923584, dt_size = 737280; >> - uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, dt_addr, it_addr; >> + const unsigned dpb_size = 15923584, ctx_size = 5287680, dt_size = 737280; >> + uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, ctx_addr, dt_addr, it_addr; >> struct amdgpu_bo_alloc_request req = {0}; >> amdgpu_bo_handle buf_handle; >> amdgpu_va_handle va_handle; >> @@ -269,8 +283,21 @@ static void amdgpu_cs_uvd_decode(void) >> memcpy(ptr, uvd_decode_msg, sizeof(uvd_create_msg)); >> if (family_id >= AMDGPU_FAMILY_VI) { >> ptr[0x10] = 7; >> - ptr[0x98] = 0xb0; >> - ptr[0x99] = 0x1; >> + ptr[0x98] = 0x00; >> + ptr[0x99] = 0x02; >> + /* chip polaris10/11 */ >> + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) { >> + /*dpb size */ >> + ptr[0x24] = 0x00; >> + ptr[0x25] = 0x94; >> + ptr[0x26] = 0x6B; >> + ptr[0x27] = 0x00; > Based on the const dpb_size a few lines above... this value is > incorrect. So either the comment is off, or one/both of the values ? > > >> + /*ctx size */ >> + ptr[0x2C] = 0x00; >> + ptr[0x2D] = 0xAF; >> + ptr[0x2E] = 0x50; >> + ptr[0x2F] = 0x00; >> + } > While this one does match ctx_size above, one should really set a > macro for these magic values and use them throughout. Also considering > that there's three almost identical places where this happens perhaps > it's better to have a common helper ? > > Regards, > Emil > _______________________________________________ > dri-devel mailing list > dri-devel@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/dri-devel
On 14 May 2016 at 19:29, Christian König <deathsimple@vodafone.de> wrote: > Am 14.05.2016 um 16:19 schrieb Emil Velikov: >> >> Hi all, >> >> On 13 May 2016 at 17:48, Alex Deucher <alexdeucher@gmail.com> wrote: >>> >>> From: Sonny Jiang <sonny.jiang@amd.com> >>> >>> Signed-off-by: Sonny Jiang <sonny.jiang@amd.com> >>> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> >>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> >>> --- >>> tests/amdgpu/cs_tests.c | 48 >>> ++++++++++++++++++++++++++++++++++++++++++------ >>> 1 file changed, 42 insertions(+), 6 deletions(-) >>> >>> diff --git a/tests/amdgpu/cs_tests.c b/tests/amdgpu/cs_tests.c >>> index c6930c0..a01ee48 100644 >>> --- a/tests/amdgpu/cs_tests.c >>> +++ b/tests/amdgpu/cs_tests.c >>> @@ -43,6 +43,8 @@ static amdgpu_device_handle device_handle; >>> static uint32_t major_version; >>> static uint32_t minor_version; >>> static uint32_t family_id; >>> +static uint32_t chip_rev; >>> +static uint32_t chip_id; >>> >>> static amdgpu_context_handle context_handle; >>> static amdgpu_bo_handle ib_handle; >>> @@ -78,6 +80,9 @@ int suite_cs_tests_init(void) >>> return CUE_SINIT_FAILED; >>> >>> family_id = device_handle->info.family_id; >>> + /* VI asic POLARIS10/11 have specific external_rev_id */ >>> + chip_rev = device_handle->info.chip_rev; >>> + chip_id = device_handle->info.chip_external_rev; >>> >>> r = amdgpu_cs_ctx_create(device_handle, &context_handle); >>> if (r) >>> @@ -200,8 +205,17 @@ static void amdgpu_cs_uvd_create(void) >>> CU_ASSERT_EQUAL(r, 0); >>> >>> memcpy(msg, uvd_create_msg, sizeof(uvd_create_msg)); >>> - if (family_id >= AMDGPU_FAMILY_VI) >>> + if (family_id >= AMDGPU_FAMILY_VI) { >>> ((uint8_t*)msg)[0x10] = 7; >>> + /* chip polaris 10/11 */ >>> + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) >>> { >>> + /* dpb size */ >>> + ((uint8_t*)msg)[0x28] = 0x00; >>> + ((uint8_t*)msg)[0x29] = 0x94; >>> + ((uint8_t*)msg)[0x2A] = 0x6B; >>> + ((uint8_t*)msg)[0x2B] = 0x00; >> >> I realise that many of the UVD stuff is 'top secret', although one >> should really try and give symbolic names for magic numbers. With them >> it's be easier and less error prone when/if the above value changes. > > > Actually we have exposed mostly everything in the UVD headers in mesa and > those binary values here are just captured example streams. > > Saying so I would also prefer that we don't patch the messages on the fly as > necessary, but rather have a full copy for each chipset family they differ. > On the other hand it's just the unit tests. > Just saw copy copy/paste job so I'd suggested unification. At the end it's up-to you guys to decide for/against and (if so) do it. Do you have any information on the dpb_size comment below ? > Regards, > Christian. > >> >>> + } >>> + } >>> >>> r = amdgpu_bo_cpu_unmap(buf_handle); >>> CU_ASSERT_EQUAL(r, 0); >>> @@ -230,8 +244,8 @@ static void amdgpu_cs_uvd_create(void) >>> >>> static void