[RFC][PATCHSET] regset ->get() rework

[RFC][PATCHSET] regset ->get() rework

[Al Viro]

There are several problems with regset API:

	regsets' methods are too generic, clumsy as hell and the inline helpers
intended to hide most of the headache have very inconvenient calling conventions.
For starters, ->get() (the method used to fetch a part of regset)
	* can have both userland and kernel destination (gets two pointers, one of
them being NULL)
	* can (theoretically) be called to fetch _any_ part of regset - possibly
starting in the middle of the damn thing.  There are very few (and very irregular)
users of that, er, feature.  Moreover, all of them know which ->get() instance they
are going to hit and quite a few ->get() instances outright say that they rely
upon nobody calling them that way.  Even more instances say nothing and still rely
upon the same.  And even the instances that *are* called that way rely upon the
knowledge of the specific offsets that might be requested.  "Brittle" doesn't
begin to describe that; it also makes the instances that attempt to handle arbitrary
offsets convoluted for no good reason.  And all the irregular callers are easily dealt
with without playing with that.
	* the helpers attempt to be too flexible and end up being too inconvenient.
They try to allow writing the pieces of data out of order - pwrite(2) rather than
write(2), if you will.  And their calling conventions are worse than those of
pwrite(2)...  It's also an attractive nuisance - if you end up having skipped
some range of offsets, there's no way for the caller to know.  Out of all
the instances only one tried to make use of those non-sequential writes - and
fucked up in process.
	* instances have to deal with the possibility of uaccess failures halfway
through.  Again, it complicates the logics for no good reason.
	* instances have no way to report how much data would they like to produce;
as the result, an additional method (->get_size()) had been invented (with only
one instance).  And handling of that thing in PTRACE_GETREGSET is broken - it's
*not* reported to the caller and there's no direct way to tell that the end of
userland buffer is left unmodified.

	Life would be much simpler if we replaced ->get() with something that
1) always writes to kernel space.
2) always writes sequentially, starting from offset 0 and up to given (bounded)
size.
3) returns an error or the amount of space left in the buffer.

	Turns out it's not that hard to do.  Callers that want the data to
go to userland can bloody well do copy_to_user() themselves.  Non-sequential
stores are simply not there anymore - not since the x86 bug got fixed last cycle.
 Callers that want non-zero starting offsets are very few and are essentially
trying to paste pieces of regset together, possibly with some parts
skipped/in different order/with different alignments/etc.  Those are easily
dealt with by providing a separate regset with the desired contents.  There are
4 locations where specific ->get() instances are called directly and 3 locations
that call ->get() instances via method.  Let's take them in order:

	[ia64] access_uarea() contains direct calls of fpregs_get() and gpregs_get().
Destination is a kernel buffer, the subset being accessed is a word at varying offset
in the corresponding regset.  That needs careful teasing apart; it's better to have
access_uarea() use the lower-level helpers instead of calling [gf]pregs_[gs]et() there.
Doable with some massage.

	[sh] dump_fpu() calls (sh) fpregs_get(); destination is a kernel buffer,
the entire regset contents gets written out.  Used by ELF_FDPIC coredumps;
ELF coredumps on sh are _not_ reaching that code.  We need to switch ELF_FDPIC
coredumps to the same machinery ELF coredumps use; then this will become dead
code.  However, as far as ->get() API change is concerned, that call is nothing
irregular - we write the entire regset out to a kernel buffer.  So we can leave
the removal of that code until we deal with ELF_FDPIC and for now just keep in
mind that this instance of ->get2() is called directly.  Just change the
arguments to match the new calling conventions and that's it.

	[x86] dump_fpu() calls (x86) fpregs_get().  That one is very simple to deal
with, seeing that the only caller used to be aout coredump support, which got (finally)
removed last year.  Dead code, IOW...

	[ELF] a couple of callers in fill_thread_core_info().  That's what ELF coredumps
normally use; destination is a kernel buffer, the entire regset gets written out.

	Finally, there's copy_regset_to_user().  That's the only place where we
ask ->get() to write to userland and it can bloody well allocate a kernel buffer,
let ->get() fill it and then do copy_to_user() of the entire thing.  That function
is used by handlers of various ptrace(2) commands (the most generic one is
PTRACE_GETREGSET).  Almost all callers ask to write a range starting at the
beginning of regset (and most of those ask to write the entire regset out);
However, there are few exceptions:

	[arm64] compat_ptrace_read_user() uses copy_regset_to_user() to fetch the
value of given register and write it to userland.  Pointless kludge, since it's
much easier to do what native arm does - supply a trivial function that returns
given register of given process and copy the result to userland, same way as we
do in other cases of compat_ptrace_read_user().

	[sparc] PTRACE_GETREGS and PTRACE_GETFPREGS cobble the layout they
want out of pieces of regsets.  It's much easier to provide a regset producing
just the layout they want and use it instead of the generic ones.

	IOW, we can regularize the callers with a bit of preliminary work,
then convert to new method with better calling conventions.  The following
series attempts to do that; it can be found in 
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git #work.regset
(and in the followups to this posting).  Based at 5.8-rc1, there are
5 short branches with preparations (#regset.{base,x86,sparc,arm64,ia64})
merged together and followed by the bulk of the series.

	Patch overviews follow:

#regset.base:
01/41) helpers for regset ->get() replacement
	We start with providing wrappers for ->get() that have saner calling
conventions than ->get() itself.  That is,
	* they always copy to kernel buffer
	* they are given the upper limit on the amount to copy and return the
amount actually copied or -E... on error.  If regset has less than the amount
required, it writes as much as it can.
	One of those wrappers copies into preexisting buffer, another allocates
and writes there (allocation size is bounded by the size limit and by the
amount regset is expected to have; user asking for half a gigabyte worth of
register contents won't trigger half a gigabyte allocation).
	Those wrappers live in (newly added) kernel/regset.c; none of them
is going to be called on sufficiently hot codepath to warrant inlining.
Initially we can't convert all callers of ->get() to those.  ELF coredump
callers can be immediately converted, but ptrace ones need more preparations.

Now we regularize the callers.  That's done on per-architecture basis:

#regset.x86
02/41) x86: copy_fpstate_to_sigframe(): have fpregs_soft_get() use kernel buffer
03/41) x86: kill dump_fpu()
	On x86 we have two irregularities: there's a part of an instance of ->get()
(helper with the identical calling conventions, that is) that gets directly called
with userland destination.  It's _not_ a fast path - we need FPU emulation to be used
in order to hit it, and at that point extra copying of 108 bytes is noise by any
definition.  Fuck it; call that ->get() instance with local variable as destination,
then copy_to_user().
	Another direct call of an instance is in dead code (dump_fpu(), not used
since the removal of aout coredumps).  That one simply gets removed.

NB: calling conventions of copy_fpstate_to_sigframe() are silly.  If you look
at the call chains, anything negative ends up with force_sigsegv(); 0 or any
positive qualifies as success.  IOW, it really returns just one bit of information,
while creating an impression that we somehow are reporting details.  We are, after
all, trying to save the userland state while entering a signal handler; whom
could we possibly return an error to?  copy_fpstate_to_sigframe(), along with
__setup_rt_frame(), etc. ought to return bool.

#regset.ia64:
04/41) [ia64] sanitize elf_access_gpreg()
05/41) [ia64] teach elf_access_reg() to handle the missing range (r16..r31)
06/41) [ia64] regularize do_gpregs_[gs]et()
07/41) [ia64] access_uarea(): stop bothering with gpregs_[gs]et()
08/41) [ia64] access_uarea(): don't bother with fpregs_[gs]et()

	The problem here is that access_uarea() (the guts of PTRAGE_POKEUSR/
PTRACE_PEEKUSR) contains direct calls of ->get() (and ->set()) instances,
for accessing single registers.  They already have a function for doing
pretty much that - access_elf_reg() (almost - it doesn't handle r16..r31,
but that's trivial to do, simplifying the logics in callers).

	gpregs_get() and gpregs_set() arrange for stack unwinder to collect
the information of where the user callee-saved registers have ended up
spilled on kernel stack.  Then they start calling access_elf_reg(), with
bloody awful switches that remap the offsets in regset to register numbers.
access_uarea(), in turn, remaps _its_ offsets into regset ones and calls
gpregs_get()/gpregs_set().

	A lot of extra complexity comes from gpregs_get() trying to bunch
the copyouts together - it puts the values of several registers into an
array, then copies them out all at once.

	In that branch:
* clean elf_access_breg() up, get rid of the switch in the damn thing.
* teach elf_access_reg() to handle the missing range of registers (r16..r31)
* regularize the after-unwind part of gpregs_get()/gpregs_set() - with
elf_access_reg() becoming able to handle that range, the bunching logics in
those become much simpler.
* switch access_uarea() to calling elf_access_reg() directly, eliminating
the calls of gpregs_get()/gpregs_set() in there.
* same for floating point registers, only there we used to call different
->get()/->set() instances.

	One note on access_uarea(): it never returns positive and its
callers choose between using "not zero" and "negative" pretty much
at random.

#regset.sparc
09/41) sparc64: switch genregs32_get() to use of get_from_target()
10/41) sparc32: get rid of odd callers of copy_regset_to_user()
11/41) sparc64: get rid of odd callers of copy_regset_to_user()
12/41) sparc32: get rid of odd callers of copy_regset_from_user()
13/41) sparc64: get rid of odd callers of copy_regset_from_user()

	First patch in the series is a  cleanup in one of the ->get()
instances (open-coded reading the target's memory, rather than using
the helper already in the same file).  That's right next to the place
that will need modifications in ->get() conversion later, so splitting
that into a separate commit makes the things easier to follow.
	Next several "paste pieces of regset together and copy them to
userland" are dealt with by adding new regsets, with ->get() instances
directly producing the layout wanted by those ptrace commands, so that
a single copy_regset_to_user() would do the right thing.  Incidentally,
that simplifies ->get() instances that used to be called by those ptrace
commands - now they are never called with non-zero offset.
	While not strictly necessary for this series, the same thing is
done to copy_regset_from_user() counterparts of those, so that all
pasting pieces of regsets together would be gone.
	NB: PTRAGE_GETREGS64 has a minor behaviour change - it used to
leave pregs->uregs[15] untouched, now it zeroes it.

#regset.arm64:
14/41) arm64: take fetching compat reg out of pt_regs into a new helper
15/41) arm64: get rid of copy_regset_to_user() in compat_ptrace_read_user()
16/41) arm64: sanitize compat_ptrace_write_user()

	arm64 compat handling of PTRACE_PEEKUSR/PTRACE_POKEUSR abuses
copy_regset_{to,from}_user().  All we want in those is access to
a single register; it's easier to lift the logics for fetching
a register out of compat_gpr_get() into a helper function, call
it in PTRACE_PEEKUSR handler and use put_user() to store its
return value; that's pretty much what the native arm ptrace does.
PTRACE_POKEUSR is, strictly speaking, unrelated to this series, but
it's too obscene to live - it has the value to store into given
register, so it does set_fs(KERNEL_DS) and calls copy_regset_from_user()
with the address of the local variable that contains the new value.
That's one hell of a way to copy a value into a field of pt_regs...

	At that point we are done with prep work.  The branches above
(#regset.{base,x86,ia64,sparc,arm64}) converge into #work.regset.

#work.regset
17/41) copy_regset_to_user(): do all copyout at once.
	By now callers of ->get() instances pass 0 as offset and
all but one (copy_regset_to_user()) have kernel-space destinations.
Convert that last caller to use of regset_get_alloc().

18/41) regset: new method and helpers for it
	This is the core of the series.  A replacement for ->get()
is introduced, hopefully with saner calling conventions.
	->get2() takes task+regset+buffer, returns the amount of
free space left in the buffer on success and -E... on error.
buffer is represented as struct membuf - a pair of (kernel) pointer
and amount of space left.  Note that buffer is passed by value -
that leads to better code generation, since the instances can
just keep the components in registers - stores to those don't need
to be visible to the caller.
	Primitives for writing to such:
		* membuf_write(buf, data, size)
		* membuf_zero(buf, size)
		* membuf_store(buf, value)
These are implemented as inlines (in case of membuf_store - a macro).
All writes are sequential; they become no-ops when there's no space
left.  Return value of all primitives is the amount of space left
after the operation, so they can be used as return values of ->get2().
    
Example of use:
// stores pt_regs of task + 64 bytes worth of zeroes + 32bit PID of task
int foo_get(struct task_struct *task, const struct regset *regset,
            struct membuf to)
{
	membuf_write(&to, task_pt_regs(task), sizeof(struct pt_regs));
	membuf_zero(&to, 64);
	return membuf_store(&to, (u32)task_tgid_vnr(task));
}

regset_get()/regset_get_alloc() taught to use that thing if present.
By the end of the series all users of ->get() will be converted;
then ->get() and ->get_size() can go.
	Note that unlike ->get() this thing always starts at offset 0
and, since it only writes to kernel buffer, can't fail on copyout.
It can, of course, fail for other reasons, but those tend to
be less numerous.
	The caller guarantees that the buffer size won't be bigger than
regset->n * regset->size.  That simplifies life for quite a few instances.


	Now we are going to convert architectures from ->get() to ->get2().

19/41) x86: switch to ->get2()
20/41) powerpc: switch to ->get2()
21/41) s390: switch to ->get2()
22/41) sparc: switch to ->get2()
23/41) mips: switch to ->get2()
24/41) arm64: switch to ->get2()
25/41) sh: convert to ->get2()
26/41) arm: switch to ->get2()
27/41) arc: switch to ->get2()
28/41) ia64: switch to ->get2()
29/41) c6x: switch to ->get2()
30/41) riscv: switch to ->get2()
31/41) openrisc: switch to ->get2()
32/41) h8300: switch to ->get2()
33/41) hexagon: switch to ->get2()
34/41) nios2: switch to ->get2()
35/41) nds32: switch to ->get2()
36/41) parisc: switch to ->get2()
37/41) xtensa: switch to ->get2()
38/41) csky: switch to ->get2()

With those done we have no ->get() instances left.  Moreover,
the only ->get_size() call is unreachable.  And no callers of
old helpers remain either.

