From: Michael Ellerman <mpe@ellerman.id.au>
To: "Michal Suchánek" <msuchanek@suse.de>
Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>,
David Hildenbrand <david@redhat.com>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
Paul Mackerras <paulus@samba.org>,
Breno Leitao <leitao@debian.org>,
Michael Neuling <mikey@neuling.org>,
Diana Craciun <diana.craciun@nxp.com>,
Firoz Khan <firoz.khan@linaro.org>, Joel Stanley <joel@jms.id.au>,
Arnd Bergmann <arnd@arndb.de>,
Nicholas Piggin <npiggin@gmail.com>,
Alexander Viro <viro@zeniv.linux.org.uk>,
Thomas Gleixner <tglx@linutronix.de>,
Allison Randal <allison@lohutok.net>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
linux-kernel@vger.kernel.org,
Hari Bathini <hbathini@linux.ibm.com>,
"Eric W. Biederman" <ebiederm@xmission.com>,
Andrew Donnellan <andrew.donnellan@au1.ibm.com>,
linux-fsdevel@vger.kernel.org,
Andrew Morton <akpm@linux-foundation.org>,
linuxppc-dev@lists.ozlabs.org
Subject: Re: [PATCH v7 3/6] powerpc/perf: consolidate read_user_stack_32
Date: Tue, 03 Sep 2019 09:45:05 +1000 [thread overview]
Message-ID: <87mufms1oe.fsf@mpe.ellerman.id.au> (raw)
In-Reply-To: <20190902102653.6d477e16@naga>
Michal Suchánek <msuchanek@suse.de> writes:
> On Mon, 02 Sep 2019 14:01:17 +1000
> Michael Ellerman <mpe@ellerman.id.au> wrote:
>> Michael Ellerman <mpe@ellerman.id.au> writes:
>> > Michal Suchanek <msuchanek@suse.de> writes:
>> ...
>> >> @@ -295,6 +279,12 @@ static inline int current_is_64bit(void)
>> >> }
>> >>
>> >> #else /* CONFIG_PPC64 */
>> >> +static int read_user_stack_slow(void __user *ptr, void *buf, int nb)
>> >> +{
>> >> + return 0;
>> >> +}
>> >> +#endif /* CONFIG_PPC64 */
>> >
>> > Ending the PPC64 else case here, and then restarting it below with an
>> > ifndef means we end up with two parts of the file that define 32-bit
>> > code, with a common chunk in the middle, which I dislike.
>> >
>> > I'd rather you add the empty read_user_stack_slow() in the existing
>> > #else section and then move read_user_stack_32() below the whole ifdef
>> > PPC64/else/endif section.
>> >
>> > Is there some reason that doesn't work?
>>
>> Gah, I missed that you split the whole file later in the series. Any
>> reason you did it in two steps rather than moving patch 6 earlier in the
>> series?
>
> To make this patch readable.
But it's not very readable :)
You also retained the comment about the 32-bit behaviour which is now a
bit confusing, because the function is used on both 32 & 64-bit.
I think moving it as I suggested in my first reply makes for a better
diff, something like eg:
diff --git a/arch/powerpc/perf/callchain.c b/arch/powerpc/perf/callchain.c
index c84bbd4298a0..82c0f81b89a5 100644
--- a/arch/powerpc/perf/callchain.c
+++ b/arch/powerpc/perf/callchain.c
@@ -165,22 +165,6 @@ static int read_user_stack_64(unsigned long __user *ptr, unsigned long *ret)
return read_user_stack_slow(ptr, ret, 8);
}
-static int read_user_stack_32(unsigned int __user *ptr, unsigned int *ret)
-{
- if ((unsigned long)ptr > TASK_SIZE - sizeof(unsigned int) ||
- ((unsigned long)ptr & 3))
- return -EFAULT;
-
- pagefault_disable();
- if (!__get_user_inatomic(*ret, ptr)) {
- pagefault_enable();
- return 0;
- }
- pagefault_enable();
-
- return read_user_stack_slow(ptr, ret, 4);
-}
-
static inline int valid_user_sp(unsigned long sp, int is_64)
{
if (!sp || (sp & 7) || sp > (is_64 ? TASK_SIZE : 0x100000000UL) - 32)
@@ -295,27 +279,6 @@ static inline int current_is_64bit(void)
}
#else /* CONFIG_PPC64 */
-/*
- * On 32-bit we just access the address and let hash_page create a
- * HPTE if necessary, so there is no need to fall back to reading
- * the page tables. Since this is called at interrupt level,
- * do_page_fault() won't treat a DSI as a page fault.
