Message ID | 1462942500-30188-1-git-send-email-rashmicy@gmail.com (mailing list archive) |
---|---|
State | Superseded, archived |
Headers | show |
On 11/05/16 14:55, Rashmica Gupta wrote: > The same logic for tm_abort appears twice, so pull it out into a > function. > > Signed-off-by: Rashmica Gupta <rashmicy@gmail.com> > --- > arch/powerpc/mm/hash_utils_64.c | 47 ++++++++++++++++++----------------------- > 1 file changed, 21 insertions(+), 26 deletions(-) > > diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c > index 7635b1c6b5da..1cef8f5aee9b 100644 > --- a/arch/powerpc/mm/hash_utils_64.c > +++ b/arch/powerpc/mm/hash_utils_64.c > @@ -1318,6 +1318,25 @@ out_exit: > local_irq_restore(flags); > } > > +#ifdef CONFIG_PPC_TRANSACTIONAL_MEM > + /* Transactions are not aborted by tlbiel, only tlbie. > + * Without, syncing a page back to a block device w/ PIO could pick up > + * transactional data (bad!) so we force an abort here. Before the > + * sync the page will be made read-only, which will flush_hash_page. > + * BIG ISSUE here: if the kernel uses a page from userspace without > + * unmapping it first, it may see the speculated version. > + */ > +static inline void abort_tm(int local) > +{ > + if (local && cpu_has_feature(CPU_FTR_TM) && > + current->thread.regs && > + MSR_TM_ACTIVE(current->thread.regs->msr)) { > + tm_enable(); > + tm_abort(TM_CAUSE_TLBI); > + } > +} While your at this do #else static inline void abort_tm(int local) { } > +#endif > + > /* WARNING: This is called from hash_low_64.S, if you change this prototype, > * do not forget to update the assembly call site ! > */ > @@ -1344,19 +1363,7 @@ void flush_hash_page(unsigned long vpn, real_pte_t pte, int psize, int ssize, > } pte_iterate_hashed_end(); > > #ifdef CONFIG_PPC_TRANSACTIONAL_MEM Then remove these extra #ifdef > - /* Transactions are not aborted by tlbiel, only tlbie. > - * Without, syncing a page back to a block device w/ PIO could pick up > - * transactional data (bad!) so we force an abort here. Before the > - * sync the page will be made read-only, which will flush_hash_page. > - * BIG ISSUE here: if the kernel uses a page from userspace without > - * unmapping it first, it may see the speculated version. > - */ > - if (local && cpu_has_feature(CPU_FTR_TM) && > - current->thread.regs && > - MSR_TM_ACTIVE(current->thread.regs->msr)) { > - tm_enable(); > - tm_abort(TM_CAUSE_TLBI); > - } > + abort_tm(local); > #endif > } > > @@ -1415,19 +1422,7 @@ void flush_hash_hugepage(unsigned long vsid, unsigned long addr, > } > tm_abort: > #ifdef CONFIG_PPC_TRANSACTIONAL_MEM Then remove these extra #ifdef > - /* Transactions are not aborted by tlbiel, only tlbie. > - * Without, syncing a page back to a block device w/ PIO could pick up > - * transactional data (bad!) so we force an abort here. Before the > - * sync the page will be made read-only, which will flush_hash_page. > - * BIG ISSUE here: if the kernel uses a page from userspace without > - * unmapping it first, it may see the speculated version. > - */ > - if (local && cpu_has_feature(CPU_FTR_TM) && > - current->thread.regs && > - MSR_TM_ACTIVE(current->thread.regs->msr)) { > - tm_enable(); > - tm_abort(TM_CAUSE_TLBI); > - } > + abort_tm(local); > #endif > return; > } >
On 11/05/16 15:03, Balbir Singh wrote: > > On 11/05/16 14:55, Rashmica Gupta wrote: >> The same logic for tm_abort appears twice, so pull it out into a >> function. >> >> Signed-off-by: Rashmica Gupta <rashmicy@gmail.com> >> --- >> arch/powerpc/mm/hash_utils_64.c | 47 ++++++++++++++++++----------------------- >> 1 file changed, 21 insertions(+), 26 deletions(-) >> >> diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c >> index 7635b1c6b5da..1cef8f5aee9b 100644 >> --- a/arch/powerpc/mm/hash_utils_64.c >> +++ b/arch/powerpc/mm/hash_utils_64.c >> @@ -1318,6 +1318,25 @@ out_exit: >> local_irq_restore(flags); >> } >> >> +#ifdef CONFIG_PPC_TRANSACTIONAL_MEM >> + /* Transactions are not aborted by tlbiel, only tlbie. >> + * Without, syncing a page back