39/41) regset: kill ->get()
40/41) regset(): kill ->get_size()
41/41) regset: kill user_regset_copyout{,_zero}()

	This is *not* all I have for regsets, but I'd rather keep
e.g. fdpic conversion to regsets as part of ELF coredump work.
This just deals with ->get().  Please, review and comment.

Diffstat:
 arch/arc/kernel/ptrace.c                    | 148 +++----
 arch/arm/kernel/ptrace.c                    |  52 +--
 arch/arm64/kernel/ptrace.c                  | 303 +++++---------
 arch/c6x/kernel/ptrace.c                    |  11 +-
 arch/csky/kernel/ptrace.c                   |  24 +-
 arch/h8300/kernel/ptrace.c                  |  17 +-
 arch/hexagon/kernel/ptrace.c                |  62 +--
 arch/ia64/kernel/ptrace.c                   | 396 +++++++------------
 arch/mips/kernel/ptrace.c                   | 204 +++-------
 arch/nds32/kernel/ptrace.c                  |   9 +-
 arch/nios2/kernel/ptrace.c                  |  51 +--
 arch/openrisc/kernel/ptrace.c               |  26 +-
 arch/parisc/kernel/ptrace.c                 |  84 +---
 arch/powerpc/kernel/ptrace/ptrace-altivec.c |  37 +-
 arch/powerpc/kernel/ptrace/ptrace-decl.h    |  44 +--
 arch/powerpc/kernel/ptrace/ptrace-novsx.c   |   5 +-
 arch/powerpc/kernel/ptrace/ptrace-spe.c     |  16 +-
 arch/powerpc/kernel/ptrace/ptrace-tm.c      | 152 +++----
 arch/powerpc/kernel/ptrace/ptrace-view.c    | 185 ++++-----
 arch/powerpc/kernel/ptrace/ptrace-vsx.c     |  13 +-
 arch/riscv/kernel/ptrace.c                  |  33 +-
 arch/s390/kernel/ptrace.c                   | 199 +++-------
 arch/sh/kernel/process_32.c                 |   5 +-
 arch/sh/kernel/ptrace_32.c                  |  48 +--
 arch/sparc/kernel/ptrace_32.c               | 265 +++++++------
 arch/sparc/kernel/ptrace_64.c               | 591 ++++++++++++++--------------
 arch/x86/include/asm/fpu/internal.h         |   1 -
 arch/x86/include/asm/fpu/regset.h           |   4 +-
 arch/x86/include/asm/fpu/xstate.h           |   4 +-
 arch/x86/kernel/fpu/regset.c                |  55 +--
 arch/x86/kernel/fpu/signal.c                |  13 +-
 arch/x86/kernel/fpu/xstate.c                | 164 ++------
 arch/x86/kernel/ptrace.c                    |  75 ++--
 arch/x86/kernel/tls.c                       |  32 +-
 arch/x86/kernel/tls.h                       |   2 +-
 arch/x86/math-emu/fpu_entry.c               |  19 +-
 arch/xtensa/kernel/ptrace.c                 |  16 +-
 fs/binfmt_elf.c                             |  54 ++-
 include/linux/regset.h                      | 218 +++-------
 kernel/Makefile                             |   2 +-
 kernel/regset.c                             |  76 ++++
 41 files changed, 1368 insertions(+), 2347 deletions(-)

[PATCH 01/41] introduction of regset ->get() wrappers, switching ELF coredumps to those

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Two new helpers: given a process and regset, dump into a buffer.
regset_get() takes a buffer and size, regset_get_alloc() takes size
and allocates a buffer.

Return value in both cases is the amount of data actually dumped in
case of success or -E...  on error.

In both cases the size is capped by regset->n * regset->size, so
->get() is called with offset 0 and size no more than what regset
expects.

binfmt_elf.c callers of ->get() are switched to using those; the other
caller (copy_regset_to_user()) will need some preparations to switch.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/binfmt_elf.c        | 54 ++++++++++++++++++++++++--------------------------
 include/linux/regset.h |  9 +++++++++
 kernel/Makefile        |  2 +-
 kernel/regset.c        | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 90 insertions(+), 29 deletions(-)
 create mode 100644 kernel/regset.c

diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index 9fe3b51c116a..80743b8957c9 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -1821,7 +1821,7 @@ static int fill_thread_core_info(struct elf_thread_core_info *t,
 				 long signr, size_t *total)
 {
 	unsigned int i;
-	unsigned int regset0_size = regset_size(t->task, &view->regsets[0]);
+	unsigned int regset0_size;
 
 	/*
 	 * NT_PRSTATUS is the one special case, because the regset data
@@ -1830,8 +1830,10 @@ static int fill_thread_core_info(struct elf_thread_core_info *t,
 	 * We assume that regset 0 is NT_PRSTATUS.
 	 */
 	fill_prstatus(&t->prstatus, t->task, signr);
-	(void) view->regsets[0].get(t->task, &view->regsets[0], 0, regset0_size,
-				    &t->prstatus.pr_reg, NULL);
+	regset0_size = regset_get(t->task, &view->regsets[0],
+		   sizeof(t->prstatus.pr_reg), &t->prstatus.pr_reg);
+	if (regset0_size < 0)
+		return 0;
 
 	fill_note(&t->notes[0], "CORE", NT_PRSTATUS,
 		  PRSTATUS_SIZE(t->prstatus, regset0_size), &t->prstatus);
@@ -1846,32 +1848,28 @@ static int fill_thread_core_info(struct elf_thread_core_info *t,
 	 */
 	for (i = 1; i < view->n; ++i) {
 		const struct user_regset *regset = &view->regsets[i];
+		int note_type = regset->core_note_type;
+		bool is_fpreg = note_type == NT_PRFPREG;
+		void *data;
+		int ret;
+
 		do_thread_regset_writeback(t->task, regset);
-		if (regset->core_note_type && regset->get &&
-		    (!regset->active || regset->active(t->task, regset) > 0)) {
-			int ret;
-			size_t size = regset_size(t->task, regset);
-			void *data = kzalloc(size, GFP_KERNEL);
-			if (unlikely(!data))
-				return 0;
-			ret = regset->get(t->task, regset,
-					  0, size, data, NULL);
-			if (unlikely(ret))
-				kfree(data);
-			else {
-				if (regset->core_note_type != NT_PRFPREG)
-					fill_note(&t->notes[i], "LINUX",
-						  regset->core_note_type,
-						  size, data);
-				else {
-					SET_PR_FPVALID(&t->prstatus,
-							1, regset0_size);
-					fill_note(&t->notes[i], "CORE",
-						  NT_PRFPREG, size, data);
-				}
-				*total += notesize(&t->notes[i]);
-			}
-		}
+		if (!note_type) // not for coredumps
+			continue;
+		if (regset->active && regset->active(t->task, regset) <= 0)
+			continue;
+
+		ret = regset_get_alloc(t->task, regset, ~0U, &data);
+		if (ret < 0)
+			continue;
+
+		if (is_fpreg)
+			SET_PR_FPVALID(&t->prstatus, 1, regset0_size);
+
+		fill_note(&t->notes[i], is_fpreg ? "CORE" : "LINUX",
+			  note_type, ret, data);
+
+		*total += notesize(&t->notes[i]);
 	}
 
 	return 1;
diff --git a/include/linux/regset.h b/include/linux/regset.h
index 46d6ae68c455..968a032922d5 100644
--- a/include/linux/regset.h
+++ b/include/linux/regset.h
@@ -353,6 +353,15 @@ static inline int user_regset_copyin_ignore(unsigned int *pos,
 	return 0;
 }
 
+extern int regset_get(struct task_struct *target,
+		      const struct user_regset *regset,
+		      unsigned int size, void *data);
+
+extern int regset_get_alloc(struct task_struct *target,
+			    const struct user_regset *regset,
+			    unsigned int size,
+			    void **data);
+
 /**
  * copy_regset_to_user - fetch a thread's user_regset data into user memory
  * @target:	thread to be examined
diff --git a/kernel/Makefile b/kernel/Makefile
index f3218bc5ec69..e6e03380a0f1 100644
--- a/kernel/Makefile
+++ b/kernel/Makefile
@@ -10,7 +10,7 @@ obj-y     = fork.o exec_domain.o panic.o \
 	    extable.o params.o \
 	    kthread.o sys_ni.o nsproxy.o \
 	    notifier.o ksysfs.o cred.o reboot.o \
-	    async.o range.o smpboot.o ucount.o
+	    async.o range.o smpboot.o ucount.o regset.o
 
 obj-$(CONFIG_MODULES) += kmod.o
 obj-$(CONFIG_MULTIUSER) += groups.o
diff --git a/kernel/regset.c b/kernel/regset.c
new file mode 100644
index 000000000000..6b39fa0993ec
--- /dev/null
+++ b/kernel/regset.c
@@ -0,0 +1,54 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include <linux/export.h>
+#include <linux/slab.h>
+#include <linux/regset.h>
+
+static int __regset_get(struct task_struct *target,
+			const struct user_regset *regset,
+			unsigned int size,
+			void **data)
+{
+	void *p = *data, *to_free = NULL;
+	int res;
+
+	if (!regset->get)
+		return -EOPNOTSUPP;
+	if (size > regset->n * regset->size)
+		size = regset->n * regset->size;
+	if (!p) {
+		to_free = p = kzalloc(size, GFP_KERNEL);
+		if (!p)
+			return -ENOMEM;
+	}
+	res = regset->get(target, regset, 0, size, p, NULL);
+	if (unlikely(res < 0)) {
+		kfree(to_free);
+		return res;
+	}
+	*data = p;
+	if (regset->get_size) { // arm64-only kludge, will go away
+		unsigned max_size = regset->get_size(target, regset);
+		if (size > max_size)
+			size = max_size;
+	}
+	return size;
+}
+
+int regset_get(struct task_struct *target,
+	       const struct user_regset *regset,
+	       unsigned int size,
+	       void *data)
+{
+	return __regset_get(target, regset, size, &data);
+}
+EXPORT_SYMBOL(regset_get);
+
+int regset_get_alloc(struct task_struct *target,
+		     const struct user_regset *regset,
+		     unsigned int size,
+		     void **data)
+{
+	*data = NULL;
+	return __regset_get(target, regset, size, data);
+}
+EXPORT_SYMBOL(regset_get_alloc);
-- 
2.11.0

[PATCH 02/41] x86: copy_fpstate_to_sigframe(): have fpregs_soft_get() use kernel buffer

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

... then copy_to_user() the results

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/x86/kernel/fpu/signal.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/arch/x86/kernel/fpu/signal.c b/arch/x86/kernel/fpu/signal.c
index 9393a445d73c..e0b832df7404 100644
--- a/arch/x86/kernel/fpu/signal.c
+++ b/arch/x86/kernel/fpu/signal.c
@@ -170,14 +170,14 @@ int copy_fpstate_to_sigframe(void __user *buf, void __user *buf_fx, int size)
 	ia32_fxstate &= (IS_ENABLED(CONFIG_X86_32) ||
 			 IS_ENABLED(CONFIG_IA32_EMULATION));
 
+	if (!static_cpu_has(X86_FEATURE_FPU)) {
+		struct user_i387_ia32_struct fp;
+		fpregs_soft_get(current, NULL, 0, sizeof(fp), &fp, NULL);
+		return copy_to_user(buf, &fp, sizeof(fp)) ? -EFAULT : 0;
+	}
+
 	if (!access_ok(buf, size))
 		return -EACCES;
-
-	if (!static_cpu_has(X86_FEATURE_FPU))
-		return fpregs_soft_get(current, NULL, 0,
-			sizeof(struct user_i387_ia32_struct), NULL,
-			(struct _fpstate_32 __user *) buf) ? -1 : 1;
-
 retry:
 	/*
 	 * Load the FPU registers if they are not valid for the current task.
-- 
2.11.0

[PATCH 03/41] x86: kill dump_fpu()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

dead since the removal of aout coredump support...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/x86/include/asm/fpu/internal.h |  1 -
 arch/x86/kernel/fpu/regset.c        | 16 ----------------
 2 files changed, 17 deletions(-)

diff --git a/arch/x86/include/asm/fpu/internal.h b/arch/x86/include/asm/fpu/internal.h
index 42159f45bf9c..0c85a08176c0 100644
--- a/arch/x86/include/asm/fpu/internal.h
+++ b/arch/x86/include/asm/fpu/internal.h
@@ -34,7 +34,6 @@ extern int  fpu__copy(struct task_struct *dst, struct task_struct *src);
 extern void fpu__clear_user_states(struct fpu *fpu);
 extern void fpu__clear_all(struct fpu *fpu);
 extern int  fpu__exception_code(struct fpu *fpu, int trap_nr);
-extern int  dump_fpu(struct pt_regs *ptregs, struct user_i387_struct *fpstate);
 
 /*
  * Boot time FPU initialization functions:
diff --git a/arch/x86/kernel/fpu/regset.c b/arch/x86/kernel/fpu/regset.c
index bd1d0649f8ce..91f80aca27fb 100644
--- a/arch/x86/kernel/fpu/regset.c
+++ b/arch/x86/kernel/fpu/regset.c
@@ -356,20 +356,4 @@ int fpregs_set(struct task_struct *target, const struct user_regset *regset,
 	return ret;
 }
 
-/*
- * FPU state for core dumps.
- * This is only used for a.out dumps now.
- * It is declared generically using elf_fpregset_t (which is
- * struct user_i387_struct) but is in fact only used for 32-bit
- * dumps, so on 64-bit it is really struct user_i387_ia32_struct.
- */
-int dump_fpu(struct pt_regs *regs, struct user_i387_struct *ufpu)
-{
-	struct task_struct *tsk = current;
-
-	return !fpregs_get(tsk, NULL, 0, sizeof(struct user_i387_ia32_struct),
-			   ufpu, NULL);
-}
-EXPORT_SYMBOL(dump_fpu);
-
 #endif	/* CONFIG_X86_32 || CONFIG_IA32_EMULATION */
-- 
2.11.0

[PATCH 04/41] [ia64] sanitize elf_access_gpreg()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

The function takes the register number, finds the corresponding field
of pt_regs for registers that are saved there or does the unwind for the
registers that end up spilled on the kernel stack.  Then it reads from
or writes to the resulting location.

