| Message ID | 726a7ce88982fba5df84e4d22b083e39b57a6bc6.1444900732.git.shengjiu.wang@freescale.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 15/10/15 10:30, Shengjiu Wang wrote: > __user_swpX_asm maybe failed in first STREX operation, emulate_swpX > will try again, but the *data has been changed in first time. which > causes the result is wrong. > This patch is to fix this issue. When STREX succeed, change the *data. > if it fail, *data is not changed. > > Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com> > --- > arch/arm/kernel/swp_emulate.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > diff --git a/arch/arm/kernel/swp_emulate.c b/arch/arm/kernel/swp_emulate.c > index 5b26e7e..c3fe769d 100644 > --- a/arch/arm/kernel/swp_emulate.c > +++ b/arch/arm/kernel/swp_emulate.c > @@ -36,10 +36,10 @@ > */ > #define __user_swpX_asm(data, addr, res, temp, B) \ > __asm__ __volatile__( \ > - " mov %2, %1\n" \ > - "0: ldrex"B" %1, [%3]\n" \ > - "1: strex"B" %0, %2, [%3]\n" \ > + "0: ldrex"B" %2, [%3]\n" \ > + "1: strex"B" %0, %1, [%3]\n" \ > " cmp %0, #0\n" \ > + " moveq %1, %2\n" \ > " movne %0, %4\n" \ > "2:\n" \ > " .section .text.fixup,\"ax\"\n" \ > I wonder what is status of this patch? I see that arm64 counterpart has landed in mainline, but I didn't manage to track down this one - is it because it never been submitted to patch system? Thanks Vladimir
On Tue, Dec 08, 2015 at 10:22:53AM +0000, Vladimir Murzin wrote: > On 15/10/15 10:30, Shengjiu Wang wrote: > > __user_swpX_asm maybe failed in first STREX operation, emulate_swpX > > will try again, but the *data has been changed in first time. which > > causes the result is wrong. > > This patch is to fix this issue. When STREX succeed, change the *data. > > if it fail, *data is not changed. > > > > Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com> > > --- > > arch/arm/kernel/swp_emulate.c | 6 +++--- > > 1 file changed, 3 insertions(+), 3 deletions(-) > > > > diff --git a/arch/arm/kernel/swp_emulate.c b/arch/arm/kernel/swp_emulate.c > > index 5b26e7e..c3fe769d 100644 > > --- a/arch/arm/kernel/swp_emulate.c > > +++ b/arch/arm/kernel/swp_emulate.c > > @@ -36,10 +36,10 @@ > > */ > > #define __user_swpX_asm(data, addr, res, temp, B) \ > > __asm__ __volatile__( \ > > - " mov %2, %1\n" \ > > - "0: ldrex"B" %1, [%3]\n" \ > > - "1: strex"B" %0, %2, [%3]\n" \ > > + "0: ldrex"B" %2, [%3]\n" \ > > + "1: strex"B" %0, %1, [%3]\n" \ > > " cmp %0, #0\n" \ > > + " moveq %1, %2\n" \ > > " movne %0, %4\n" \ > > "2:\n" \ > > " .section .text.fixup,\"ax\"\n" \ > > > > I wonder what is status of this patch? I see that arm64 counterpart has > landed in mainline, but I didn't manage to track down this one - is it > because it never been submitted to patch system? Most probably.
On 08/12/15 10:44, Russell King - ARM Linux wrote: > On Tue, Dec 08, 2015 at 10:22:53AM +0000, Vladimir Murzin wrote: >> On 15/10/15 10:30, Shengjiu Wang wrote: >>> __user_swpX_asm maybe failed in first STREX operation, emulate_swpX >>> will try again, but the *data has been changed in first time. which >>> causes the result is wrong. >>> This patch is to fix this issue. When STREX succeed, change the *data. >>> if it fail, *data is not changed. >>> >>> Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com> >>> --- >>> arch/arm/kernel/swp_emulate.c | 6 +++--- >>> 1 file changed, 3 insertions(+), 3 deletions(-) >>> >>> diff --git a/arch/arm/kernel/swp_emulate.c b/arch/arm/kernel/swp_emulate.c >>> index 5b26e7e..c3fe769d 100644 >>> --- a/arch/arm/kernel/swp_emulate.c >>> +++ b/arch/arm/kernel/swp_emulate.c >>> @@ -36,10 +36,10 @@ >>> */ >>> #define __user_swpX_asm(data, addr, res, temp, B) \ >>> __asm__ __volatile__( \ >>> - " mov %2, %1\n" \ >>> - "0: ldrex"B" %1, [%3]\n" \ >>> - "1: strex"B" %0, %2, [%3]\n" \ >>> + "0: ldrex"B" %2, [%3]\n" \ >>> + "1: strex"B" %0, %1, [%3]\n" \ >>> " cmp %0, #0\n" \ >>> + " moveq %1, %2\n" \ >>> " movne %0, %4\n" \ >>> "2:\n" \ >>> " .section .text.fixup,\"ax\"\n" \ >>> >> >> I wonder what is status of this patch? I see that arm64 counterpart has >> landed in mainline, but I didn't manage to track down this one - is it >> because it never been submitted to patch system? > > Most probably. > I've just submitted (on behalf of Shengjiu) it as 8475/1 with assumption you have no concerns about the patch. Thanks Vladimir
diff --git a/arch/arm/kernel/swp_emulate.c b/arch/arm/kernel/swp_emulate.c index 5b26e7e..c3fe769d 100644 --- a/arch/arm/kernel/swp_emulate.c +++ b/arch/arm/kernel/swp_emulate.c @@ -36,10 +36,10 @@ */ #define __user_swpX_asm(data, addr, res, temp, B) \ __asm__ __volatile__( \ - " mov %2, %1\n" \ - "0: ldrex"B" %1, [%3]\n" \ - "1: strex"B" %0, %2, [%3]\n" \ + "0: ldrex"B" %2, [%3]\n" \ + "1: strex"B" %0, %1, [%3]\n" \ " cmp %0, #0\n" \ + " moveq %1, %2\n" \ " movne %0, %4\n" \ "2:\n" \ " .section .text.fixup,\"ax\"\n" \
__user_swpX_asm maybe failed in first STREX operation, emulate_swpX will try again, but the *data has been changed in first time. which causes the result is wrong. This patch is to fix this issue. When STREX succeed, change the *data. if it fail, *data is not changed. Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com> --- arch/arm/kernel/swp_emulate.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)