- */
-static int read_user_stack_32(unsigned int __user *ptr, unsigned int *ret)
-{
- int rc;
-
- if ((unsigned long)ptr > TASK_SIZE - sizeof(unsigned int) ||
- ((unsigned long)ptr & 3))
- return -EFAULT;
-
- pagefault_disable();
- rc = __get_user_inatomic(*ret, ptr);
- pagefault_enable();
-
- return rc;
-}
-
static inline void perf_callchain_user_64(struct perf_callchain_entry_ctx *entry,
struct pt_regs *regs)
{
@@ -333,6 +296,11 @@ static inline int valid_user_sp(unsigned long sp, int is_64)
return 1;
}
+static int read_user_stack_slow(void __user *ptr, void *buf, int nb)
+{
+ return 0;
+}
+
#define __SIGNAL_FRAMESIZE32 __SIGNAL_FRAMESIZE
#define sigcontext32 sigcontext
#define mcontext32 mcontext
@@ -341,6 +309,33 @@ static inline int valid_user_sp(unsigned long sp, int is_64)
#endif /* CONFIG_PPC64 */
+static int read_user_stack_32(unsigned int __user *ptr, unsigned int *ret)
+{
+ int rc;
+
+ if ((unsigned long)ptr > TASK_SIZE - sizeof(unsigned int) ||
+ ((unsigned long)ptr & 3))
+ return -EFAULT;
+
+ pagefault_disable();
+ rc = __get_user_inatomic(*ret, ptr);
+ pagefault_enable();
+
+ /*
+ * On 32-bit we just access the address and let hash_page() create a
+ * HPTE if necessary, so there is no need to fall back to reading the
+ * page tables. Since this is called at interrupt level, do_page_fault()
+ * won't treat a DSI as a page fault.
+ *
+ * On 64-bit if the access faults we fall back to
+ * read_user_stack_slow(), see the comment there for more details.
+ */
+ if (IS_ENABLED(CONFIG_PPC64) && rc)
+ return read_user_stack_slow(ptr, ret, 4);
+
+ return rc;
+}
+
/*
* Layout for non-RT signal frames
*/
cheers
WARNING: multiple messages have this Message-ID (diff)
From: Michael Ellerman <mpe@ellerman.id.au>
To: "Michal Suchánek" <msuchanek@suse.de>
Cc: linuxppc-dev@lists.ozlabs.org,
Madhavan Srinivasan <maddy@linux.vnet.ibm.com>,
David Hildenbrand <david@redhat.com>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
Paul Mackerras <paulus@samba.org>,
Breno Leitao <leitao@debian.org>,
Michael Neuling <mikey@neuling.org>,
Diana Craciun <diana.craciun@nxp.com>,
Firoz Khan <firoz.khan@linaro.org>,
Hari Bathini <hbathini@linux.ibm.com>,
Joel Stanley <joel@jms.id.au>, Arnd Bergmann <arnd@arndb.de>,
Nicholas Piggin <npiggin@gmail.com>,
Alexander Viro <viro@zeniv.linux.org.uk>,
Thomas Gleixner <tglx@linutronix.de>,
Allison Randal <allison@lohutok.net>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
linux-kernel@vger.kernel.org,
"Eric W. Biederman" <ebiederm@xmission.com>,
Andrew Donnellan <andrew.donnellan@au1.ibm.com>,
linux-fsdevel@vger.kernel.org,
Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH v7 3/6] powerpc/perf: consolidate read_user_stack_32
Date: Tue, 03 Sep 2019 09:45:05 +1000 [thread overview]
Message-ID: <87mufms1oe.fsf@mpe.ellerman.id.au> (raw)
In-Reply-To: <20190902102653.6d477e16@naga>
Michal Suchánek <msuchanek@suse.de> writes:
> On Mon, 02 Sep 2019 14:01:17 +1000
> Michael Ellerman <mpe@ellerman.id.au> wrote:
>> Michael Ellerman <mpe@ellerman.id.au> writes:
>> > Michal Suchanek <msuchanek@suse.de> writes:
>> ...