Unfortunately, finding the required pt_regs field is done by rather
horrible switch.  It's microoptimized in all the wrong places - it
even uses the knowledge that fields for r8..r11 follow each other
in pt_regs layout, while r12..r13 are not adjacent to those, etc.

All of that is to encode the mapping from register numbers to offsets +
the information that r4..r7 are not to be found in pt_regs.

It's deeply in nasal demon territory, at that - the games it plays
with pointer arithmetics on addresses of structure members are
undefined behaviour.

Valid C ends up with better code in this case: just initialize a constant
array with offsets of relevant pt_regs fields and we don't need that
switch anymore.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/ia64/kernel/ptrace.c | 57 +++++++++++++++++++----------------------------
 1 file changed, 23 insertions(+), 34 deletions(-)

diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c
index 82aaacf64583..b9d068903b98 100644
--- a/arch/ia64/kernel/ptrace.c
+++ b/arch/ia64/kernel/ptrace.c
@@ -1273,52 +1273,41 @@ struct regset_getset {
 	int ret;
 };
 
+static const ptrdiff_t pt_offsets[16] =
+{
+#define R(n) offsetof(struct pt_regs, r##n)
+	[0] = -1, R(1), R(2), R(3),
+	[4] = -1, [5] = -1, [6] = -1, [7] = -1,
+	R(8), R(9), R(10), R(11), R(12), R(13), R(14), R(15),
+#undef R
+};
+
 static int
 access_elf_gpreg(struct task_struct *target, struct unw_frame_info *info,
 		unsigned long addr, unsigned long *data, int write_access)
 {
-	struct pt_regs *pt;
-	unsigned long *ptr = NULL;
-	int ret;
-	char nat = 0;
+	struct pt_regs *pt = task_pt_regs(target);
+	unsigned reg = addr / sizeof(unsigned long);
+	ptrdiff_t d = pt_offsets[reg];
 
-	pt = task_pt_regs(target);
-	switch (addr) {
-	case ELF_GR_OFFSET(1):
-		ptr = &pt->r1;
-		break;
-	case ELF_GR_OFFSET(2):
-	case ELF_GR_OFFSET(3):
-		ptr = (void *)&pt->r2 + (addr - ELF_GR_OFFSET(2));
-		break;
-	case ELF_GR_OFFSET(4) ... ELF_GR_OFFSET(7):
+	if (d >= 0) {
+		unsigned long *ptr = (void *)pt + d;
+		if (write_access)
+			*ptr = *data;
+		else
+			*data = *ptr;
+		return 0;
+	} else {
+		char nat = 0;
 		if (write_access) {
 			/* read NaT bit first: */
 			unsigned long dummy;
-
-			ret = unw_get_gr(info, addr/8, &dummy, &nat);
+			int ret = unw_get_gr(info, reg, &dummy, &nat);
 			if (ret < 0)
 				return ret;
 		}
-		return unw_access_gr(info, addr/8, data, &nat, write_access);
-	case ELF_GR_OFFSET(8) ... ELF_GR_OFFSET(11):
-		ptr = (void *)&pt->r8 + addr - ELF_GR_OFFSET(8);
-		break;
-	case ELF_GR_OFFSET(12):
-	case ELF_GR_OFFSET(13):
-		ptr = (void *)&pt->r12 + addr - ELF_GR_OFFSET(12);
-		break;
-	case ELF_GR_OFFSET(14):
-		ptr = &pt->r14;
-		break;
-	case ELF_GR_OFFSET(15):
-		ptr = &pt->r15;
+		return unw_access_gr(info, reg, data, &nat, write_access);
 	}
-	if (write_access)
-		*ptr = *data;
-	else
-		*data = *ptr;
-	return 0;
 }
 
 static int
-- 
2.11.0

[PATCH 05/41] [ia64] teach elf_access_reg() to handle the missing range (r16..r31)

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Now it's easy to make elf_access_gpreg() handle the rest of global
registers (r16..r31).  That gets rid of the hole in the registers
elf_access_reg() can handle, which will allow to simplify its callers
later in the series.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/ia64/kernel/ptrace.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c
index b9d068903b98..37b5140d0dfe 100644
--- a/arch/ia64/kernel/ptrace.c
+++ b/arch/ia64/kernel/ptrace.c
@@ -1273,12 +1273,14 @@ struct regset_getset {
 	int ret;
 };
 
-static const ptrdiff_t pt_offsets[16] =
+static const ptrdiff_t pt_offsets[32] =
 {
 #define R(n) offsetof(struct pt_regs, r##n)
 	[0] = -1, R(1), R(2), R(3),
 	[4] = -1, [5] = -1, [6] = -1, [7] = -1,
 	R(8), R(9), R(10), R(11), R(12), R(13), R(14), R(15),
+	R(16), R(17), R(18), R(19), R(20), R(21), R(22), R(23),
+	R(24), R(25), R(26), R(27), R(28), R(29), R(30), R(31),
 #undef R
 };
 
@@ -1479,7 +1481,7 @@ static int
 access_elf_reg(struct task_struct *target, struct unw_frame_info *info,
 		unsigned long addr, unsigned long *data, int write_access)
 {
-	if (addr >= ELF_GR_OFFSET(1) && addr <= ELF_GR_OFFSET(15))
+	if (addr >= ELF_GR_OFFSET(1) && addr <= ELF_GR_OFFSET(31))
 		return access_elf_gpreg(target, info, addr, data, write_access);
 	else if (addr >= ELF_BR_OFFSET(0) && addr <= ELF_BR_OFFSET(7))
 		return access_elf_breg(target, info, addr, data, write_access);
-- 
2.11.0

[PATCH 06/41] [ia64] regularize do_gpregs_[gs]et()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

now access_elf_reg() does the right thing for everything other than
r0, we can simplify do_grepgs_[gs]et()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/ia64/kernel/ptrace.c | 155 ++++++++++------------------------------------
 1 file changed, 31 insertions(+), 124 deletions(-)

diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c
index 37b5140d0dfe..be635f6f93c9 100644
--- a/arch/ia64/kernel/ptrace.c
+++ b/arch/ia64/kernel/ptrace.c
@@ -1491,10 +1491,7 @@ access_elf_reg(struct task_struct *target, struct unw_frame_info *info,
 
 void do_gpregs_get(struct unw_frame_info *info, void *arg)
 {
-	struct pt_regs *pt;
 	struct regset_getset *dst = arg;
-	elf_greg_t tmp[16];
-	unsigned int i, index, min_copy;
 
 	if (unw_unwind_to_user(info) < 0)
 		return;
@@ -1517,160 +1514,70 @@ void do_gpregs_get(struct unw_frame_info *info, void *arg)
 						      &dst->u.get.kbuf,
 						      &dst->u.get.ubuf,
 						      0, ELF_GR_OFFSET(1));
-		if (dst->ret || dst->count == 0)
-			return;
-	}
-
-	/* gr1 - gr15 */
-	if (dst->count > 0 && dst->pos < ELF_GR_OFFSET(16)) {
-		index = (dst->pos - ELF_GR_OFFSET(1)) / sizeof(elf_greg_t);
-		min_copy = ELF_GR_OFFSET(16) > (dst->pos + dst->count) ?
-			 (dst->pos + dst->count) : ELF_GR_OFFSET(16);
-		for (i = dst->pos; i < min_copy; i += sizeof(elf_greg_t),
-				index++)
-			if (access_elf_reg(dst->target, info, i,
-						&tmp[index], 0) < 0) {
-				dst->ret = -EIO;
-				return;
-			}
-		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
-				&dst->u.get.kbuf, &dst->u.get.ubuf, tmp,
-				ELF_GR_OFFSET(1), ELF_GR_OFFSET(16));
-		if (dst->ret || dst->count == 0)
-			return;
-	}
-
-	/* r16-r31 */
-	if (dst->count > 0 && dst->pos < ELF_NAT_OFFSET) {
-		pt = task_pt_regs(dst->target);
-		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
-				&dst->u.get.kbuf, &dst->u.get.ubuf, &pt->r16,
-				ELF_GR_OFFSET(16), ELF_NAT_OFFSET);
-		if (dst->ret || dst->count == 0)
+		if (dst->ret)
 			return;
 	}
 
-	/* nat, pr, b0 - b7 */
-	if (dst->count > 0 && dst->pos < ELF_CR_IIP_OFFSET) {
-		index = (dst->pos - ELF_NAT_OFFSET) / sizeof(elf_greg_t);
-		min_copy = ELF_CR_IIP_OFFSET > (dst->pos + dst->count) ?
-			 (dst->pos + dst->count) : ELF_CR_IIP_OFFSET;
-		for (i = dst->pos; i < min_copy; i += sizeof(elf_greg_t),
-				index++)
-			if (access_elf_reg(dst->target, info, i,
-						&tmp[index], 0) < 0) {
+	while (dst->count && dst->pos < ELF_AR_END_OFFSET) {
+		unsigned int n, from, to;
+		elf_greg_t tmp[16];
+
+		from = dst->pos;
+		to = from + min(dst->count, (unsigned)sizeof(tmp));
+		if (to > ELF_AR_END_OFFSET)
+			to = ELF_AR_END_OFFSET;
+		for (n = 0; from < to; from += sizeof(elf_greg_t), n++) {
+			if (access_elf_reg(dst->target, info, from,
+						&tmp[n], 0) < 0) {
 				dst->ret = -EIO;
 				return;
 			}
+		}
 		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
 				&dst->u.get.kbuf, &dst->u.get.ubuf, tmp,
-				ELF_NAT_OFFSET, ELF_CR_IIP_OFFSET);
-		if (dst->ret || dst->count == 0)
+				dst->pos, to);
+		if (dst->ret)
 			return;
 	}
-
-	/* ip cfm psr ar.rsc ar.bsp ar.bspstore ar.rnat
-	 * ar.ccv ar.unat ar.fpsr ar.pfs ar.lc ar.ec ar.csd ar.ssd
-	 */
-	if (dst->count > 0 && dst->pos < (ELF_AR_END_OFFSET)) {
-		index = (dst->pos - ELF_CR_IIP_OFFSET) / sizeof(elf_greg_t);
-		min_copy = ELF_AR_END_OFFSET > (dst->pos + dst->count) ?
-			 (dst->pos + dst->count) : ELF_AR_END_OFFSET;
-		for (i = dst->pos; i < min_copy; i += sizeof(elf_greg_t),
-				index++)
-			if (access_elf_reg(dst->target, info, i,
-						&tmp[index], 0) < 0) {
-				dst->ret = -EIO;
-				return;
-			}
-		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
-				&dst->u.get.kbuf, &dst->u.get.ubuf, tmp,
-				ELF_CR_IIP_OFFSET, ELF_AR_END_OFFSET);
-	}
 }
 
 void do_gpregs_set(struct unw_frame_info *info, void *arg)
 {
-	struct pt_regs *pt;
 	struct regset_getset *dst = arg;
-	elf_greg_t tmp[16];
-	unsigned int i, index;
 
 	if (unw_unwind_to_user(info) < 0)
 		return;
 
+	if (!dst->count)
+		return;
 	/* Skip r0 */
-	if (dst->count > 0 && dst->pos < ELF_GR_OFFSET(1)) {
+	if (dst->pos < ELF_GR_OFFSET(1)) {
 		dst->ret = user_regset_copyin_ignore(&dst->pos, &dst->count,
 						       &dst->u.set.kbuf,
 						       &dst->u.set.ubuf,
 						       0, ELF_GR_OFFSET(1));
-		if (dst->ret || dst->count == 0)
-			return;
-	}
-
-	/* gr1-gr15 */
-	if (dst->count > 0 && dst->pos < ELF_GR_OFFSET(16)) {
-		i = dst->pos;
-		index = (dst->pos - ELF_GR_OFFSET(1)) / sizeof(elf_greg_t);
-		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
-				&dst->u.set.kbuf, &dst->u.set.ubuf, tmp,
-				ELF_GR_OFFSET(1), ELF_GR_OFFSET(16));
 		if (dst->ret)
 			return;
-		for ( ; i < dst->pos; i += sizeof(elf_greg_t), index++)
-			if (access_elf_reg(dst->target, info, i,
-						&tmp[index], 1) < 0) {
-				dst->ret = -EIO;
-				return;
-			}
-		if (dst->count == 0)
-			return;
-	}
-
-	/* gr16-gr31 */
-	if (dst->count > 0 && dst->pos < ELF_NAT_OFFSET) {
-		pt = task_pt_regs(dst->target);
-		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
-				&dst->u.set.kbuf, &dst->u.set.ubuf, &pt->r16,
-				ELF_GR_OFFSET(16), ELF_NAT_OFFSET);
-		if (dst->ret || dst->count == 0)
-			return;
 	}
 
-	/* nat, pr, b0 - b7 */
-	if (dst->count > 0 && dst->pos < ELF_CR_IIP_OFFSET) {
-		i = dst->pos;
-		index = (dst->pos - ELF_NAT_OFFSET) / sizeof(elf_greg_t);
-		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
-				&dst->u.set.kbuf, &dst->u.set.ubuf, tmp,
-				ELF_NAT_OFFSET, ELF_CR_IIP_OFFSET);
-		if (dst->ret)
-			return;
-		for (; i < dst->pos; i += sizeof(elf_greg_t), index++)
-			if (access_elf_reg(dst->target, info, i,
-						&tmp[index], 1) < 0) {
-				dst->ret = -EIO;
-				return;
-			}
-		if (dst->count == 0)
-			return;
-	}
+	while (dst->count && dst->pos < ELF_AR_END_OFFSET) {
+		unsigned int n, from, to;
+		elf_greg_t tmp[16];
 
-	/* ip cfm psr ar.rsc ar.bsp ar.bspstore ar.rnat
-	 * ar.ccv ar.unat ar.fpsr ar.pfs ar.lc ar.ec ar.csd ar.ssd
-	 */
-	if (dst->count > 0 && dst->pos < (ELF_AR_END_OFFSET)) {
-		i = dst->pos;
-		index = (dst->pos - ELF_CR_IIP_OFFSET) / sizeof(elf_greg_t);
+		from = dst->pos;
+		to = from + sizeof(tmp);
+		if (to > ELF_AR_END_OFFSET)
+			to = ELF_AR_END_OFFSET;
+		/* get up to 16 values */
 		dst->ret = user_regset_copyin(&dst->pos, &dst->count,
 				&dst->u.set.kbuf, &dst->u.set.ubuf, tmp,
-				ELF_CR_IIP_OFFSET, ELF_AR_END_OFFSET);
+				from, to);
 		if (dst->ret)
 			return;
-		for ( ; i < dst->pos; i += sizeof(elf_greg_t), index++)
-			if (access_elf_reg(dst->target, info, i,
-						&tmp[index], 1) < 0) {
+		/* now copy them into registers */
+		for (n = 0; from < dst->pos; from += sizeof(elf_greg_t), n++)
+			if (access_elf_reg(dst->target, info, from,
+						&tmp[n], 1) < 0) {
 				dst->ret = -EIO;
 				return;
 			}
-- 
2.11.0

[PATCH 07/41] [ia64] access_uarea(): stop bothering with gpregs_[gs]et()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

We know this won't be called for child == current, so we don't need
to bother with callbacks, etc. - just do unw_init_from_blocked_task(),
unw_unwind_to_user() and do the payload of gpregs_[gs]et().  For
one register.  Which is to say, access_elf_reg().

