| Message ID | 1414403901-32648-1-git-send-email-alex.hung@canonical.com |
|---|---|
| State | Accepted |
| Headers | show |
On 27/10/14 09:58, Alex Hung wrote: > Kernel patches also fix it such as 21126b296e6f5f172e72da7cebb42b87fbd19b3a > and 6d1490e248190646773083df60b25ffe14f74a4b as there is no 32 bit in ACPI. I'm not entirely sure what problem this fix actually addresses. Can you explain a bit further? thanks, Colin > > Signed-off-by: Alex Hung <alex.hung@canonical.com> > --- > data/klog.json | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/data/klog.json b/data/klog.json > index 003136a..d41d4e0 100644 > --- a/data/klog.json > +++ b/data/klog.json > @@ -1844,9 +1844,9 @@ > }, > { > "compare_mode": "regex", > - "log_level": "LOG_LEVEL_HIGH", > + "log_level": "LOG_LEVEL_MEDIUM", > "pattern": "32-bit FADT register is too long .* to convert to GAS struct - 255 bits max, truncating", > - "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated. This is firmware bug.", > + "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated.", > "label": "KlogAcpiFadtGasBitWidthTooLong" > }, > { >
On 11/13/2014 08:00 PM, Colin Ian King wrote: > On 27/10/14 09:58, Alex Hung wrote: >> Kernel patches also fix it such as 21126b296e6f5f172e72da7cebb42b87fbd19b3a >> and 6d1490e248190646773083df60b25ffe14f74a4b as there is no 32 bit in ACPI. > > I'm not entirely sure what problem this fix actually addresses. Can you > explain a bit further? > > thanks, Colin > It is common to see kernel complains about the size of GPE0 in recent systems as the example below. It is Intel's BIOS reference code that defines such a size. In ACPI spec, GPE0_BLK_LEN is defined as "Number of bytes decoded by GPE0_BLK ... is a non-negative multiple of 2." and is not limited to 255 bits. New kernel (ACPICA) also fixes this error message. Because it is not a bug in firmware, it is a good idea to remove it from HIGH so users (and cert) will not mistakenly think this is a firmware bug that needs fixing right away. =========================================================== FAILED [HIGH] KlogAcpiFadtGasBitWidthTooLong: Test 1, HIGH Kernel message: [0.000000] ACPI Error: Gpe0Block - 32-bit FADT register is too long (32 bytes, 256 bits) to convert to GAS struct - 255 bits max, truncating (20131115/tbfadt-202) The disassemabled FACP by fwts is shown here: FACP @ 99ffb000 (268 bytes) ---- Fixed ACPI Description Table: [0x000 0000 4] Signature: FACP [0x004 0004 4] Length: 0x0000010c [0x008 0008 1] Revision: 0x05 [0x009 0009 1] Checksum: 0x92 [0x00a 0010 6] OEM ID: HPQOEM [0x010 0016 8] OEM Table ID: 2248 [0x018 0024 4] OEM Revision: 0x00000000 [0x01c 0028 4] Creator ID: HP [0x020 0032 4] Creator Revision: 0x00000001 [0x024 0036 4] FACS Address: 0x99f76000 [0x028 0040 4] DSDT Address: 0x99fd0000 [0x02c 0044 1] Model: 0x00 [0x02d 0045 1] PM Profile: 0x02 (Mobile) [0x02e 0046 2] SCI Interrupt: 0x0009 [0x030 0048 4] SMI Command Port: 0x000000b2 [0x034 0052 1] ACPI Enable Value: 0xf0 [0x035 0053 1] ACPI Disable Value: 0xf1 [0x036 0054 1] S4BIOS Command: 0xf2 [0x037 0055 1] P-State Control: 0x00 [0x038 0056 4] PM1A Event Block: 0x00001800 [0x03c 0060 4] PM1B Event Block: 0x00000000 [0x040 0064 4] PM1A Control Block: 0x00001804 [0x044 0068 4] PM1B Control Block: 0x00000000 [0x048 0072 4] PM2 Control Block: 0x00001850 [0x04c 0076 4] PM Timer Block: 0x00001808 [0x050 0080 4] GPE0 Block: 0x00001880 [0x054 0084 4] GPE1 Block: 0x00000000 [0x058 0088 1] PM1 Event Block Length: 0x04 [0x059 0089 1] PM1 Control Block Length: 0x02 [0x05a 0090 1] PM2 Control Block Length: 0x01 [0x05b 0091 1] PM Timer Block Length: 0x04 [0x05c 0092 1] GPE0 Block Length: 0x20 [0x05d 0093 1] GPE1 Block Length: 0x00 [0x05f 0095 1] _CST Support: 0x00 [0x060 0096 2] C2 Latency: 0x0065 [0x062 0098 2] C3 Latency: 0x0039 [0x064 0100 2] CPU Cache Size: 0x0000 [0x066 0102 2] CPU Cache Flush Stride: 0x0000 [0x068 0104 1] Duty Cycle Offset: 0x01 [0x069 0105 1] Duty Cycle Width: 0x03 [0x06a 0106 1] RTC Day Alarm Index: 0x0d [0x06b 0107 1] RTC Month Alarm Index: 0x00 [0x06c 0108 1] RTC Century Index: 0x32 [0x06d 0109 2] IA-PC Boot Flags (see below): 0x0013 .... (skipped) =========================================================== >> >> Signed-off-by: Alex Hung <alex.hung@canonical.com> >> --- >> data/klog.json | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/data/klog.json b/data/klog.json >> index 003136a..d41d4e0 100644 >> --- a/data/klog.json >> +++ b/data/klog.json >> @@ -1844,9 +1844,9 @@ >> }, >> { >> "compare_mode": "regex", >> - "log_level": "LOG_LEVEL_HIGH", >> + "log_level": "LOG_LEVEL_MEDIUM", >> "pattern": "32-bit FADT register is too long .* to convert to GAS struct - 255 bits max, truncating", >> - "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated. This is firmware bug.", >> + "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated.", >> "label": "KlogAcpiFadtGasBitWidthTooLong" >> }, >> { >> > >