amdgpu_cs_uvd_decode(void) >>> { >>> - const unsigned dpb_size = 15923584, dt_size = 737280; >>> - uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, dt_addr, it_addr; >>> + const unsigned dpb_size = 15923584, ctx_size = 5287680, dt_size = >>> 737280; >>> + uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, ctx_addr, dt_addr, >>> it_addr; >>> struct amdgpu_bo_alloc_request req = {0}; >>> amdgpu_bo_handle buf_handle; >>> amdgpu_va_handle va_handle; >>> @@ -269,8 +283,21 @@ static void amdgpu_cs_uvd_decode(void) >>> memcpy(ptr, uvd_decode_msg, sizeof(uvd_create_msg)); >>> if (family_id >= AMDGPU_FAMILY_VI) { >>> ptr[0x10] = 7; >>> - ptr[0x98] = 0xb0; >>> - ptr[0x99] = 0x1; >>> + ptr[0x98] = 0x00; >>> + ptr[0x99] = 0x02; >>> + /* chip polaris10/11 */ >>> + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) >>> { >>> + /*dpb size */ >>> + ptr[0x24] = 0x00; >>> + ptr[0x25] = 0x94; >>> + ptr[0x26] = 0x6B; >>> + ptr[0x27] = 0x00; >> >> Based on the const dpb_size a few lines above... this value is >> incorrect. So either the comment is off, or one/both of the values ? >> Namely this one ? Thanks Emil
Hi Emil, These values are just for this clip using on this test. We have given method in both kernel and mesa codes on how to calculate dpb size and context size. I calculated them from the sample clip and used only for this test. Regards, Sonny
diff --git a/tests/amdgpu/cs_tests.c b/tests/amdgpu/cs_tests.c index c6930c0..a01ee48 100644 --- a/tests/amdgpu/cs_tests.c +++ b/tests/amdgpu/cs_tests.c @@ -43,6 +43,8 @@ static amdgpu_device_handle device_handle; static uint32_t major_version; static uint32_t minor_version; static uint32_t family_id; +static uint32_t chip_rev; +static uint32_t chip_id; static amdgpu_context_handle context_handle; static amdgpu_bo_handle ib_handle; @@ -78,6 +80,9 @@ int suite_cs_tests_init(void) return CUE_SINIT_FAILED; family_id = device_handle->info.family_id; + /* VI asic POLARIS10/11 have specific external_rev_id */ + chip_rev = device_handle->info.chip_rev; + chip_id = device_handle->info.chip_external_rev; r = amdgpu_cs_ctx_create(device_handle, &context_handle); if (r) @@ -200,8 +205,17 @@ static void amdgpu_cs_uvd_create(void) CU_ASSERT_EQUAL(r, 0); memcpy(msg, uvd_create_msg, sizeof(uvd_create_msg)); - if (family_id >= AMDGPU_FAMILY_VI) + if (family_id >= AMDGPU_FAMILY_VI) { ((uint8_t*)msg)[0x10] = 7; + /* chip polaris 10/11 */ + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) { + /* dpb size */ + ((uint8_t*)msg)[0x28] = 0x00; + ((uint8_t*)msg)[0x29] = 0x94; + ((uint8_t*)msg)[0x2A] = 0x6B; + ((uint8_t*)msg)[0x2B] = 0x00; + } + } r = amdgpu_bo_cpu_unmap(buf_handle); CU_ASSERT_EQUAL(r, 0); @@ -230,8 +244,8 @@ static void amdgpu_cs_uvd_create(void) static void amdgpu_cs_uvd_decode(void) { - const unsigned dpb_size = 15923584, dt_size = 737280; - uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, dt_addr, it_addr; + const unsigned dpb_size = 15923584, ctx_size = 5287680, dt_size = 737280; + uint64_t msg_addr, fb_addr, bs_addr, dpb_addr, ctx_addr, dt_addr, it_addr; struct amdgpu_bo_alloc_request req = {0}; amdgpu_bo_handle buf_handle; amdgpu_va_handle va_handle; @@ -269,8 +283,21 @@ static void amdgpu_cs_uvd_decode(void) memcpy(ptr, uvd_decode_msg, sizeof(uvd_create_msg)); if (family_id >= AMDGPU_FAMILY_VI) { ptr[0x10] = 7; - ptr[0x98] = 0xb0; - ptr[0x99] = 0x1; + ptr[0x98] = 0x00; + ptr[0x99] = 0x02; + /* chip polaris10/11 */ + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) { + /*dpb size */ + ptr[0x24] = 0x00; + ptr[0x25] = 0x94; + ptr[0x26] = 0x6B; + ptr[0x27] = 0x00; + /*ctx size */ + ptr[0x2C] = 0x00; + ptr[0x2D] = 0xAF; + ptr[0x2E] = 0x50; + ptr[0x2F] = 0x00; + } } ptr += 4*1024; @@ -301,6 +328,12 @@ static void amdgpu_cs_uvd_decode(void) } else bs_addr = fb_addr + 4*1024; dpb_addr = ALIGN(bs_addr + sizeof(uvd_bitstream), 4*1024); + + if ((family_id >= AMDGPU_FAMILY_VI) && + (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A)) { + ctx_addr = ALIGN(dpb_addr + 0x006B9400, 4*1024); + } + dt_addr = ALIGN(dpb_addr + dpb_size, 4*1024); i = 0; @@ -309,8 +342,11 @@ static void amdgpu_cs_uvd_decode(void) uvd_cmd(dt_addr, 0x2, &i); uvd_cmd(fb_addr, 0x3, &i); uvd_cmd(bs_addr, 0x100, &i); - if (family_id >= AMDGPU_FAMILY_VI) + if (family_id >= AMDGPU_FAMILY_VI) { uvd_cmd(it_addr, 0x204, &i); + if (chip_id == chip_rev+0x50 || chip_id == chip_rev+0x5A) + uvd_cmd(ctx_addr, 0x206, &i); +} ib_cpu[i++] = 0x3BC6; ib_cpu[i++] = 0x1; for (; i % 16; ++i)