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/ia64/kernel/ptrace.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c
index be635f6f93c9..e0813c8e4b47 100644
--- a/arch/ia64/kernel/ptrace.c
+++ b/arch/ia64/kernel/ptrace.c
@@ -1936,15 +1936,14 @@ access_uarea(struct task_struct *child, unsigned long addr,
 	}
 
 	if (pos != -1) {
-		if (write_access)
-			ret = gpregs_set(child, NULL, pos,
-				sizeof(unsigned long), data, NULL);
-		else
-			ret = gpregs_get(child, NULL, pos,
-				sizeof(unsigned long), data, NULL);
-		if (ret != 0)
-			return -1;
-		return 0;
+		struct unw_frame_info info;
+
+		memset(&info, 0, sizeof(info));
+		unw_init_from_blocked_task(&info, child);
+		if (unw_unwind_to_user(&info) < 0)
+			return 0;
+
+		return access_elf_reg(child, &info, pos, data, write_access);
 	}
 
 	/* access debug registers */
-- 
2.11.0

[PATCH 08/41] [ia64] access_uarea(): don't bother with fpregs_[gs]et()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

similar to previous commit...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/ia64/kernel/ptrace.c | 42 +++++++++++++++++++++++++++++++++---------
 1 file changed, 33 insertions(+), 9 deletions(-)

diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c
index e0813c8e4b47..eae7b094f608 100644
--- a/arch/ia64/kernel/ptrace.c
+++ b/arch/ia64/kernel/ptrace.c
@@ -1811,7 +1811,6 @@ access_uarea(struct task_struct *child, unsigned long addr,
 	      unsigned long *data, int write_access)
 {
 	unsigned int pos = -1; /* an invalid value */
-	int ret;
 	unsigned long *ptr, regnum;
 
 	if ((addr & 0x7) != 0) {
@@ -1843,14 +1842,39 @@ access_uarea(struct task_struct *child, unsigned long addr,
 	}
 
 	if (pos != -1) {
-		if (write_access)
-			ret = fpregs_set(child, NULL, pos,
-				sizeof(unsigned long), data, NULL);
-		else
-			ret = fpregs_get(child, NULL, pos,
-				sizeof(unsigned long), data, NULL);
-		if (ret != 0)
-			return -1;
+		unsigned reg = pos / sizeof(elf_fpreg_t);
+		int which_half = (pos / sizeof(unsigned long)) & 1;
+
+		if (reg < 32) { /* fr2-fr31 */
+			struct unw_frame_info info;
+			elf_fpreg_t fpreg;
+
+			memset(&info, 0, sizeof(info));
+			unw_init_from_blocked_task(&info, child);
+			if (unw_unwind_to_user(&info) < 0)
+				return 0;
+
+			if (unw_get_fr(&info, reg, &fpreg))
+				return -1;
+			if (write_access) {
+				fpreg.u.bits[which_half] = *data;
+				if (unw_set_fr(&info, reg, fpreg))
+					return -1;
+			} else {
+				*data = fpreg.u.bits[which_half];
+			}
+		} else { /* fph */
+			elf_fpreg_t *p = &child->thread.fph[reg - 32];
+			unsigned long *bits = &p->u.bits[which_half];
+
+			ia64_sync_fph(child);
+			if (write_access)
+				*bits = *data;
+			else if (child->thread.flags & IA64_THREAD_FPH_VALID)
+				*data = *bits;
+			else
+				*data = 0;
+		}
 		return 0;
 	}
 
-- 
2.11.0

[PATCH 09/41] sparc64: switch genregs32_get() to use of get_from_target()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

... for fetching the register window from target's stack, rather
than open-coding it.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/sparc/kernel/ptrace_64.c | 59 ++++++++++++-------------------------------
 1 file changed, 16 insertions(+), 43 deletions(-)

diff --git a/arch/sparc/kernel/ptrace_64.c b/arch/sparc/kernel/ptrace_64.c
index 7122efb4b1cc..f7b2ddfc81d6 100644
--- a/arch/sparc/kernel/ptrace_64.c
+++ b/arch/sparc/kernel/ptrace_64.c
@@ -518,10 +518,10 @@ static int genregs32_get(struct task_struct *target,
 			 void *kbuf, void __user *ubuf)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	compat_ulong_t __user *reg_window;
 	compat_ulong_t *k = kbuf;
 	compat_ulong_t __user *u = ubuf;
-	compat_ulong_t reg;
+	u32 uregs[16];
+	u32 reg;
 
 	if (target == current)
 		flushw_user();
@@ -533,52 +533,25 @@ static int genregs32_get(struct task_struct *target,
 		for (; count > 0 && pos < 16; count--)
 			*k++ = regs->u_regs[pos++];
 
-		reg_window = (compat_ulong_t __user *) regs->u_regs[UREG_I6];
-		reg_window -= 16;
-		if (target == current) {
-			for (; count > 0 && pos < 32; count--) {
-				if (get_user(*k++, &reg_window[pos++]))
-					return -EFAULT;
-			}
-		} else {
-			for (; count > 0 && pos < 32; count--) {
-				if (access_process_vm(target,
-						      (unsigned long)
-						      &reg_window[pos],
-						      k, sizeof(*k),
-						      FOLL_FORCE)
-				    != sizeof(*k))
-					return -EFAULT;
-				k++;
-				pos++;
-			}
+		if (count && pos < 32) {
+			if (get_from_target(target, regs->u_regs[UREG_I6],
+					uregs, sizeof(uregs)))
+				return -EFAULT;
+			for (; count > 0 && pos < 32; count--)
+				*k++ = uregs[pos++ - 16];
+
 		}
 	} else {
-		for (; count > 0 && pos < 16; count--) {
+		for (; count > 0 && pos < 16; count--)
 			if (put_user((compat_ulong_t) regs->u_regs[pos++], u++))
 				return -EFAULT;
-		}
-
-		reg_window = (compat_ulong_t __user *) regs->u_regs[UREG_I6];
-		reg_window -= 16;
-		if (target == current) {
-			for (; count > 0 && pos < 32; count--) {
-				if (get_user(reg, &reg_window[pos++]) ||
-				    put_user(reg, u++))
-					return -EFAULT;
-			}
-		} else {
-			for (; count > 0 && pos < 32; count--) {
-				if (access_process_vm(target,
-						      (unsigned long)
-						      &reg_window[pos++],
-						      &reg, sizeof(reg),
-						      FOLL_FORCE)
-				    != sizeof(reg))
-					return -EFAULT;
-				if (put_user(reg, u++))
+		if (count && pos < 32) {
+			if (get_from_target(target, regs->u_regs[UREG_I6],
+					uregs, sizeof(uregs)))
+				return -EFAULT;
+			for (; count > 0 && pos < 32; count--)
+				if (put_user(uregs[pos++ - 16], u++))
 					return -EFAULT;
-			}
 		}
 	}
 	while (count > 0) {
-- 
2.11.0

[PATCH 10/41] sparc32: get rid of odd callers of copy_regset_to_user()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

the life is much simpler if copy_regset_to_user() (and ->get())
gets called only with pos == 0; sparc32 PTRACE_GETREGS and
PTRACE_GETFPREGS are among the few things that use it to fetch
pieces of regset _not_ starting at the beginning.  It's actually
easier to define a separate regset that would provide what
we need, rather than trying to cobble that from the one
PTRACE_GETREGSET uses.

Extra ->get() instances do not amount to much code and once
we get the conversion of ->get() to new API (dependent upon the
lack of weird callers of ->get()) they'll shrink a lot, along
with the rest of ->get() instances...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/sparc/kernel/ptrace_32.c | 117 +++++++++++++++++++++++++++++-------------
 1 file changed, 82 insertions(+), 35 deletions(-)

diff --git a/arch/sparc/kernel/ptrace_32.c b/arch/sparc/kernel/ptrace_32.c
index 47eb315d411c..f72b7d2c4716 100644
--- a/arch/sparc/kernel/ptrace_32.c
+++ b/arch/sparc/kernel/ptrace_32.c
@@ -99,15 +99,13 @@ static int genregs32_get(struct task_struct *target,
 	if (ret || !count)
 		return ret;
 
-	if (pos < 32 * sizeof(u32)) {
-		if (regwindow32_get(target, regs, uregs))
-			return -EFAULT;
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  uregs,
-					  16 * sizeof(u32), 32 * sizeof(u32));
-		if (ret || !count)
-			return ret;
-	}
+	if (regwindow32_get(target, regs, uregs))
+		return -EFAULT;
+	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				  uregs,
+				  16 * sizeof(u32), 32 * sizeof(u32));
+	if (ret)
+		return ret;
 
 	uregs[0] = regs->psr;
 	uregs[1] = regs->pc;
@@ -288,6 +286,73 @@ static const struct user_regset sparc32_regsets[] = {
 	},
 };
 
+static int getregs_get(struct task_struct *target,
+			 const struct user_regset *regset,
+			 unsigned int pos, unsigned int count,
+			 void *kbuf, void __user *ubuf)
+{
+	const struct pt_regs *regs = target->thread.kregs;
+	u32 v[4];
+	int ret;
+
+	if (target == current)
+		flush_user_windows();
+
+	v[0] = regs->psr;
+	v[1] = regs->pc;
+	v[2] = regs->npc;
+	v[3] = regs->y;
+	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				  v,
+				  0 * sizeof(u32), 4 * sizeof(u32));
+	if (ret)
+		return ret;
+
+	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				  regs->u_regs + 1,
+				  4 * sizeof(u32), 19 * sizeof(u32));
+}
+
+static int getfpregs_get(struct task_struct *target,
+			const struct user_regset *regset,
+			unsigned int pos, unsigned int count,
+			void *kbuf, void __user *ubuf)
+{
+	const unsigned long *fpregs = target->thread.float_regs;
+	int ret = 0;
+
+#if 0
+	if (target == current)
+		save_and_clear_fpu();
+#endif
+
+	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				  fpregs,
+				  0, 32 * sizeof(u32));
+	if (ret)
+		return ret;
+	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				  &target->thread.fsr,
+				  32 * sizeof(u32), 33 * sizeof(u32));
+	if (ret)
+		return ret;
+	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
+				  33 * sizeof(u32), 68 * sizeof(u32));
+}
+
+static const struct user_regset ptrace32_regsets[] = {
+	[REGSET_GENERAL] = {
+		.n = 19, .size = sizeof(u32), .get = getregs_get,
+	},
+	[REGSET_FP] = {
+		.n = 68, .size = sizeof(u32), .get = getfpregs_get,
+	},
+};
+
+static const struct user_regset_view ptrace32_view = {
+	.regsets = ptrace32_regsets, .n = ARRAY_SIZE(ptrace32_regsets)
+};
+
 static const struct user_regset_view user_sparc32_view = {
 	.name = "sparc", .e_machine = EM_SPARC,
 	.regsets = sparc32_regsets, .n = ARRAY_SIZE(sparc32_regsets)
@@ -327,15 +392,10 @@ long arch_ptrace(struct task_struct *child, long request,
 
 	switch(request) {
 	case PTRACE_GETREGS: {
-		ret = copy_regset_to_user(child, view, REGSET_GENERAL,
-					  32 * sizeof(u32),
-					  4 * sizeof(u32),
-					  &pregs->psr);
-		if (!ret)
-			copy_regset_to_user(child, view, REGSET_GENERAL,
-					    1 * sizeof(u32),
-					    15 * sizeof(u32),
-					    &pregs->u_regs[0]);
+		ret = copy_regset_to_user(child, &ptrace32_view,
+					  REGSET_GENERAL, 0,
+					  19 * sizeof(u32),
+					  pregs);
 		break;
 	}
 