On Mon, Oct 27, 2014 at 5:58 PM, Alex Hung <alex.hung@canonical.com> wrote: > Kernel patches also fix it such as 21126b296e6f5f172e72da7cebb42b87fbd19b3a > and 6d1490e248190646773083df60b25ffe14f74a4b as there is no 32 bit in ACPI. > > Signed-off-by: Alex Hung <alex.hung@canonical.com> > --- > data/klog.json | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/data/klog.json b/data/klog.json > index 003136a..d41d4e0 100644 > --- a/data/klog.json > +++ b/data/klog.json > @@ -1844,9 +1844,9 @@ > }, > { > "compare_mode": "regex", > - "log_level": "LOG_LEVEL_HIGH", > + "log_level": "LOG_LEVEL_MEDIUM", > "pattern": "32-bit FADT register is too long .* to convert to GAS struct - 255 bits max, truncating", > - "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated. This is firmware bug.", > + "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated.", > "label": "KlogAcpiFadtGasBitWidthTooLong" > }, > { > -- > 1.9.1 > Acked-by: Keng-Yu Lin <kengyu@canonical.com>
On 2014年12月03日 14:37, Keng-Yu Lin wrote: > On Mon, Oct 27, 2014 at 5:58 PM, Alex Hung <alex.hung@canonical.com> wrote: >> Kernel patches also fix it such as 21126b296e6f5f172e72da7cebb42b87fbd19b3a >> and 6d1490e248190646773083df60b25ffe14f74a4b as there is no 32 bit in ACPI. >> >> Signed-off-by: Alex Hung <alex.hung@canonical.com> >> --- >> data/klog.json | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/data/klog.json b/data/klog.json >> index 003136a..d41d4e0 100644 >> --- a/data/klog.json >> +++ b/data/klog.json >> @@ -1844,9 +1844,9 @@ >> }, >> { >> "compare_mode": "regex", >> - "log_level": "LOG_LEVEL_HIGH", >> + "log_level": "LOG_LEVEL_MEDIUM", >> "pattern": "32-bit FADT register is too long .* to convert to GAS struct - 255 bits max, truncating", >> - "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated. This is firmware bug.", >> + "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated.", >> "label": "KlogAcpiFadtGasBitWidthTooLong" >> }, >> { >> -- >> 1.9.1 >> > Acked-by: Ivan Hu <ivan.hu@canonical.com>
On 27/10/14 09:58, Alex Hung wrote: > Kernel patches also fix it such as 21126b296e6f5f172e72da7cebb42b87fbd19b3a > and 6d1490e248190646773083df60b25ffe14f74a4b as there is no 32 bit in ACPI. > > Signed-off-by: Alex Hung <alex.hung@canonical.com> > --- > data/klog.json | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/data/klog.json b/data/klog.json > index 003136a..d41d4e0 100644 > --- a/data/klog.json > +++ b/data/klog.json > @@ -1844,9 +1844,9 @@ > }, > { > "compare_mode": "regex", > - "log_level": "LOG_LEVEL_HIGH", > + "log_level": "LOG_LEVEL_MEDIUM", > "pattern": "32-bit FADT register is too long .* to convert to GAS struct - 255 bits max, truncating", > - "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated. This is firmware bug.", > + "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated.", > "label": "KlogAcpiFadtGasBitWidthTooLong" > }, > { > Apologies, I thought I replied with an ACK, apparently, I didn't. Sorry for the delay. Acked-by: Colin Ian King <colin.king@canonical.com>
diff --git a/data/klog.json b/data/klog.json index 003136a..d41d4e0 100644 --- a/data/klog.json +++ b/data/klog.json @@ -1844,9 +1844,9 @@ }, { "compare_mode": "regex", - "log_level": "LOG_LEVEL_HIGH", + "log_level": "LOG_LEVEL_MEDIUM", "pattern": "32-bit FADT register is too long .* to convert to GAS struct - 255 bits max, truncating", - "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated. This is firmware bug.", + "advice": "The Generic Address Structure bit_width was greater than 31 which is too large for a 32 bit ACPI FADT register and was truncated.", "label": "KlogAcpiFadtGasBitWidthTooLong" }, {
Kernel patches also fix it such as 21126b296e6f5f172e72da7cebb42b87fbd19b3a and 6d1490e248190646773083df60b25ffe14f74a4b as there is no 32 bit in ACPI. Signed-off-by: Alex Hung <alex.hung@canonical.com> --- data/klog.json | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)