>> >> @@ -295,6 +279,12 @@ static inline int current_is_64bit(void)
>> >> }
>> >>
>> >> #else /* CONFIG_PPC64 */
>> >> +static int read_user_stack_slow(void __user *ptr, void *buf, int nb)
>> >> +{
>> >> + return 0;
>> >> +}
>> >> +#endif /* CONFIG_PPC64 */
>> >
>> > Ending the PPC64 else case here, and then restarting it below with an
>> > ifndef means we end up with two parts of the file that define 32-bit
>> > code, with a common chunk in the middle, which I dislike.
>> >
>> > I'd rather you add the empty read_user_stack_slow() in the existing
>> > #else section and then move read_user_stack_32() below the whole ifdef
>> > PPC64/else/endif section.
>> >
>> > Is there some reason that doesn't work?
>>
>> Gah, I missed that you split the whole file later in the series. Any
>> reason you did it in two steps rather than moving patch 6 earlier in the
>> series?
>
> To make this patch readable.
But it's not very readable :)
You also retained the comment about the 32-bit behaviour which is now a
bit confusing, because the function is used on both 32 & 64-bit.
I think moving it as I suggested in my first reply makes for a better
diff, something like eg:
diff --git a/arch/powerpc/perf/callchain.c b/arch/powerpc/perf/callchain.c
index c84bbd4298a0..82c0f81b89a5 100644
--- a/arch/powerpc/perf/callchain.c
+++ b/arch/powerpc/perf/callchain.c
@@ -165,22 +165,6 @@ static int read_user_stack_64(unsigned long __user *ptr, unsigned long *ret)
return read_user_stack_slow(ptr, ret, 8);
}
-static int read_user_stack_32(unsigned int __user *ptr, unsigned int *ret)
-{
- if ((unsigned long)ptr > TASK_SIZE - sizeof(unsigned int) ||
- ((unsigned long)ptr & 3))
- return -EFAULT;
-
- pagefault_disable();
- if (!__get_user_inatomic(*ret, ptr)) {
- pagefault_enable();
- return 0;
- }
- pagefault_enable();
-
- return read_user_stack_slow(ptr, ret, 4);
-}
-
static inline int valid_user_sp(unsigned long sp, int is_64)
{
if (!sp || (sp & 7) || sp > (is_64 ? TASK_SIZE : 0x100000000UL) - 32)
@@ -295,27 +279,6 @@ static inline int current_is_64bit(void)
}
#else /* CONFIG_PPC64 */
-/*
- * On 32-bit we just access the address and let hash_page create a
- * HPTE if necessary, so there is no need to fall back to reading
- * the page tables. Since this is called at interrupt level,
- * do_page_fault() won't treat a DSI as a page fault.
- */
-static int read_user_stack_32(unsigned int __user *ptr, unsigned int *ret)
-{
- int rc;
-
- if ((unsigned long)ptr > TASK_SIZE - sizeof(unsigned int) ||
- ((unsigned long)ptr & 3))
- return -EFAULT;
-
- pagefault_disable();
- rc = __get_user_inatomic(*ret, ptr);
- pagefault_enable();
-
- return rc;
-}
-
static inline void perf_callchain_user_64(struct perf_callchain_entry_ctx *entry,
struct pt_regs *regs)
{
@@ -333,6 +296,11 @@ static inline int valid_user_sp(unsigned long sp, int is_64)
return 1;
}
+static int read_user_stack_slow(void __user *ptr, void *buf, int nb)
+{
+ return 0;
+}
+
#define __SIGNAL_FRAMESIZE32 __SIGNAL_FRAMESIZE
#define sigcontext32 sigcontext
#define mcontext32 mcontext
@@ -341,6 +309,33 @@ static inline int valid_user_sp(unsigned long sp, int is_64)
#endif /* CONFIG_PPC64 */
+static int read_user_stack_32(unsigned int __user *ptr, unsigned int *ret)
+{
+ int rc;
+
+ if ((unsigned long)ptr > TASK_SIZE - sizeof(unsigned int) ||
+ ((unsigned long)ptr & 3))
+ return -EFAULT;
+
+ pagefault_disable();
+ rc = __get_user_inatomic(*ret, ptr);
+ pagefault_enable();
+
+ /*
+ * On 32-bit we just access the address and let hash_page() create a
+ * HPTE if necessary, so there is no need to fall back to reading the
+ * page tables. Since this is called at interrupt level, do_page_fault()
+ * won't treat a DSI as a page fault.
+ *
+ * On 64-bit if the access faults we fall back to
+ * read_user_stack_slow(), see the comment there for more details.