@@ -353,23 +413,10 @@ long arch_ptrace(struct task_struct *child, long request,
 	}
 
 	case PTRACE_GETFPREGS: {
-		ret = copy_regset_to_user(child, view, REGSET_FP,
-					  0 * sizeof(u32),
-					  32 * sizeof(u32),
-					  &fps->regs[0]);
-		if (!ret)
-			ret = copy_regset_to_user(child, view, REGSET_FP,
-						  33 * sizeof(u32),
-						  1 * sizeof(u32),
-						  &fps->fsr);
-
-		if (!ret) {
-			if (__put_user(0, &fps->fpqd) ||
-			    __put_user(0, &fps->flags) ||
-			    __put_user(0, &fps->extra) ||
-			    clear_user(fps->fpq, sizeof(fps->fpq)))
-				ret = -EFAULT;
-		}
+		ret = copy_regset_to_user(child, &ptrace32_view,
+					  REGSET_FP, 0,
+					  68 * sizeof(u32),
+					  fps);
 		break;
 	}
 
-- 
2.11.0

[PATCH 11/41] sparc64: get rid of odd callers of copy_regset_to_user()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

same as for sparc32, and that's it - no more caller of ->get() with
non-zero pos.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/sparc/kernel/ptrace_64.c | 175 ++++++++++++++++++++++++++++++++----------
 1 file changed, 136 insertions(+), 39 deletions(-)

diff --git a/arch/sparc/kernel/ptrace_64.c b/arch/sparc/kernel/ptrace_64.c
index f7b2ddfc81d6..1b1910b67ca4 100644
--- a/arch/sparc/kernel/ptrace_64.c
+++ b/arch/sparc/kernel/ptrace_64.c
@@ -258,7 +258,7 @@ static int genregs64_get(struct task_struct *target,
 	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
 				  regs->u_regs,
 				  0, 16 * sizeof(u64));
-	if (!ret && count && pos < (32 * sizeof(u64))) {
+	if (!ret && count) {
 		struct reg_window window;
 
 		if (regwindow64_get(target, regs, &window))
@@ -506,6 +506,57 @@ static const struct user_regset sparc64_regsets[] = {
 	},
 };
 
+static int getregs64_get(struct task_struct *target,
+			 const struct user_regset *regset,
+			 unsigned int pos, unsigned int count,
+			 void *kbuf, void __user *ubuf)
+{
+	const struct pt_regs *regs = task_pt_regs(target);
+	int ret;
+
+	if (target == current)
+		flushw_user();
+
+	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				  regs->u_regs + 1,
+				  0, 15 * sizeof(u64));
+	if (!ret)
+		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
+				  15 * sizeof(u64), 16 * sizeof(u64));
+	if (!ret) {
+		/* TSTATE, TPC, TNPC */
+		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+					  &regs->tstate,
+					  16 * sizeof(u64),
+					  19 * sizeof(u64));
+	}
+	if (!ret) {
+		unsigned long y = regs->y;
+
+		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+					  &y,
+					  19 * sizeof(u64),
+					  20 * sizeof(u64));
+	}
+	return ret;
+}
+
+static const struct user_regset ptrace64_regsets[] = {
+	/* Format is:
+	 *      G1 --> G7
+	 *      O0 --> O7
+	 *	0
+	 *      TSTATE, TPC, TNPC, Y
+	 */
+	[REGSET_GENERAL] = {
+		.n = 20, .size = sizeof(u64), .get = getregs64_get,
+	},
+};
+
+static const struct user_regset_view ptrace64_view = {
+	.regsets = ptrace64_regsets, .n = ARRAY_SIZE(ptrace64_regsets)
+};
+
 static const struct user_regset_view user_sparc64_view = {
 	.name = "sparc64", .e_machine = EM_SPARCV9,
 	.regsets = sparc64_regsets, .n = ARRAY_SIZE(sparc64_regsets)
@@ -533,7 +584,7 @@ static int genregs32_get(struct task_struct *target,
 		for (; count > 0 && pos < 16; count--)
 			*k++ = regs->u_regs[pos++];
 
-		if (count && pos < 32) {
+		if (count) {
 			if (get_from_target(target, regs->u_regs[UREG_I6],
 					uregs, sizeof(uregs)))
 				return -EFAULT;
@@ -545,7 +596,7 @@ static int genregs32_get(struct task_struct *target,
 		for (; count > 0 && pos < 16; count--)
 			if (put_user((compat_ulong_t) regs->u_regs[pos++], u++))
 				return -EFAULT;
-		if (count && pos < 32) {
+		if (count) {
 			if (get_from_target(target, regs->u_regs[UREG_I6],
 					uregs, sizeof(uregs)))
 				return -EFAULT;
@@ -840,6 +891,76 @@ static const struct user_regset sparc32_regsets[] = {
 	},
 };
 
+static int getregs_get(struct task_struct *target,
+		       const struct user_regset *regset,
+		       unsigned int pos, unsigned int count,
+		       void *kbuf, void __user *ubuf)
+{
+	const struct pt_regs *regs = task_pt_regs(target);
+	u32 uregs[19];
+	int i;
+
+	if (target == current)
+		flushw_user();
+
+	uregs[0] = tstate_to_psr(regs->tstate);
+	uregs[1] = regs->tpc;
+	uregs[2] = regs->tnpc;
+	uregs[3] = regs->y;
+	for (i = 1; i < 16; i++)
+		uregs[3 + i] = regs->u_regs[i];
+	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				   uregs,
+				   0, 19 * sizeof(u32));
+}
+
+static int getfpregs_get(struct task_struct *target,
+			const struct user_regset *regset,
+			unsigned int pos, unsigned int count,
+			void *kbuf, void __user *ubuf)
+{
+	const unsigned long *fpregs = task_thread_info(target)->fpregs;
+	unsigned long fprs;
+	compat_ulong_t fsr;
+	int ret = 0;
+
+	if (target == current)
+		save_and_clear_fpu();
+
+	fprs = task_thread_info(target)->fpsaved[0];
+	if (fprs & FPRS_FEF) {
+		fsr = task_thread_info(target)->xfsr[0];
+	} else {
+		fsr = 0;
+	}
+
+	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+				  fpregs,
+				  0, 32 * sizeof(u32));
+	if (!ret)
+		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
+					  &fsr,
+					  32 * sizeof(u32),
+					  33 * sizeof(u32));
+	if (!ret)
+		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
+					  33 * sizeof(u32), 68 * sizeof(u32));
+	return ret;
+}
+
+static const struct user_regset ptrace32_regsets[] = {
+	[REGSET_GENERAL] = {
+		.n = 19, .size = sizeof(u32), .get = getregs_get,
+	},
+	[REGSET_FP] = {
+		.n = 68, .size = sizeof(u32), .get = getfpregs_get,
+	},
+};
+
+static const struct user_regset_view ptrace32_view = {
+	.regsets = ptrace32_regsets, .n = ARRAY_SIZE(ptrace32_regsets)
+};
+
 static const struct user_regset_view user_sparc32_view = {
 	.name = "sparc", .e_machine = EM_SPARC,
 	.regsets = sparc32_regsets, .n = ARRAY_SIZE(sparc32_regsets)
@@ -889,15 +1010,10 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
 		break;
 
 	case PTRACE_GETREGS:
-		ret = copy_regset_to_user(child, view, REGSET_GENERAL,
-					  32 * sizeof(u32),
-					  4 * sizeof(u32),
-					  &pregs->psr);
-		if (!ret)
-			ret = copy_regset_to_user(child, view, REGSET_GENERAL,
-						  1 * sizeof(u32),
-						  15 * sizeof(u32),
-						  &pregs->u_regs[0]);
+		ret = copy_regset_to_user(child, &ptrace32_view,
+					  REGSET_GENERAL, 0,
+					  19 * sizeof(u32),
+					  pregs);
 		break;
 
 	case PTRACE_SETREGS:
@@ -913,22 +1029,10 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
 		break;
 
 	case PTRACE_GETFPREGS:
-		ret = copy_regset_to_user(child, view, REGSET_FP,
-					  0 * sizeof(u32),
-					  32 * sizeof(u32),
-					  &fps->regs[0]);
-		if (!ret)
-			ret = copy_regset_to_user(child, view, REGSET_FP,
-						  33 * sizeof(u32),
-						  1 * sizeof(u32),
-						  &fps->fsr);
-		if (!ret) {
-			if (__put_user(0, &fps->flags) ||
-			    __put_user(0, &fps->extra) ||
-			    __put_user(0, &fps->fpqd) ||
-			    clear_user(&fps->fpq[0], 32 * sizeof(unsigned int)))
-				ret = -EFAULT;
-		}
+		ret = copy_regset_to_user(child, &ptrace32_view,
+					  REGSET_FP, 0,
+					  68 * sizeof(u32),
+					  fps);
 		break;
 
 	case PTRACE_SETFPREGS:
@@ -999,17 +1103,10 @@ long arch_ptrace(struct task_struct *child, long request,
 		break;
 
 	case PTRACE_GETREGS64:
-		ret = copy_regset_to_user(child, view, REGSET_GENERAL,
-					  1 * sizeof(u64),
-					  15 * sizeof(u64),
-					  &pregs->u_regs[0]);
-		if (!ret) {
-			/* XXX doesn't handle 'y' register correctly XXX */
-			ret = copy_regset_to_user(child, view, REGSET_GENERAL,
-						  32 * sizeof(u64),
-						  4 * sizeof(u64),
-						  &pregs->tstate);
-		}
+		ret = copy_regset_to_user(child, &ptrace64_view,
+					  REGSET_GENERAL, 0,
+					  19 * sizeof(u64),
+					  pregs);
 		break;
 
 	case PTRACE_SETREGS64:
-- 
2.11.0

[PATCH 12/41] sparc32: get rid of odd callers of copy_regset_from_user()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/sparc/kernel/ptrace_32.c | 109 ++++++++++++++++++++++++++++--------------
 1 file changed, 73 insertions(+), 36 deletions(-)

diff --git a/arch/sparc/kernel/ptrace_32.c b/arch/sparc/kernel/ptrace_32.c
index f72b7d2c4716..0856e0104539 100644
--- a/arch/sparc/kernel/ptrace_32.c
+++ b/arch/sparc/kernel/ptrace_32.c
@@ -137,19 +137,18 @@ static int genregs32_set(struct task_struct *target,
 	if (ret || !count)
 		return ret;
 
-	if (pos < 32 * sizeof(u32)) {
-		if (regwindow32_get(target, regs, uregs))
-			return -EFAULT;
-		ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
-					 uregs,
-					 16 * sizeof(u32), 32 * sizeof(u32));
-		if (ret)
-			return ret;
-		if (regwindow32_set(target, regs, uregs))
-			return -EFAULT;
-		if (!count)
-			return 0;
-	}
+	if (regwindow32_get(target, regs, uregs))
+		return -EFAULT;
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 uregs,
+				 16 * sizeof(u32), 32 * sizeof(u32));
+	if (ret)
+		return ret;
+	if (regwindow32_set(target, regs, uregs))
+		return -EFAULT;
+	if (!count)
+		return 0;
+
 	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
 				 &psr,
 				 32 * sizeof(u32), 33 * sizeof(u32));
@@ -241,13 +240,11 @@ static int fpregs32_set(struct task_struct *target,
 		user_regset_copyin_ignore(&pos, &count, &kbuf, &ubuf,
 					  32 * sizeof(u32),
 					  33 * sizeof(u32));
-	if (!ret && count > 0) {
+	if (!ret)
 		ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
 					 &target->thread.fsr,
 					 33 * sizeof(u32),
 					 34 * sizeof(u32));
-	}
-
 	if (!ret)
 		ret = user_regset_copyin_ignore(&pos, &count, &kbuf, &ubuf,
 						34 * sizeof(u32), -1);
@@ -313,6 +310,31 @@ static int getregs_get(struct task_struct *target,
 				  4 * sizeof(u32), 19 * sizeof(u32));
 }
 
+static int setregs_set(struct task_struct *target,
+			 const struct user_regset *regset,
+			 unsigned int pos, unsigned int count,
+			 const void *kbuf, const void __user *ubuf)
+{
+	struct pt_regs *regs = target->thread.kregs;
+	u32 v[4];
+	int ret;
+
+	if (target == current)
+		flush_user_windows();
+
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 v,
+				 0, 4 * sizeof(u32));
+	regs->psr = (regs->psr & ~(PSR_ICC | PSR_SYSCALL)) |
+		    (v[0] & (PSR_ICC | PSR_SYSCALL));
+	regs->pc = v[1];
+	regs->npc = v[2];
+	regs->y = v[3];
+	return user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 regs->u_regs + 1,
+				 4 * sizeof(u32) , 19 * sizeof(u32));
+}
+
 static int getfpregs_get(struct task_struct *target,
 			const struct user_regset *regset,
 			unsigned int pos, unsigned int count,
@@ -340,12 +362,37 @@ static int getfpregs_get(struct task_struct *target,
 				  33 * sizeof(u32), 68 * sizeof(u32));
 }
 
+static int setfpregs_set(struct task_struct *target,
+			const struct user_regset *regset,
+			unsigned int pos, unsigned int count,
+			const void *kbuf, const void __user *ubuf)
+{
+	unsigned long *fpregs = target->thread.float_regs;
+	int ret;
+
+#if 0
+	if (target == current)
+		save_and_clear_fpu();
+#endif
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 fpregs,
+				 0, 32 * sizeof(u32));
+	if (ret)
+		return ret;
+	return user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 &target->thread.fsr,
+				 32 * sizeof(u32),
+				 33 * sizeof(u32));
+}
+
 static const struct user_regset ptrace32_regsets[] = {
 	[REGSET_GENERAL] = {
-		.n = 19, .size = sizeof(u32), .get = getregs_get,
+		.n = 19, .size = sizeof(u32),
+		.get = getregs_get, .set = setregs_set,
 	},
 	[REGSET_FP] = {
-		.n = 68, .size = sizeof(u32), .get = getfpregs_get,
+		.n = 68, .size = sizeof(u32),
+		.get = getfpregs_get, .set = setfpregs_set,
 	},
 };
 
