| Message ID | 20240903045247.275642-1-changqing.li@windriver.com |
|---|---|
| State | New |
| Headers | show |
| Series | [scarthgap] mariadb: fix runtime failure on riscv | expand |
do we need this on master as well ? if so please rebase on master-next or master On Mon, Sep 2, 2024 at 9:52 PM Changqing Li via lists.openembedded.org <changqing.li=windriver.com@lists.openembedded.org> wrote: > > From: Changqing Li <changqing.li@windriver.com> > > Starting with Linux 6.6, RDCYCLE is a privileged instruction on RISC-V > and can't be used directly from userland. This causes 'systemctl start > mysqld.service' failed with error: > [ 1456.918172] mariadbd[12115]: unhandled signal 4 code 0x1 at 0x000055558689d134 in mariadbd[555585bfa000+14a7000] > [ 1456.921772] CPU: 1 PID: 12115 Comm: mariadbd Not tainted 6.6.43-yocto-standard #1 > [ 1456.922327] Hardware name: riscv-virtio,qemu (DT) > [ 1456.923045] epc : 000055558689d134 ra : 000055558620ea48 sp : 00007fffdc487770 > [ 1456.923525] gp : 00005555872ec400 tp : 00007fff89560780 t0 : 0000555587be32e8 > [ 1456.923951] t1 : 0000555586886042 t2 : 000000002d6a89f0 s0 : 00007fffdc4877b0 > > Signed-off-by: Changqing Li <changqing.li@windriver.com> > --- > meta-oe/recipes-dbs/mysql/mariadb.inc | 1 + > ...RISC-V-use-RDTIME-instead-of-RDCYCLE.patch | 66 +++++++++++++++++++ > 2 files changed, 67 insertions(+) > create mode 100644 meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch > > diff --git a/meta-oe/recipes-dbs/mysql/mariadb.inc b/meta-oe/recipes-dbs/mysql/mariadb.inc > index 33da32fb2..124a49ff9 100644 > --- a/meta-oe/recipes-dbs/mysql/mariadb.inc > +++ b/meta-oe/recipes-dbs/mysql/mariadb.inc > @@ -25,6 +25,7 @@ SRC_URI = "https://archive.mariadb.org/${BP}/source/${BP}.tar.gz \ > file://0001-Add-missing-includes-cstdint-and-cstdio.patch \ > file://0001-Remove-the-compile_time_assert-lines.patch \ > file://0001-MDEV-33439-Fix-build-with-libxml2-2.12.patch \ > + file://0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch \ > " > SRC_URI:append:libc-musl = " file://ppc-remove-glibc-dep.patch" > SRC_URI[sha256sum] = "5239a245ed90517e96396605cd01ccd8f73cd7442d1b3076b6ffe258110e5157" > diff --git a/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch > new file mode 100644 > index 000000000..25a2c399c > --- /dev/null > +++ b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch > @@ -0,0 +1,66 @@ > +From 342f0dd9b4f9fc49dcb589cd98933ea330de55d8 Mon Sep 17 00:00:00 2001 > +From: Aurelien Jarno <aurelien@aurel32.net> > +Date: Thu, 4 Jan 2024 11:30:34 +0100 > +Subject: [PATCH] RISC-V: use RDTIME instead of RDCYCLE > + > +Starting with Linux 6.6 [1], RDCYCLE is a privileged instruction on > +RISC-V and can't be used directly from userland. There is a sysctl > +option to change that as a transition period, but it will eventually > +disappear. > + > +Use RDTIME instead, which while less accurate has the advantage of being > +synchronized between CPU (and thus monotonic) and of constant frequency. > + > +[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=cc4c07c89aada16229084eeb93895c95b7eabaa3 > + > +Upstream-Status: Backport [https://github.com/MariaDB/server/commit/656f8867720efc1b4dd0969319f35a3e1a2a005e] > +Signed-off-by: Changqing Li <changqing.li@windriver.com> > +--- > + include/my_rdtsc.h | 12 ++++++------ > + 1 file changed, 6 insertions(+), 6 deletions(-) > + > +diff --git a/include/my_rdtsc.h b/include/my_rdtsc.h > +index 8b9b0046bc0..21e44847d9a 100644 > +--- a/include/my_rdtsc.h > ++++ b/include/my_rdtsc.h > +@@ -111,7 +111,7 @@ C_MODE_START > + On AARCH64, we use the generic timer base register. We override clang > + implementation for aarch64 as it access a PMU register which is not > + guaranteed to be active. > +- On RISC-V, we use the rdcycle instruction to read from mcycle register. > ++ On RISC-V, we use the rdtime instruction to read from mtime register. > + > + Sadly, we have nothing for the Digital Alpha, MIPS, Motorola m68k, > + HP PA-RISC or other non-mainstream (or obsolete) processors. > +@@ -211,15 +211,15 @@ static inline ulonglong my_timer_cycles(void) > + } > + #elif defined(__riscv) > + #define MY_TIMER_ROUTINE_CYCLES MY_TIMER_ROUTINE_RISCV > +- /* Use RDCYCLE (and RDCYCLEH on riscv32) */ > ++ /* Use RDTIME (and RDTIMEH on riscv32) */ > + { > + # if __riscv_xlen == 32 > + ulong result_lo, result_hi0, result_hi1; > + /* Implemented in assembly because Clang insisted on branching. */ > + __asm __volatile__( > +- "rdcycleh %0\n" > +- "rdcycle %1\n" > +- "rdcycleh %2\n" > ++ "rdtimeh %0\n" > ++ "rdtime %1\n" > ++ "rdtimeh %2\n" > + "sub %0, %0, %2\n" > + "seqz %0, %0\n" > + "sub %0, zero, %0\n" > +@@ -228,7 +228,7 @@ static inline ulonglong my_timer_cycles(void) > + return (static_cast<ulonglong>(result_hi1) << 32) | result_lo; > + # else > + ulonglong result; > +- __asm __volatile__("rdcycle %0" : "=r"(result)); > ++ __asm __volatile__("rdtime %0" : "=r"(result)); > + return result; > + } > + # endif > +-- > +2.25.1 > + > -- > 2.46.0 > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#112078): https://lists.openembedded.org/g/openembedded-devel/message/112078 > Mute This Topic: https://lists.openembedded.org/mt/108240112/1997914 > Group Owner: openembedded-devel+owner@lists.openembedded.org > Unsubscribe: https://lists.openembedded.org/g/openembedded-devel/unsub [raj.khem@gmail.com] > -=-=-=-=-=-=-=-=-=-=-=- >
On 9/3/24 22:03, Khem Raj wrote: > CAUTION: This email comes from a non Wind River email account! > Do not click links or open attachments unless you recognize the sender and know the content is safe. > > do we need this on master as well ? if so please rebase on master-next > or master No, master is mariadb 10.11.9, this issue is fixed since 10.11.8. //Changqing > On Mon, Sep 2, 2024 at 9:52 PM Changqing Li via lists.openembedded.org > <changqing.li=windriver.com@lists.openembedded.org> wrote: >> From: Changqing Li<changqing.li@windriver.com> >> >> Starting with Linux 6.6, RDCYCLE is a privileged instruction on RISC-V >> and can't be used directly from userland. This causes 'systemctl start >> mysqld.service' failed with error: >> [ 1456.918172] mariadbd[12115]: unhandled signal 4 code 0x1 at 0x000055558689d134 in mariadbd[555585bfa000+14a7000] >> [ 1456.921772] CPU: 1 PID: 12115 Comm: mariadbd Not tainted 6.6.43-yocto-standard #1 >> [ 1456.922327] Hardware name: riscv-virtio,qemu (DT) >> [ 1456.923045] epc : 000055558689d134 ra : 000055558620ea48 sp : 00007fffdc487770 >> [ 1456.923525] gp : 00005555872ec400 tp : 00007fff89560780 t0 : 0000555587be32e8 >> [ 1456.923951] t1 : 0000555586886042 t2 : 000000002d6a89f0 s0 : 00007fffdc4877b0 >> >> Signed-off-by: Changqing Li<changqing.li@windriver.com> >> --- >> meta-oe/recipes-dbs/mysql/mariadb.inc | 1 + >> ...RISC-V-use-RDTIME-instead-of-RDCYCLE.patch | 66 +++++++++++++++++++ >> 2 files