+ */
+ if (IS_ENABLED(CONFIG_PPC64) && rc)
+ return read_user_stack_slow(ptr, ret, 4);
+
+ return rc;
+}
+
/*
* Layout for non-RT signal frames
*/
cheers
next prev parent reply other threads:[~2019-09-02 23:47 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-30 21:03 [PATCH v7 0/6] Disable compat cruft on ppc64le v7 Michal Suchanek
2019-08-30 21:03 ` Michal Suchanek
2019-08-30 21:03 ` [PATCH v7 1/6] powerpc: Add back __ARCH_WANT_SYS_LLSEEK macro Michal Suchanek
2019-08-30 21:03 ` Michal Suchanek
2019-08-30 21:03 ` [PATCH v7 2/6] powerpc: move common register copy functions from signal_32.c to signal.c Michal Suchanek
2019-08-30 21:03 ` Michal Suchanek
2019-08-30 21:03 ` [PATCH v7 3/6] powerpc/perf: consolidate read_user_stack_32 Michal Suchanek
2019-08-30 21:03 ` Michal Suchanek
2019-09-02 3:53 ` Michael Ellerman
2019-09-02 3:53 ` Michael Ellerman
2019-09-02 4:01 ` Michael Ellerman
2019-09-02 4:01 ` Michael Ellerman
2019-09-02 8:26 ` Michal Suchánek
2019-09-02 8:26 ` Michal Suchánek
2019-09-02 23:45 ` Michael Ellerman [this message]
2019-09-02 23:45 ` Michael Ellerman
2019-08-30 21:03 ` [PATCH v7 4/6] powerpc/64: make buildable without CONFIG_COMPAT Michal Suchanek
2019-08-30 21:03 ` Michal Suchanek
2019-08-31 6:43 ` Christophe Leroy
2019-08-31 6:43 ` Christophe Leroy
2019-08-31 13:02 ` Michal Suchánek
2019-08-31 13:02 ` Michal Suchánek
2019-08-30 21:03 ` [PATCH v7 5/6] powerpc/64: Make COMPAT user-selectable disabled on littleendian by default Michal Suchanek
2019-08-30 21:03 ` Michal Suchanek
2019-09-02 2:03 ` Michael Ellerman
2019-09-02 2:03 ` Michael Ellerman
2019-09-02 9:42 ` Michal Suchánek
2019-09-02 9:42 ` Michal Suchánek
2019-09-03 0:00 ` Michael Ellerman
2019-09-03 0:00 ` Michael Ellerman
2019-09-14 10:22 ` Michal Suchánek
2019-09-14 10:22 ` Michal Suchánek
2019-09-18 5:36 ` Michael Ellerman
2019-09-18 5:36 ` Michael Ellerman
2019-09-02 13:00 ` Segher Boessenkool
2019-09-02 13:00 ` Segher Boessenkool
2019-09-02 23:53 ` Michael Ellerman
2019-09-02 23:53 ` Michael Ellerman
2019-09-03 5:21 ` Christophe Leroy
2019-09-03 5:21 ` Christophe Leroy
2019-08-30 21:03 ` [PATCH v7 6/6] powerpc/perf: split callchain.c by bitness Michal Suchanek
2019-08-30 21:03 ` Michal Suchanek
2019-08-31 18:51 ` Michal Suchánek
2019-08-31 18:51 ` Michal Suchánek
2019-08-31 6:41 ` [PATCH v7 0/6] Disable compat cruft on ppc64le v7 Christophe Leroy
2019-08-31 6:41 ` Christophe Leroy
2019-08-31 18:03 ` Michal Suchánek
2019-08-31 18:03 ` Michal Suchánek
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87mufms1oe.fsf@mpe.ellerman.id.au \
--to=mpe@ellerman.id.au \
--cc=akpm@linux-foundation.org \
--cc=allison@lohutok.net \
--cc=andrew.donnellan@au1.ibm.com \
--cc=arnd@arndb.de \
--cc=david@redhat.com \
--cc=diana.craciun@nxp.com \
--cc=ebiederm@xmission.com \
--cc=firoz.khan@linaro.org \
--cc=gregkh@linuxfoundation.org \
--cc=hbathini@linux.ibm.com \
--cc=heiko.carstens@de.ibm.com \
--cc=joel@jms.id.au \
--cc=leitao@debian.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=maddy@linux.vnet.ibm.com \
--cc=mikey@neuling.org \
--cc=msuchanek@suse.de \
--cc=npiggin@gmail.com \
--cc=paulus@samba.org \
--cc=tglx@linutronix.de \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.