@@ -400,15 +447,10 @@ long arch_ptrace(struct task_struct *child, long request,
 	}
 
 	case PTRACE_SETREGS: {
-		ret = copy_regset_from_user(child, view, REGSET_GENERAL,
-					    32 * sizeof(u32),
-					    4 * sizeof(u32),
-					    &pregs->psr);
-		if (!ret)
-			copy_regset_from_user(child, view, REGSET_GENERAL,
-					      1 * sizeof(u32),
-					      15 * sizeof(u32),
-					      &pregs->u_regs[0]);
+		ret = copy_regset_from_user(child, &ptrace32_view,
+					    REGSET_GENERAL, 0,
+					    19 * sizeof(u32),
+					    pregs);
 		break;
 	}
 
@@ -421,15 +463,10 @@ long arch_ptrace(struct task_struct *child, long request,
 	}
 
 	case PTRACE_SETFPREGS: {
-		ret = copy_regset_from_user(child, view, REGSET_FP,
-					    0 * sizeof(u32),
-					    32 * sizeof(u32),
-					    &fps->regs[0]);
-		if (!ret)
-			ret = copy_regset_from_user(child, view, REGSET_FP,
-						    33 * sizeof(u32),
-						    1 * sizeof(u32),
-						    &fps->fsr);
+		ret = copy_regset_from_user(child, &ptrace32_view,
+					  REGSET_FP, 0,
+					  33 * sizeof(u32),
+					  fps);
 		break;
 	}
 
-- 
2.11.0

[PATCH 13/41] sparc64: get rid of odd callers of copy_regset_from_user()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/sparc/kernel/ptrace_64.c | 177 ++++++++++++++++++++++++++++++++++--------
 1 file changed, 144 insertions(+), 33 deletions(-)

diff --git a/arch/sparc/kernel/ptrace_64.c b/arch/sparc/kernel/ptrace_64.c
index 1b1910b67ca4..3c9eee12102a 100644
--- a/arch/sparc/kernel/ptrace_64.c
+++ b/arch/sparc/kernel/ptrace_64.c
@@ -541,6 +541,60 @@ static int getregs64_get(struct task_struct *target,
 	return ret;
 }
 
+static int setregs64_set(struct task_struct *target,
+			 const struct user_regset *regset,
+			 unsigned int pos, unsigned int count,
+			 const void *kbuf, const void __user *ubuf)
+{
+	struct pt_regs *regs = task_pt_regs(target);
+	unsigned long y = regs->y;
+	unsigned long tstate;
+	int ret;
+
+	if (target == current)
+		flushw_user();
+
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 regs->u_regs + 1,
+				 0 * sizeof(u64),
+				 15 * sizeof(u64));
+	if (ret)
+		return ret;
+	ret =user_regset_copyin_ignore(&pos, &count, &kbuf, &ubuf,
+				 15 * sizeof(u64), 16 * sizeof(u64));
+	if (ret)
+		return ret;
+	/* TSTATE */
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 &tstate,
+				 16 * sizeof(u64),
+				 17 * sizeof(u64));
+	if (ret)
+		return ret;
+	/* Only the condition codes and the "in syscall"
+	 * state can be modified in the %tstate register.
+	 */
+	tstate &= (TSTATE_ICC | TSTATE_XCC | TSTATE_SYSCALL);
+	regs->tstate &= ~(TSTATE_ICC | TSTATE_XCC | TSTATE_SYSCALL);
+	regs->tstate |= tstate;
+
+	/* TPC, TNPC */
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 &regs->tpc,
+				 17 * sizeof(u64),
+				 19 * sizeof(u64));
+	if (ret)
+		return ret;
+	/* Y */
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 &y,
+				 19 * sizeof(u64),
+				 20 * sizeof(u64));
+	if (!ret)
+		regs->y = y;
+	return ret;
+}
+
 static const struct user_regset ptrace64_regsets[] = {
 	/* Format is:
 	 *      G1 --> G7
@@ -549,7 +603,8 @@ static const struct user_regset ptrace64_regsets[] = {
 	 *      TSTATE, TPC, TNPC, Y
 	 */
 	[REGSET_GENERAL] = {
-		.n = 20, .size = sizeof(u64), .get = getregs64_get,
+		.n = 20, .size = sizeof(u64),
+		.get = getregs64_get, .set = setregs64_set,
 	},
 };
 
@@ -914,6 +969,40 @@ static int getregs_get(struct task_struct *target,
 				   0, 19 * sizeof(u32));
 }
 
+static int setregs_set(struct task_struct *target,
+			 const struct user_regset *regset,
+			 unsigned int pos, unsigned int count,
+			 const void *kbuf, const void __user *ubuf)
+{
+	struct pt_regs *regs = task_pt_regs(target);
+	unsigned long tstate;
+	u32 uregs[19];
+	int i, ret;
+
+	if (target == current)
+		flushw_user();
+
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 uregs,
+				 0, 19 * sizeof(u32));
+	if (ret)
+		return ret;
+
+	tstate = regs->tstate;
+	tstate &= ~(TSTATE_ICC | TSTATE_XCC | TSTATE_SYSCALL);
+	tstate |= psr_to_tstate_icc(uregs[0]);
+	if (uregs[0] & PSR_SYSCALL)
+		tstate |= TSTATE_SYSCALL;
+	regs->tstate = tstate;
+	regs->tpc = uregs[1];
+	regs->tnpc = uregs[2];
+	regs->y = uregs[3];
+
+	for (i = 1; i < 15; i++)
+		regs->u_regs[i] = uregs[3 + i];
+	return 0;
+}
+
 static int getfpregs_get(struct task_struct *target,
 			const struct user_regset *regset,
 			unsigned int pos, unsigned int count,
@@ -948,12 +1037,52 @@ static int getfpregs_get(struct task_struct *target,
 	return ret;
 }
 
+static int setfpregs_set(struct task_struct *target,
+			const struct user_regset *regset,
+			unsigned int pos, unsigned int count,
+			const void *kbuf, const void __user *ubuf)
+{
+	unsigned long *fpregs = task_thread_info(target)->fpregs;
+	unsigned long fprs;
+	int ret;
+
+	if (target == current)
+		save_and_clear_fpu();
+
+	fprs = task_thread_info(target)->fpsaved[0];
+
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+				 fpregs,
+				 0, 32 * sizeof(u32));
+	if (!ret) {
+		compat_ulong_t fsr;
+		unsigned long val;
+
+		ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf,
+					 &fsr,
+					 32 * sizeof(u32),
+					 33 * sizeof(u32));
+		if (!ret) {
+			val = task_thread_info(target)->xfsr[0];
+			val &= 0xffffffff00000000UL;
+			val |= fsr;
+			task_thread_info(target)->xfsr[0] = val;
+		}
+	}
+
+	fprs |= (FPRS_FEF | FPRS_DL);
+	task_thread_info(target)->fpsaved[0] = fprs;
+	return ret;
+}
+
 static const struct user_regset ptrace32_regsets[] = {
 	[REGSET_GENERAL] = {
-		.n = 19, .size = sizeof(u32), .get = getregs_get,
+		.n = 19, .size = sizeof(u32),
+		.get = getregs_get, .set = setregs_set,
 	},
 	[REGSET_FP] = {
-		.n = 68, .size = sizeof(u32), .get = getfpregs_get,
+		.n = 68, .size = sizeof(u32),
+		.get = getfpregs_get, .set = setfpregs_set,
 	},
 };
 
@@ -992,7 +1121,6 @@ struct compat_fps {
 long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
 			compat_ulong_t caddr, compat_ulong_t cdata)
 {
-	const struct user_regset_view *view = task_user_regset_view(current);
 	compat_ulong_t caddr2 = task_pt_regs(current)->u_regs[UREG_I4];
 	struct pt_regs32 __user *pregs;
 	struct compat_fps __user *fps;
@@ -1017,15 +1145,10 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
 		break;
 
 	case PTRACE_SETREGS:
-		ret = copy_regset_from_user(child, view, REGSET_GENERAL,
-					    32 * sizeof(u32),
-					    4 * sizeof(u32),
-					    &pregs->psr);
-		if (!ret)
-			ret = copy_regset_from_user(child, view, REGSET_GENERAL,
-						    1 * sizeof(u32),
-						    15 * sizeof(u32),
-						    &pregs->u_regs[0]);
+		ret = copy_regset_from_user(child, &ptrace32_view,
+					  REGSET_GENERAL, 0,
+					  19 * sizeof(u32),
+					  pregs);
 		break;
 
 	case PTRACE_GETFPREGS:
@@ -1036,15 +1159,10 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
 		break;
 
 	case PTRACE_SETFPREGS:
-		ret = copy_regset_from_user(child, view, REGSET_FP,
-					    0 * sizeof(u32),
-					    32 * sizeof(u32),
-					    &fps->regs[0]);
-		if (!ret)
-			ret = copy_regset_from_user(child, view, REGSET_FP,
-						    33 * sizeof(u32),
-						    1 * sizeof(u32),
-						    &fps->fsr);
+		ret = copy_regset_from_user(child, &ptrace32_view,
+					  REGSET_FP, 0,
+					  33 * sizeof(u32),
+					  fps);
 		break;
 
 	case PTRACE_READTEXT:
@@ -1110,17 +1228,10 @@ long arch_ptrace(struct task_struct *child, long request,
 		break;
 
 	case PTRACE_SETREGS64:
-		ret = copy_regset_from_user(child, view, REGSET_GENERAL,
-					    1 * sizeof(u64),
-					    15 * sizeof(u64),
-					    &pregs->u_regs[0]);
-		if (!ret) {
-			/* XXX doesn't handle 'y' register correctly XXX */
-			ret = copy_regset_from_user(child, view, REGSET_GENERAL,
-						    32 * sizeof(u64),
-						    4 * sizeof(u64),
-						    &pregs->tstate);
-		}
+		ret = copy_regset_from_user(child, &ptrace64_view,
+					  REGSET_GENERAL, 0,
+					  19 * sizeof(u64),
+					  pregs);
 		break;
 
 	case PTRACE_GETFPREGS64:
-- 
2.11.0

[PATCH 14/41] arm64: take fetching compat reg out of pt_regs into a new helper

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/arm64/kernel/ptrace.c | 34 +++++++++++++++++-----------------
 1 file changed, 17 insertions(+), 17 deletions(-)

diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index 68b7f34a08f5..5b837741ab76 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -1237,6 +1237,22 @@ enum compat_regset {
 	REGSET_COMPAT_VFP,
 };
 
+static inline compat_ulong_t compat_get_user_reg(struct task_struct *task, int idx)
+{
+	struct pt_regs *regs = task_pt_regs(task);
+
+	switch (idx) {
+	case 15:
+		return regs->pc;
+	case 16:
+		return pstate_to_compat_psr(regs->pstate);
+	case 17:
+		return regs->orig_x0;
+	default:
+		return regs->regs[idx];
+	}
+}
+
 static int compat_gpr_get(struct task_struct *target,
 			  const struct user_regset *regset,
 			  unsigned int pos, unsigned int count,
@@ -1255,23 +1271,7 @@ static int compat_gpr_get(struct task_struct *target,
 		return -EIO;
 
 	for (i = 0; i < num_regs; ++i) {
-		unsigned int idx = start + i;
-		compat_ulong_t reg;
-
-		switch (idx) {
-		case 15:
-			reg = task_pt_regs(target)->pc;
-			break;
-		case 16:
-			reg = task_pt_regs(target)->pstate;
-			reg = pstate_to_compat_psr(reg);
-			break;
-		case 17:
-			reg = task_pt_regs(target)->orig_x0;
-			break;
-		default:
-			reg = task_pt_regs(target)->regs[idx];
-		}
+		compat_ulong_t reg = compat_get_user_reg(target, start + i);
 
 		if (kbuf) {
 			memcpy(kbuf, &reg, sizeof(reg));
-- 
2.11.0

[PATCH 15/41] arm64: get rid of copy_regset_to_user() in compat_ptrace_read_user()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/arm64/kernel/ptrace.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index 5b837741ab76..d5f3da5197a1 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -1541,9 +1541,7 @@ static int compat_ptrace_read_user(struct task_struct *tsk, compat_ulong_t off,
 	else if (off == COMPAT_PT_TEXT_END_ADDR)
 		tmp = tsk->mm->end_code;
 	else if (off < sizeof(compat_elf_gregset_t))
-		return copy_regset_to_user(tsk, &user_aarch32_view,
-					   REGSET_COMPAT_GPR, off,
-					   sizeof(compat_ulong_t), ret);
+		tmp = compat_get_user_reg(tsk, off >> 2);
 	else if (off >= COMPAT_USER_SZ)
 		return -EIO;
 	else
-- 
2.11.0

[PATCH 16/41] arm64: sanitize compat_ptrace_write_user()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

don't bother with copy_regset_from_user() (not to mention
set_fs())

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/arm64/kernel/ptrace.c | 29 ++++++++++++++++++++---------
 1 file changed, 20 insertions(+), 9 deletions(-)

diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index d5f3da5197a1..9f769e862f68 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -1553,8 +1553,8 @@ static int compat_ptrace_read_user(struct task_struct *tsk, compat_ulong_t off,
 static int compat_ptrace_write_user(struct task_struct *tsk, compat_ulong_t off,
 				    compat_ulong_t val)
 {
-	int ret;
-	mm_segment_t old_fs = get_fs();
+	struct pt_regs newregs = *task_pt_regs(tsk);
+	unsigned int idx = off / 4;
 
 	if (off & 3 || off >= COMPAT_USER_SZ)
 		return -EIO;
@@ -1562,14 +1562,25 @@ static int compat_ptrace_write_user(struct task_struct *tsk, compat_ulong_t off,
 	if (off >= sizeof(compat_elf_gregset_t))
 		return 0;
 
-	set_fs(KERNEL_DS);
-	ret = copy_regset_from_user(tsk, &user_aarch32_view,
-				    REGSET_COMPAT_GPR, off,
-				    sizeof(compat_ulong_t),
-				    &val);
-	set_fs(old_fs);
+	switch (idx) {
+	case 15:
+		newregs.pc = val;
+		break;
+	case 16:
+		newregs.pstate = compat_psr_to_pstate(val);
+		break;
+	case 17:
+		newregs.orig_x0 = val;
+		break;
+	default:
+		newregs.regs[idx] = val;
+	}
+
+	if (!valid_user_regs(&newregs.user_regs, tsk))
+		return -EINVAL;
 
-	return ret;
+	*task_pt_regs(tsk) = newregs;
+	return 0;
 }
 
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
-- 
2.11.0

[PATCH 17/41] copy_regset_to_user(): do all copyout at once.