changed, 67 insertions(+) >> create mode 100644 meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch >> >> diff --git a/meta-oe/recipes-dbs/mysql/mariadb.inc b/meta-oe/recipes-dbs/mysql/mariadb.inc >> index 33da32fb2..124a49ff9 100644 >> --- a/meta-oe/recipes-dbs/mysql/mariadb.inc >> +++ b/meta-oe/recipes-dbs/mysql/mariadb.inc >> @@ -25,6 +25,7 @@ SRC_URI ="https://archive.mariadb.org/${BP}/source/${BP}.tar.gz \ >> file://0001-Add-missing-includes-cstdint-and-cstdio.patch \ >> file://0001-Remove-the-compile_time_assert-lines.patch \ >> file://0001-MDEV-33439-Fix-build-with-libxml2-2.12.patch \ + >> file://0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch \ " >> SRC_URI:append:libc-musl =" file://ppc-remove-glibc-dep.patch" >> SRC_URI[sha256sum] = "5239a245ed90517e96396605cd01ccd8f73cd7442d1b3076b6ffe258110e5157" >> diff --git a/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch >> new file mode 100644 >> index 000000000..25a2c399c >> --- /dev/null >> +++ b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch >> @@ -0,0 +1,66 @@ >> +From 342f0dd9b4f9fc49dcb589cd98933ea330de55d8 Mon Sep 17 00:00:00 2001 >> +From: Aurelien Jarno<aurelien@aurel32.net> >> +Date: Thu, 4 Jan 2024 11:30:34 +0100 >> +Subject: [PATCH] RISC-V: use RDTIME instead of RDCYCLE >> + >> +Starting with Linux 6.6 [1], RDCYCLE is a privileged instruction on >> +RISC-V and can't be used directly from userland. There is a sysctl >> +option to change that as a transition period, but it will eventually >> +disappear. >> + >> +Use RDTIME instead, which while less accurate has the advantage of being >> +synchronized between CPU (and thus monotonic) and of constant frequency. >> + >> +[1]https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=cc4c07c89aada16229084eeb93895c95b7eabaa3 >> + >> +Upstream-Status: Backport [https://github.com/MariaDB/server/commit/656f8867720efc1b4dd0969319f35a3e1a2a005e] >> +Signed-off-by: Changqing Li<changqing.li@windriver.com> >> +--- >> + include/my_rdtsc.h | 12 ++++++------ >> + 1 file changed, 6 insertions(+), 6 deletions(-) >> + >> +diff --git a/include/my_rdtsc.h b/include/my_rdtsc.h >> +index 8b9b0046bc0..21e44847d9a 100644 >> +--- a/include/my_rdtsc.h >> ++++ b/include/my_rdtsc.h >> +@@ -111,7 +111,7 @@ C_MODE_START >> + On AARCH64, we use the generic timer base register. We override clang >> + implementation for aarch64 as it access a PMU register which is not >> + guaranteed to be active. >> +- On RISC-V, we use the rdcycle instruction to read from mcycle register. >> ++ On RISC-V, we use the rdtime instruction to read from mtime register. >> + >> + Sadly, we have nothing for the Digital Alpha, MIPS, Motorola m68k, >> + HP PA-RISC or other non-mainstream (or obsolete) processors. >> +@@ -211,15 +211,15 @@ static inline ulonglong my_timer_cycles(void) >> + } >> + #elif defined(__riscv) >> + #define MY_TIMER_ROUTINE_CYCLES MY_TIMER_ROUTINE_RISCV >> +- /* Use RDCYCLE (and RDCYCLEH on riscv32) */ >> ++ /* Use RDTIME (and RDTIMEH on riscv32) */ >> + { >> + # if __riscv_xlen == 32 >> + ulong result_lo, result_hi0, result_hi1; >> + /* Implemented in assembly because Clang insisted on branching. */ >> + __asm __volatile__( >> +- "rdcycleh %0\n" >> +- "rdcycle %1\n" >> +- "rdcycleh %2\n" >> ++ "rdtimeh %0\n" >> ++ "rdtime %1\n" >> ++ "rdtimeh %2\n" >> + "sub %0, %0, %2\n" >> + "seqz %0, %0\n" >> + "sub %0, zero, %0\n" >> +@@ -228,7 +228,7 @@ static inline