to a block device w/ PIO could pick up >> + * transactional data (bad!) so we force an abort here. Before the >> + * sync the page will be made read-only, which will flush_hash_page. >> + * BIG ISSUE here: if the kernel uses a page from userspace without >> + * unmapping it first, it may see the speculated version. >> + */ >> +static inline void abort_tm(int local) >> +{ >> + if (local && cpu_has_feature(CPU_FTR_TM) && >> + current->thread.regs && >> + MSR_TM_ACTIVE(current->thread.regs->msr)) { >> + tm_enable(); >> + tm_abort(TM_CAUSE_TLBI); >> + } >> +} > While your at this do > > #else > > static inline void abort_tm(int local) > { > } If I'm doing that, wouldn't it make more sense to write: +static inline void abort_tm(int local) +{ +#ifdef CONFIG_PPC_TRANSACTIONAL_MEM + if (local && cpu_has_feature(CPU_FTR_TM) && + current->thread.regs && + MSR_TM_ACTIVE(current->thread.regs->msr)) { + tm_enable(); + tm_abort(TM_CAUSE_TLBI); + } +#endif +} >> +#endif >> + >> /* WARNING: This is called from hash_low_64.S, if you change this prototype, >> * do not forget to update the assembly call site ! >> */ >> @@ -1344,19 +1363,7 @@ void flush_hash_page(unsigned long vpn, real_pte_t pte, int psize, int ssize, >> } pte_iterate_hashed_end(); >> >> #ifdef CONFIG_PPC_TRANSACTIONAL_MEM > Then remove these extra #ifdef >> - /* Transactions are not aborted by tlbiel, only tlbie. >> - * Without, syncing a page back to a block device w/ PIO could pick up >> - * transactional data (bad!) so we force an abort here. Before the >> - * sync the page will be made read-only, which will flush_hash_page. >> - * BIG ISSUE here: if the kernel uses a page from userspace without >> - * unmapping it first, it may see the speculated version. >> - */ >> - if (local && cpu_has_feature(CPU_FTR_TM) && >> - current->thread.regs && >> - MSR_TM_ACTIVE(current->thread.regs->msr)) { >> - tm_enable(); >> - tm_abort(TM_CAUSE_TLBI); >> - } >> + abort_tm(local); >> #endif >> } >> >> @@ -1415,19 +1422,7 @@ void flush_hash_hugepage(unsigned long vsid, unsigned long addr, >> } >> tm_abort: >> #ifdef CONFIG_PPC_TRANSACTIONAL_MEM > Then remove these extra #ifdef >> - /* Transactions are not aborted by tlbiel, only tlbie. >> - * Without, syncing a page back to a block device w/ PIO could pick up >> - * transactional data (bad!) so we force an abort here. Before the >> - * sync the page will be made read-only, which will flush_hash_page. >> - * BIG ISSUE here: if the kernel uses a page from userspace without >> - * unmapping it first, it may see the speculated version. >> - */ >> - if (local && cpu_has_feature(CPU_FTR_TM) && >> - current->thread.regs && >> - MSR_TM_ACTIVE(current->thread.regs->msr)) { >> - tm_enable(); >> - tm_abort(TM_CAUSE_TLBI); >> - } >> + abort_tm(local); >> #endif >> return; >> } >>
On Wed, 2016-05-11 at 17:44 +1000, Rashmica wrote: > On 11/05/16 15:03, Balbir Singh wrote: > > On 11/05/16 14:55, Rashmica Gupta wrote: > > > The same logic for tm_abort appears twice, so pull it out into a > > > function. > > > > > > diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c > > > index 7635b1c6b5da..1cef8f5aee9b 100644 > > > --- a/arch/powerpc/mm/hash_utils_64.c > > > +++ b/arch/powerpc/mm/hash_utils_64.c > > > @@ -1318,6 +1318,25 @@ out_exit: > > > local_irq_restore(flags); > > > } > > > > > > +#ifdef CONFIG_PPC_TRANSACTIONAL_MEM > > > + /* Transactions are not aborted by tlbiel, only tlbie. Can you fix up the comment formatting while you're there, should be: /* * Transactions are not aborted by tlbiel, only tlbie. > > > + * Without, syncing a page back to a block device w/ PIO could pick up > > > + * transactional data (bad!) so we force an abort here. Before the > > > + * sync the page will be made read-only, which will flush_hash_page. > > > + * BIG ISSUE here: if the kernel uses a page from userspace without > > > + * unmapping it first, it may see the speculated version. > > > + */ > > > +static inline void abort_tm(int local) > > > +{ > > > + if (local && cpu_has_feature(CPU_FTR_TM) && > > > + current->thread.regs && > > > + MSR_TM_ACTIVE(current->thread.regs->msr)) { > > > + tm_enable(); > > > + tm_abort(TM_CAUSE_TLBI); > > > + } > > > +} > > While your at this do > > > > #else > > > > static inline void abort_tm(int local) > > { > > } > If I'm doing that, wouldn't it make more sense to write: > > +static inline void abort_tm(int local) It needs a better name. Perhaps tlb_flush_abort_tm() ? > +{ > +#ifdef CONFIG_PPC_TRANSACTIONAL_MEM > + if (local && cpu_has_feature(CPU_FTR_TM) && > + current->thread.regs && > + MSR_TM_ACTIVE(current->thread.regs->msr)) { > + tm_enable(); > + tm_abort(TM_CAUSE_TLBI); > + } > +#endif > +} In this case it would make some sense. But the goal is "no #ifdef's in C code", which that would violate. And if you have two functions that need to be inside the #ifdef then the above doesn't work anymore. So the preferred style is: #ifdef CONFIG_FOO static void foo(int bar) { ... } #else /* !CONFIG_FOO */ static inline foo(int bar) { } #endif And when it expands you end up with: #ifdef CONFIG_FOO static void foo(int bar) { ... } static void foo_new(int bar) { ... } #else /* !CONFIG_FOO */ static inline foo(int bar) { } static inline foo_new(int bar) { } #endif cheers
diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c index 7635b1c6b5da..1cef8f5aee9b 100644 --- a/arch/powerpc/mm/hash_utils_64.c +++ b/arch/powerpc/mm/hash_utils_64.c @@ -1318,6 +1318,25 @@ out_exit: local_irq_restore(flags); } +#ifdef CONFIG_PPC_TRANSACTIONAL_MEM + /* Transactions are not aborted by tlbiel, only tlbie. + * Without, syncing a page back to a block device w/ PIO could pick up + * transactional data (bad!) so we force an abort here. Before the + * sync the page will be made read-only, which will flush_hash_page. + * BIG ISSUE here: if the kernel uses a page from userspace without + * unmapping it first, it may see the speculated version. + */ +static inline void abort_tm(int local) +{ + if (local && cpu_has_feature(CPU_FTR_TM) && + current->thread.regs && + MSR_TM_ACTIVE(current->thread.regs->msr)) { + tm_enable(); + tm_abort(TM_CAUSE_TLBI); + } +} +#endif + /* WARNING: This is called from hash_low_64.S, if you change this prototype, * do not forget to update the assembly call site ! */ @@ -1344,19 +1363,7 @@ void flush_hash_page(unsigned long vpn, real_pte_t pte, int psize, int ssize, } pte_iterate_hashed_end(); #ifdef CONFIG_PPC_TRANSACTIONAL_MEM - /* Transactions are not aborted by tlbiel, only tlbie. - * Without, syncing a page back to a block device w/ PIO could pick up - * transactional data (bad!) so we force an abort here. Before the - * sync the page will be made read-only, which will flush_hash_page. - * BIG ISSUE here: if the kernel uses a page from userspace without - * unmapping it first, it may see the speculated version. - */ - if (local && cpu_has_feature(CPU_FTR_TM) && - current->thread.regs && - MSR_TM_ACTIVE(current->thread.regs->msr)) { - tm_enable(); - tm_abort(TM_CAUSE_TLBI); - } + abort_tm(local); #endif } @@ -1415,19 +1422,7 @@ void flush_hash_hugepage(unsigned long vsid, unsigned long addr, } tm_abort: #ifdef CONFIG_PPC_TRANSACTIONAL_MEM - /* Transactions are not aborted by tlbiel, only tlbie. - * Without, syncing a page back to a block device w/ PIO could pick up - * transactional data (bad!) so we force an abort here. Before the - * sync the page will be made read-only, which will flush_hash_page. - * BIG ISSUE here: if the kernel uses a page from userspace without - * unmapping it first, it may see the speculated version. - */ - if (local && cpu_has_feature(CPU_FTR_TM) && - current->thread.regs && - MSR_TM_ACTIVE(current->thread.regs->msr)) { - tm_enable(); - tm_abort(TM_CAUSE_TLBI); - } + abort_tm(local); #endif return; }
The same logic for tm_abort appears twice, so pull it out into a function. Signed-off-by: Rashmica Gupta <rashmicy@gmail.com> --- arch/powerpc/mm/hash_utils_64.c | 47 ++++++++++++++++++----------------------- 1 file changed, 21 insertions(+), 26 deletions(-)