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Turn copy_regset_to_user() into regset_get_alloc() + copy_to_user().
Now all ->get() calls have a kernel buffer as destination.

Note that we'd already eliminated the callers of copy_regset_to_user()
with non-zero offset; now that argument is simply unused.

Uninlined, while we are at it.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 include/linux/regset.h | 29 ++++-------------------------
 kernel/regset.c        | 26 ++++++++++++++++++++++++++
 2 files changed, 30 insertions(+), 25 deletions(-)

diff --git a/include/linux/regset.h b/include/linux/regset.h
index 968a032922d5..af57c1db1924 100644
--- a/include/linux/regset.h
+++ b/include/linux/regset.h
@@ -362,31 +362,10 @@ extern int regset_get_alloc(struct task_struct *target,
 			    unsigned int size,
 			    void **data);
 
-/**
- * copy_regset_to_user - fetch a thread's user_regset data into user memory
- * @target:	thread to be examined
- * @view:	&struct user_regset_view describing user thread machine state
- * @setno:	index in @view->regsets
- * @offset:	offset into the regset data, in bytes
- * @size:	amount of data to copy, in bytes
- * @data:	user-mode pointer to copy into
- */
-static inline int copy_regset_to_user(struct task_struct *target,
-				      const struct user_regset_view *view,
-				      unsigned int setno,
-				      unsigned int offset, unsigned int size,
-				      void __user *data)
-{
-	const struct user_regset *regset = &view->regsets[setno];
-
-	if (!regset->get)
-		return -EOPNOTSUPP;
-
-	if (!access_ok(data, size))
-		return -EFAULT;
-
-	return regset->get(target, regset, offset, size, NULL, data);
-}
+extern int copy_regset_to_user(struct task_struct *target,
+			       const struct user_regset_view *view,
+			       unsigned int setno, unsigned int offset,
+			       unsigned int size, void __user *data);
 
 /**
  * copy_regset_from_user - store into thread's user_regset data from user memory
diff --git a/kernel/regset.c b/kernel/regset.c
index 6b39fa0993ec..0a610983ce43 100644
--- a/kernel/regset.c
+++ b/kernel/regset.c
@@ -52,3 +52,29 @@ int regset_get_alloc(struct task_struct *target,
 	return __regset_get(target, regset, size, data);
 }
 EXPORT_SYMBOL(regset_get_alloc);
+
+/**
+ * copy_regset_to_user - fetch a thread's user_regset data into user memory
+ * @target:	thread to be examined
+ * @view:	&struct user_regset_view describing user thread machine state
+ * @setno:	index in @view->regsets
+ * @offset:	offset into the regset data, in bytes
+ * @size:	amount of data to copy, in bytes
+ * @data:	user-mode pointer to copy into
+ */
+int copy_regset_to_user(struct task_struct *target,
+			const struct user_regset_view *view,
+			unsigned int setno,
+			unsigned int offset, unsigned int size,
+			void __user *data)
+{
+	const struct user_regset *regset = &view->regsets[setno];
+	void *buf;
+	int ret;
+
+	ret = regset_get_alloc(target, regset, size, &buf);
+	if (ret > 0)
+		ret = copy_to_user(data, buf, ret) ? -EFAULT : 0;
+	kfree(buf);
+	return ret;
+}
-- 
2.11.0

[PATCH 18/41] regset: new method and helpers for it

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

->get2() takes task+regset+buffer, returns the amount of free space
left in the buffer on success and -E... on error.

buffer is represented as struct membuf - a pair of (kernel) pointer
and amount of space left

Primitives for writing to such:
	* membuf_write(buf, data, size)
	* membuf_zero(buf, size)
	* membuf_store(buf, value)

These are implemented as inlines (in case of membuf_store - a macro).
All writes are sequential; they become no-ops when there's no space
left.  Return value of all primitives is the amount of space left
after the operation, so they can be used as return values of ->get2().

Example of use:

// stores pt_regs of task + 64 bytes worth of zeroes + 32bit PID of task
int foo_get(struct task_struct *task, const struct regset *regset,
	    struct membuf to)
{
	membuf_write(&to, task_pt_regs(task), sizeof(struct pt_regs));
	membuf_zero(&to, 64);
	return membuf_store(&to, (u32)task_tgid_vnr(task));
}

regset_get()/regset_get_alloc() taught to use that thing if present.
By the end of the series all users of ->get() will be converted;
then ->get() and ->get_size() can go.

Note that unlike ->get() this thing always starts at offset 0 and,
since it only writes to kernel buffer, can't fail on copyout.
It can, of course, fail for other reasons, but those tend to
be less numerous.

The caller guarantees that the buffer size won't be bigger than
regset->n * regset->size.  That simplifies life for quite a few
instances.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 include/linux/regset.h | 51 ++++++++++++++++++++++++++++++++++++++++++++++++++
 kernel/regset.c        | 12 +++++++++++-
 2 files changed, 62 insertions(+), 1 deletion(-)

diff --git a/include/linux/regset.h b/include/linux/regset.h
index af57c1db1924..567a4b63f469 100644
--- a/include/linux/regset.h
+++ b/include/linux/regset.h
@@ -17,6 +17,52 @@
 struct task_struct;
 struct user_regset;
 
+struct membuf {
+	void *p;
+	size_t left;
+};
+
+static inline int membuf_zero(struct membuf *s, size_t size)
+{
+	if (s->left) {
+		if (size > s->left)
+			size = s->left;
+		memset(s->p, 0, size);
+		s->p += size;
+		s->left -= size;
+	}
+	return s->left;
+}
+
+static inline int membuf_write(struct membuf *s, const void *v, size_t size)
+{
+	if (s->left) {
+		if (size > s->left)
+			size = s->left;
+		memcpy(s->p, v, size);
+		s->p += size;
+		s->left -= size;
+	}
+	return s->left;
+}
+
+/* current s->p must be aligned for v; v must be a scalar */
+#define membuf_store(s, v)				\
+({							\
+	struct membuf *__s = (s);			\
+        if (__s->left) {				\
+		typeof(v) __v = (v);			\
+		size_t __size = sizeof(__v);		\
+		if (unlikely(__size > __s->left)) {	\
+			__size = __s->left;		\
+			memcpy(__s->p, &__v, __size);	\
+		} else {				\
+			*(typeof(__v + 0) *)__s->p = __v;	\
+		}					\
+		__s->p += __size;			\
+		__s->left -= __size;			\
+	}						\
+	__s->left;})
 
 /**
  * user_regset_active_fn - type of @active function in &struct user_regset
@@ -57,6 +103,10 @@ typedef int user_regset_get_fn(struct task_struct *target,
 			       unsigned int pos, unsigned int count,
 			       void *kbuf, void __user *ubuf);
 
+typedef int user_regset_get2_fn(struct task_struct *target,
+			       const struct user_regset *regset,
+			       struct membuf to);
+
 /**
  * user_regset_set_fn - type of @set function in &struct user_regset
  * @target:	thread being examined
@@ -186,6 +236,7 @@ typedef unsigned int user_regset_get_size_fn(struct task_struct *target,
  */
 struct user_regset {
 	user_regset_get_fn		*get;
+	user_regset_get2_fn		*get2;
 	user_regset_set_fn		*set;
 	user_regset_active_fn		*active;
 	user_regset_writeback_fn	*writeback;
diff --git a/kernel/regset.c b/kernel/regset.c
index 0a610983ce43..c548fb9f0943 100644
--- a/kernel/regset.c
+++ b/kernel/regset.c
@@ -11,7 +11,7 @@ static int __regset_get(struct task_struct *target,
 	void *p = *data, *to_free = NULL;
 	int res;
 
-	if (!regset->get)
+	if (!regset->get && !regset->get2)
 		return -EOPNOTSUPP;
 	if (size > regset->n * regset->size)
 		size = regset->n * regset->size;
@@ -20,6 +20,16 @@ static int __regset_get(struct task_struct *target,
 		if (!p)
 			return -ENOMEM;
 	}
+	if (regset->get2) {
+		res = regset->get2(target, regset,
+				   (struct membuf){.p = p, .left = size});
+		if (res < 0) {
+			kfree(to_free);
+			return res;
+		}
+		*data = p;
+		return size - res;
+	}
 	res = regset->get(target, regset, 0, size, p, NULL);
 	if (unlikely(res < 0)) {
 		kfree(to_free);
-- 
2.11.0

[PATCH 19/41] x86: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

	All instances of ->get() in arch/x86 switched; that might or might
not be worth splitting up.  Notes:

	* for xstateregs_get() the amount we want to store is determined at
the boot time; see init_xstate_size() and update_regset_xstate_info() for
details.  task->thread.fpu.state.xsave ends with a flexible array member and
the amount of data in it depends upon the FPU features supported/enabled.

	* fpregs_get() writes slightly less than full ->thread.fpu.state.fsave
(the last word is not copied); we pass the full size of state.fsave and let
membuf_write() trim to the amount declared by regset - __regset_get() will
make sure that the space in buffer is no more than that.

	* copy_xstate_to_user() and its helpers are gone now.

	* fpregs_soft_get() was getting user_regset_copyout() arguments
wrong.  Since "x86: x86 user_regset math_emu" back in 2008...  I really
doubt that it's worth splitting out for -stable, though - you need
a 486SX box for that to trigger...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/x86/include/asm/fpu/regset.h |   4 +-
 arch/x86/include/asm/fpu/xstate.h |   4 +-
 arch/x86/kernel/fpu/regset.c      |  39 ++++-----
 arch/x86/kernel/fpu/signal.c      |   3 +-
 arch/x86/kernel/fpu/xstate.c      | 164 +++++++-------------------------------
 arch/x86/kernel/ptrace.c          |  75 ++++++-----------
 arch/x86/kernel/tls.c             |  32 ++------
 arch/x86/kernel/tls.h             |   2 +-
 arch/x86/math-emu/fpu_entry.c     |  19 ++---
 9 files changed, 83 insertions(+), 259 deletions(-)

diff --git a/arch/x86/include/asm/fpu/regset.h b/arch/x86/include/asm/fpu/regset.h
index d5bdffb9d27f..4f928d6a367b 100644
--- a/arch/x86/include/asm/fpu/regset.h
+++ b/arch/x86/include/asm/fpu/regset.h
@@ -8,8 +8,8 @@
 #include <linux/regset.h>
 
 extern user_regset_active_fn regset_fpregs_active, regset_xregset_fpregs_active;
-extern user_regset_get_fn fpregs_get, xfpregs_get, fpregs_soft_get,
-				xstateregs_get;
+extern user_regset_get2_fn fpregs_get, xfpregs_get, fpregs_soft_get,
+				 xstateregs_get;
 extern user_regset_set_fn fpregs_set, xfpregs_set, fpregs_soft_set,
 				 xstateregs_set;
 
diff --git a/arch/x86/include/asm/fpu/xstate.h b/arch/x86/include/asm/fpu/xstate.h
index 422d8369012a..f691ea1bc086 100644
--- a/arch/x86/include/asm/fpu/xstate.h
+++ b/arch/x86/include/asm/fpu/xstate.h
@@ -71,8 +71,8 @@ extern void __init update_regset_xstate_info(unsigned int size,
 void *get_xsave_addr(struct xregs_state *xsave, int xfeature_nr);
 const void *get_xsave_field_ptr(int xfeature_nr);
 int using_compacted_format(void);
-int copy_xstate_to_kernel(void *kbuf, struct xregs_state *xsave, unsigned int offset, unsigned int size);
-int copy_xstate_to_user(void __user *ubuf, struct xregs_state *xsave, unsigned int offset, unsigned int size);
+struct membuf;
+void copy_xstate_to_kernel(struct membuf to, struct xregs_state *xsave);
 int copy_kernel_to_xstate(struct xregs_state *xsave, const void *kbuf);
 int copy_user_to_xstate(struct xregs_state *xsave, const void __user *ubuf);
 void copy_supervisor_to_kernel(struct xregs_state *xsave);
diff --git a/arch/x86/kernel/fpu/regset.c b/arch/x86/kernel/fpu/regset.c
index 91f80aca27fb..c413756ba89f 100644
--- a/arch/x86/kernel/fpu/regset.c
+++ b/arch/x86/kernel/fpu/regset.c
@@ -27,8 +27,7 @@ int regset_xregset_fpregs_active(struct task_struct *target, const struct user_r
 }
 
 int xfpregs_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count,
-		void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
 	struct fpu *fpu = &target->thread.fpu;
 