ulonglong my_timer_cycles(void) >> + return (static_cast<ulonglong>(result_hi1) << 32) | result_lo; >> + # else >> + ulonglong result; >> +- __asm __volatile__("rdcycle %0" : "=r"(result)); >> ++ __asm __volatile__("rdtime %0" : "=r"(result)); >> + return result; >> + } >> + # endif >> +-- >> +2.25.1 >> + >> -- >> 2.46.0 >> >> >> -=-=-=-=-=-=-=-=-=-=-=- >> Links: You receive all messages sent to this group. >> View/Reply Online (#112078):https://lists.openembedded.org/g/openembedded-devel/message/112078 >> Mute This Topic:https://lists.openembedded.org/mt/108240112/1997914 >> Group Owner:openembedded-devel+owner@lists.openembedded.org >> Unsubscribe:https://lists.openembedded.org/g/openembedded-devel/unsub [raj.khem@gmail.com] >> -=-=-=-=-=-=-=-=-=-=-=- >>
On Tue, Sep 3, 2024 at 5:15 PM Changqing Li <changqing.li@windriver.com> wrote: > > On 9/3/24 22:03, Khem Raj wrote: > > CAUTION: This email comes from a non Wind River email account! > Do not click links or open attachments unless you recognize the sender and know the content is safe. > > do we need this on master as well ? if so please rebase on master-next > or master > > No, master is mariadb 10.11.9, this issue is fixed since 10.11.8. > Yeah I realised thanks for confirmation > //Changqing > > On Mon, Sep 2, 2024 at 9:52 PM Changqing Li via lists.openembedded.org<changqing.li=windriver.com@lists.openembedded.org> <changqing.li=windriver.com@lists.openembedded.org> wrote: > > From: Changqing Li <changqing.li@windriver.com> <changqing.li@windriver.com> > > Starting with Linux 6.6, RDCYCLE is a privileged instruction on RISC-V > and can't be used directly from userland. This causes 'systemctl start > mysqld.service' failed with error: > [ 1456.918172] mariadbd[12115]: unhandled signal 4 code 0x1 at 0x000055558689d134 in mariadbd[555585bfa000+14a7000] > [ 1456.921772] CPU: 1 PID: 12115 Comm: mariadbd Not tainted 6.6.43-yocto-standard #1 > [ 1456.922327] Hardware name: riscv-virtio,qemu (DT) > [ 1456.923045] epc : 000055558689d134 ra : 000055558620ea48 sp : 00007fffdc487770 > [ 1456.923525] gp : 00005555872ec400 tp : 00007fff89560780 t0 : 0000555587be32e8 > [ 1456.923951] t1 : 0000555586886042 t2 : 000000002d6a89f0 s0 : 00007fffdc4877b0 > > Signed-off-by: Changqing Li <changqing.li@windriver.com> <changqing.li@windriver.com> > --- > meta-oe/recipes-dbs/mysql/mariadb.inc | 1 + > ...RISC-V-use-RDTIME-instead-of-RDCYCLE.patch | 66 +++++++++++++++++++ > 2 files changed, 67 insertions(+) > create mode 100644 meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch > > diff --git a/meta-oe/recipes-dbs/mysql/mariadb.inc b/meta-oe/recipes-dbs/mysql/mariadb.inc > index 33da32fb2..124a49ff9 100644 > --- a/meta-oe/recipes-dbs/mysql/mariadb.inc > +++ b/meta-oe/recipes-dbs/mysql/mariadb.inc > @@ -25,6 +25,7 @@ SRC_URI = "https://archive.mariadb.org/${BP}/source/${BP}.tar.gz \ > file://0001-Add-missing-includes-cstdint-and-cstdio.patch \ > file://0001-Remove-the-compile_time_assert-lines.patch \ > file://0001-MDEV-33439-Fix-build-with-libxml2-2.12.patch \ > + file://0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch \ > " <https://archive.mariadb.org/$%7BBP%7D/source/$%7BBP%7D.tar.gz%5Cfile://0001-Add-missing-includes-cstdint-and-cstdio.patch%5Cfile://0001-Remove-the-compile_time_assert-lines.patch%5Cfile://0001-MDEV-33439-Fix-build-with-libxml2-2.12.patch%5C+file://0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch%5C> > SRC_URI:append:libc-musl = " file://ppc-remove-glibc-dep.patch" > SRC_URI[sha256sum] = "5239a245ed90517e96396605cd01ccd8f73cd7442d1b3076b6ffe258110e5157" > diff --git a/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch > new file mode 100644 > index 000000000..25a2c399c > --- /dev/null > +++ b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch > @@ -0,0 +1,66 @@ > +From 342f0dd9b4f9fc49dcb589cd98933ea330de55d8 Mon Sep 17 00:00:00 2001 > +From: Aurelien Jarno <aurelien@aurel32.net> <aurelien@aurel32.net> > +Date: Thu, 4 Jan 2024 11:30:34 +0100 > +Subject: [PATCH] RISC-V: use RDTIME instead of RDCYCLE > + > +Starting with Linux 6.6 [1], RDCYCLE is a privileged instruction on > +RISC-V and can't be used directly from userland. There is a sysctl > +option to change that as a transition period, but it will eventually > +disappear. > + > +Use RDTIME instead, which while less accurate has the advantage of being > +synchronized between CPU (and thus monotonic) and of constant frequency. > + > +[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=cc4c07c89aada16229084eeb93895c95b7eabaa3 > + > +Upstream-Status: Backport [https://github.com/MariaDB/server/commit/656f8867720efc1b4dd0969319f35a3e1a2a005e] > +Signed-off-by: Changqing Li <changqing.li@windriver.com> <changqing.li@windriver.com> > +--- > + include/my_rdtsc.h | 12 ++++++------ > + 1 file changed, 6 insertions(+), 6 deletions(-) > + > +diff --git a/include/my_rdtsc.h b/include/my_rdtsc.h > +index 8b9b0046bc0..21e44847d9a 100644 > +--- a/include/my_rdtsc.h > ++++ b/include/my_rdtsc.h > +@@ -111,7 +111,7 @@ C_MODE_START > + On AARCH64, we use the generic timer base register. We override clang > + implementation for aarch64 as it access a PMU register which is not > + guaranteed to be active. > +- On RISC-V, we use the rdcycle instruction to read from mcycle register. > ++ On RISC-V, we use the rdtime instruction to read from mtime register. > + > + Sadly, we have nothing for the Digital Alpha, MIPS, Motorola m68k, > + HP PA-RISC or other non-mainstream (or obsolete) processors. > +@@ -211,15 +211,15 @@ static inline ulonglong my_timer_cycles(void) > + } > + #elif defined(__riscv) > + #define MY_TIMER_ROUTINE_CYCLES MY_TIMER_ROUTINE_RISCV > +- /* Use RDCYCLE (and RDCYCLEH on riscv32) */ > ++ /* Use RDTIME (and RDTIMEH on riscv32) */ > + { > + # if __riscv_xlen == 32 > + ulong result_lo, result_hi0, result_hi1; > + /* Implemented in assembly because Clang insisted on branching. */ > + __asm __volatile__( > +- "rdcycleh %0\n" > +- "rdcycle %1\n" > +- "rdcycleh %2\n" > ++ "rdtimeh %0\n" > ++ "rdtime %1\n" > ++ "rdtimeh %2\n" > + "sub %0, %0, %2\n" > + "seqz %0, %0\n" > + "sub %0, zero, %0\n" > +@@ -228,7 +228,7 @@ static inline ulonglong my_timer_cycles(void) > + return (static_cast<ulonglong>(result_hi1) << 32) | result_lo; > + # else > + ulonglong result; > +- __asm __volatile__("rdcycle %0" : "=r"(result)); > ++ __asm __volatile__("rdtime %0" : "=r"(result)); > + return result; > + } > + # endif > +-- > +2.25.1 > + > -- > 2.46.0 > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#112078): https://lists.openembedded.org/g/openembedded-devel/message/112078 > Mute This Topic: https://lists.openembedded.org/mt/108240112/1997914 > Group Owner: openembedded-devel+owner@lists.openembedded.org > Unsubscribe: https://lists.openembedded.org/g/openembedded-devel/unsub [raj.khem@gmail.com] > -=-=-=-=-=-=-=-=-=-=-=- > > >
diff --git a/meta-oe/recipes-dbs/mysql/mariadb.inc b/meta-oe/recipes-dbs/mysql/mariadb.inc index 33da32fb2..124a49ff9 100644 --- a/meta-oe/recipes-dbs/mysql/mariadb.inc +++ b/meta-oe/recipes-dbs/mysql/mariadb.inc @@ -25,6 +25,7 @@ SRC_URI = "https://archive.mariadb.org/${BP}/source/${BP}.tar.gz \ file://0001-Add-missing-includes-cstdint-and-cstdio.patch \ file://0001-Remove-the-compile_time_assert-lines.patch \ file://0001-MDEV-33439-Fix-build-with-libxml2-2.12.patch \ + file://0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch \ " SRC_URI:append:libc-musl = " file://ppc-remove-glibc-dep.patch" SRC_URI[sha256sum] = "5239a245ed90517e96396605cd01ccd8f73cd7442d1b3076b6ffe258110e5157" diff --git a/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch new file mode 100644 index 000000000..25a2c399c --- /dev/null +++ b/meta-oe/recipes-dbs/mysql/mariadb/0001-RISC-V-use-RDTIME-instead-of-RDCYCLE.patch @@ -0,0 +1,66 @@ +From 342f0dd9b4f9fc49dcb589cd98933ea330de55d8 Mon Sep 17 00:00:00 2001 +From: Aurelien Jarno <aurelien@aurel32.net> +Date: Thu, 4 Jan 2024 11:30:34 +0100 +Subject: [PATCH] RISC-V: use RDTIME instead of RDCYCLE + +Starting with Linux 6.6 [1], RDCYCLE is a privileged instruction on +RISC-V and can't be used directly from userland. There is a sysctl +option to change that as a transition period, but it will eventually +disappear. + +Use RDTIME instead, which while less accurate has the advantage of being +synchronized between CPU (and thus monotonic) and of constant frequency. + +[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=cc4c07c89aada16229084eeb93895c95b7eabaa3 + +Upstream-Status: Backport [https://github.com/MariaDB/server/commit/656f8867720efc1b4dd0969319f35a3e1a2a005e] +Signed-off-by: Changqing Li <changqing.li@windriver.com> +--- + include/my_rdtsc.h | 12 ++++++------ + 1 file changed, 6 insertions(+), 6 deletions(-) + +diff --git a/include/my_rdtsc.h b/include/my_rdtsc.h +index 8b9b0046bc0..21e44847d9a 100644 +--- a/include/my_rdtsc.h ++++ b/include/my_rdtsc.h +@@ -111,7 +111,7 @@ C_MODE_START + On AARCH64, we use the generic timer base register. We override clang + implementation for aarch64 as it access a PMU register which is not + guaranteed to be active. +- On RISC-V, we use the rdcycle instruction to read from mcycle register. ++ On RISC-V, we use the rdtime instruction to read from mtime register. + + Sadly, we have nothing for the Digital Alpha, MIPS, Motorola m68k, + HP PA-RISC or other non-mainstream (or obsolete) processors. +@@ -211,15 +211,15 @@ static inline ulonglong my_timer_cycles(void) + } + #elif defined(__riscv) + #define MY_TIMER_ROUTINE_CYCLES MY_TIMER_ROUTINE_RISCV +- /* Use RDCYCLE (and RDCYCLEH on riscv32) */ ++ /* Use RDTIME (and RDTIMEH on riscv32) */ + { + # if __riscv_xlen == 32 + ulong result_lo, result_hi0, result_hi1; + /* Implemented in assembly because Clang insisted on branching. */ + __asm __volatile__( +- "rdcycleh %0\n" +- "rdcycle %1\n" +- "rdcycleh %2\n" ++ "rdtimeh %0\n" ++ "rdtime %1\n" ++ "rdtimeh %2\n" + "sub %0, %0, %2\n" + "seqz %0, %0\n" + "sub %0, zero, %0\n" +@@ -228,7 +228,7 @@ static inline ulonglong my_timer_cycles(void) + return (static_cast<ulonglong>(result_hi1) << 32) | result_lo; + # else + ulonglong result; +- __asm __volatile__("rdcycle %0" : "=r"(result)); ++ __asm __volatile__("rdtime %0" : "=r"(result)); + return result; + } + # endif +-- +2.25.1 +