@@ -38,8 +37,7 @@ int xfpregs_get(struct task_struct *target, const struct user_regset *regset,
 	fpu__prepare_read(fpu);
 	fpstate_sanitize_xstate(fpu);
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &fpu->state.fxsave, 0, -1);
+	return membuf_write(&to, &fpu->state.fxsave, sizeof(struct fxregs_state));
 }
 
 int xfpregs_set(struct task_struct *target, const struct user_regset *regset,
@@ -74,12 +72,10 @@ int xfpregs_set(struct task_struct *target, const struct user_regset *regset,
 }
 
 int xstateregs_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count,
-		void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
 	struct fpu *fpu = &target->thread.fpu;
 	struct xregs_state *xsave;
-	int ret;
 
 	if (!boot_cpu_has(X86_FEATURE_XSAVE))
 		return -ENODEV;
@@ -89,10 +85,8 @@ int xstateregs_get(struct task_struct *target, const struct user_regset *regset,
 	fpu__prepare_read(fpu);
 
 	if (using_compacted_format()) {
-		if (kbuf)
-			ret = copy_xstate_to_kernel(kbuf, xsave, pos, count);
-		else
-			ret = copy_xstate_to_user(ubuf, xsave, pos, count);
+		copy_xstate_to_kernel(to, xsave);
+		return 0;
 	} else {
 		fpstate_sanitize_xstate(fpu);
 		/*
@@ -105,9 +99,8 @@ int xstateregs_get(struct task_struct *target, const struct user_regset *regset,
 		/*
 		 * Copy the xstate memory layout.
 		 */
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, xsave, 0, -1);
+		return membuf_write(&to, xsave, fpu_user_xstate_size);
 	}
-	return ret;
 }
 
 int xstateregs_set(struct task_struct *target, const struct user_regset *regset,
@@ -293,8 +286,7 @@ void convert_to_fxsr(struct fxregs_state *fxsave,
 }
 
 int fpregs_get(struct task_struct *target, const struct user_regset *regset,
-	       unsigned int pos, unsigned int count,
-	       void *kbuf, void __user *ubuf)
+	       struct membuf to)
 {
 	struct fpu *fpu = &target->thread.fpu;
 	struct user_i387_ia32_struct env;
@@ -302,23 +294,22 @@ int fpregs_get(struct task_struct *target, const struct user_regset *regset,
 	fpu__prepare_read(fpu);
 
 	if (!boot_cpu_has(X86_FEATURE_FPU))
-		return fpregs_soft_get(target, regset, pos, count, kbuf, ubuf);
+		return fpregs_soft_get(target, regset, to);
 
-	if (!boot_cpu_has(X86_FEATURE_FXSR))
-		return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					   &fpu->state.fsave, 0,
-					   -1);
+	if (!boot_cpu_has(X86_FEATURE_FXSR)) {
+		return membuf_write(&to, &fpu->state.fsave,
+				    sizeof(struct fregs_state));
+	}
 
 	fpstate_sanitize_xstate(fpu);
 
-	if (kbuf && pos == 0 && count == sizeof(env)) {
-		convert_from_fxsr(kbuf, target);
+	if (to.left == sizeof(env)) {
+		convert_from_fxsr(to.p, target);
 		return 0;
 	}
 
 	convert_from_fxsr(&env, target);
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &env, 0, -1);
+	return membuf_write(&to, &env, sizeof(env));
 }
 
 int fpregs_set(struct task_struct *target, const struct user_regset *regset,
diff --git a/arch/x86/kernel/fpu/signal.c b/arch/x86/kernel/fpu/signal.c
index e0b832df7404..a4ec65317a7f 100644
--- a/arch/x86/kernel/fpu/signal.c
+++ b/arch/x86/kernel/fpu/signal.c
@@ -172,7 +172,8 @@ int copy_fpstate_to_sigframe(void __user *buf, void __user *buf_fx, int size)
 
 	if (!static_cpu_has(X86_FEATURE_FPU)) {
 		struct user_i387_ia32_struct fp;
-		fpregs_soft_get(current, NULL, 0, sizeof(fp), &fp, NULL);
+		fpregs_soft_get(current, NULL, (struct membuf){.p = &fp,
+						.left = sizeof(fp)});
 		return copy_to_user(buf, &fp, sizeof(fp)) ? -EFAULT : 0;
 	}
 
diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
index bda2e5eaca0e..c80d4734c1f6 100644
--- a/arch/x86/kernel/fpu/xstate.c
+++ b/arch/x86/kernel/fpu/xstate.c
@@ -1009,32 +1009,20 @@ static inline bool xfeatures_mxcsr_quirk(u64 xfeatures)
 	return true;
 }
 
-static void fill_gap(unsigned to, void **kbuf, unsigned *pos, unsigned *count)
+static void fill_gap(struct membuf *to, unsigned *last, unsigned offset)
 {
-	if (*pos < to) {
-		unsigned size = to - *pos;
-
-		if (size > *count)
-			size = *count;
-		memcpy(*kbuf, (void *)&init_fpstate.xsave + *pos, size);
-		*kbuf += size;
-		*pos += size;
-		*count -= size;
-	}
+	if (*last >= offset)
+		return;
+	membuf_write(to, (void *)&init_fpstate.xsave + *last, offset - *last);
+	*last = offset;
 }
 
-static void copy_part(unsigned offset, unsigned size, void *from,
-			void **kbuf, unsigned *pos, unsigned *count)
+static void copy_part(struct membuf *to, unsigned *last, unsigned offset,
+		      unsigned size, void *from)
 {
-	fill_gap(offset, kbuf, pos, count);
-	if (size > *count)
-		size = *count;
-	if (size) {
-		memcpy(*kbuf, from, size);
-		*kbuf += size;
-		*pos += size;
-		*count -= size;
-	}
+	fill_gap(to, last, offset);
+	membuf_write(to, from, size);
+	*last = offset + size;
 }
 
 /*
@@ -1044,20 +1032,15 @@ static void copy_part(unsigned offset, unsigned size, void *from,
  * It supports partial copy but pos always starts from zero. This is called
  * from xstateregs_get() and there we check the CPU has XSAVES.
  */
-int copy_xstate_to_kernel(void *kbuf, struct xregs_state *xsave, unsigned int offset_start, unsigned int size_total)
+void copy_xstate_to_kernel(struct membuf to, struct xregs_state *xsave)
 {
 	struct xstate_header header;
 	const unsigned off_mxcsr = offsetof(struct fxregs_state, mxcsr);
-	unsigned count = size_total;
+	unsigned size = to.left;
+	unsigned last = 0;
 	int i;
 
 	/*
-	 * Currently copy_regset_to_user() starts from pos 0:
-	 */
-	if (unlikely(offset_start != 0))
-		return -EFAULT;
-
-	/*
 	 * The destination is a ptrace buffer; we put in only user xstates:
 	 */
 	memset(&header, 0, sizeof(header));
@@ -1065,27 +1048,26 @@ int copy_xstate_to_kernel(void *kbuf, struct xregs_state *xsave, unsigned int of
 	header.xfeatures &= xfeatures_mask_user();
 
 	if (header.xfeatures & XFEATURE_MASK_FP)
-		copy_part(0, off_mxcsr,
-			  &xsave->i387, &kbuf, &offset_start, &count);
+		copy_part(&to, &last, 0, off_mxcsr, &xsave->i387);
 	if (header.xfeatures & (XFEATURE_MASK_SSE | XFEATURE_MASK_YMM))
-		copy_part(off_mxcsr, MXCSR_AND_FLAGS_SIZE,
-			  &xsave->i387.mxcsr, &kbuf, &offset_start, &count);
+		copy_part(&to, &last, off_mxcsr,
+			  MXCSR_AND_FLAGS_SIZE, &xsave->i387.mxcsr);
 	if (header.xfeatures & XFEATURE_MASK_FP)
-		copy_part(offsetof(struct fxregs_state, st_space), 128,
-			  &xsave->i387.st_space, &kbuf, &offset_start, &count);
+		copy_part(&to, &last, offsetof(struct fxregs_state, st_space),
+			  128, &xsave->i387.st_space);
 	if (header.xfeatures & XFEATURE_MASK_SSE)
-		copy_part(xstate_offsets[XFEATURE_MASK_SSE], 256,
-			  &xsave->i387.xmm_space, &kbuf, &offset_start, &count);
+		copy_part(&to, &last, xstate_offsets[XFEATURE_MASK_SSE],
+			  256, &xsave->i387.xmm_space);
 	/*
 	 * Fill xsave->i387.sw_reserved value for ptrace frame:
 	 */
-	copy_part(offsetof(struct fxregs_state, sw_reserved), 48,
-		  xstate_fx_sw_bytes, &kbuf, &offset_start, &count);
+	copy_part(&to, &last, offsetof(struct fxregs_state, sw_reserved),
+		  48, xstate_fx_sw_bytes);
 	/*
 	 * Copy xregs_state->header:
 	 */
-	copy_part(offsetof(struct xregs_state, header), sizeof(header),
-		  &header, &kbuf, &offset_start, &count);
+	copy_part(&to, &last, offsetof(struct xregs_state, header),
+		  sizeof(header), &header);
 
 	for (i = FIRST_EXTENDED_XFEATURE; i < XFEATURE_MAX; i++) {
 		/*
@@ -1094,104 +1076,12 @@ int copy_xstate_to_kernel(void *kbuf, struct xregs_state *xsave, unsigned int of
 		if ((header.xfeatures >> i) & 1) {
 			void *src = __raw_xsave_addr(xsave, i);
 
-			copy_part(xstate_offsets[i], xstate_sizes[i],
-				  src, &kbuf, &offset_start, &count);
+			copy_part(&to, &last, xstate_offsets[i],
+				  xstate_sizes[i], src);
 		}
 
 	}
-	fill_gap(size_total, &kbuf, &offset_start, &count);
-
-	return 0;
-}
-
-static inline int
-__copy_xstate_to_user(void __user *ubuf, const void *data, unsigned int offset, unsigned int size, unsigned int size_total)
-{
-	if (!size)
-		return 0;
-
-	if (offset < size_total) {
-		unsigned int copy = min(size, size_total - offset);
-
-		if (__copy_to_user(ubuf + offset, data, copy))
-			return -EFAULT;
-	}
-	return 0;
-}
-
-/*
- * Convert from kernel XSAVES compacted format to standard format and copy
- * to a user-space buffer. It supports partial copy but pos always starts from
- * zero. This is called from xstateregs_get() and there we check the CPU
- * has XSAVES.
- */
-int copy_xstate_to_user(void __user *ubuf, struct xregs_state *xsave, unsigned int offset_start, unsigned int size_total)
-{
-	unsigned int offset, size;
-	int ret, i;
-	struct xstate_header header;
-
-	/*
-	 * Currently copy_regset_to_user() starts from pos 0:
-	 */
-	if (unlikely(offset_start != 0))
-		return -EFAULT;
-
-	/*
-	 * The destination is a ptrace buffer; we put in only user xstates:
-	 */
-	memset(&header, 0, sizeof(header));
-	header.xfeatures = xsave->header.xfeatures;
-	header.xfeatures &= xfeatures_mask_user();
-
-	/*
-	 * Copy xregs_state->header:
-	 */
-	offset = offsetof(struct xregs_state, header);
-	size = sizeof(header);
-
-	ret = __copy_xstate_to_user(ubuf, &header, offset, size, size_total);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < XFEATURE_MAX; i++) {
-		/*
-		 * Copy only in-use xstates:
-		 */
-		if ((header.xfeatures >> i) & 1) {
-			void *src = __raw_xsave_addr(xsave, i);
-
-			offset = xstate_offsets[i];
-			size = xstate_sizes[i];
-
-			/* The next component has to fit fully into the output buffer: */
-			if (offset + size > size_total)
-				break;
-
-			ret = __copy_xstate_to_user(ubuf, src, offset, size, size_total);
-			if (ret)
-				return ret;
-		}
-
-	}
-
-	if (xfeatures_mxcsr_quirk(header.xfeatures)) {
-		offset = offsetof(struct fxregs_state, mxcsr);
-		size = MXCSR_AND_FLAGS_SIZE;
-		__copy_xstate_to_user(ubuf, &xsave->i387.mxcsr, offset, size, size_total);
-	}
-
-	/*
-	 * Fill xsave->i387.sw_reserved value for ptrace frame:
-	 */
-	offset = offsetof(struct fxregs_state, sw_reserved);
-	size = sizeof(xstate_fx_sw_bytes);
-
-	ret = __copy_xstate_to_user(ubuf, xstate_fx_sw_bytes, offset, size, size_total);
-	if (ret)
-		return ret;
-
-	return 0;
+	fill_gap(&to, &last, size);
 }
 
 /*
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index 44130588987f..81725806d61d 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -433,26 +433,12 @@ static unsigned long getreg(struct task_struct *task, unsigned long offset)
 
 static int genregs_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+		       struct membuf to)
 {
-	if (kbuf) {
-		unsigned long *k = kbuf;
-		while (count >= sizeof(*k)) {
-			*k++ = getreg(target, pos);
-			count -= sizeof(*k);
-			pos += sizeof(*k);
-		}
-	} else {
-		unsigned long __user *u = ubuf;
-		while (count >= sizeof(*u)) {
-			if (__put_user(getreg(target, pos), u++))
-				return -EFAULT;
-			count -= sizeof(*u);
-			pos += sizeof(*u);
-		}
-	}
+	int reg;
 
+	for (reg = 0; to.left; reg++)
+		membuf_store(&to, getreg(target, reg * sizeof(unsigned long)));
 	return 0;
 }
 
@@ -716,16 +702,14 @@ static int ioperm_active(struct task_struct *target,
 
 static int ioperm_get(struct task_struct *target,
 		      const struct user_regset *regset,
-		      unsigned int pos, unsigned int count,
-		      void *kbuf, void __user *ubuf)
+		      struct membuf to)
 {
 	struct io_bitmap *iobm = target->thread.io_bitmap;
 
 	if (!iobm)
 		return -ENXIO;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   iobm->bitmap, 0, IO_BITMAP_BYTES);
+	return membuf_write(&to, iobm->bitmap, IO_BITMAP_BYTES);
 }
 
 /*
@@ -1003,28 +987,15 @@ static int getreg32(struct task_struct *child, unsigned regno, u32 *val)
 
 static int genregs32_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
-	if (kbuf) {
-		compat_ulong_t *k = kbuf;
-		while (count >= sizeof(*k)) {
-			getreg32(target, pos, k++);
-			count -= sizeof(*k);
-			pos += sizeof(*k);
-		}
-	} else {
-		compat_ulong_t __user *u = ubuf;
-		while (count >= sizeof(*u)) {
-			compat_ulong_t word;
-			getreg32(target, pos, &word);
-			if (__put_user(word, u++))
-				return -EFAULT;
-			count -= sizeof(*u);
-			pos += sizeof(*u);
-		}
-	}
+	int reg;
 
+	for (reg = 0; to.left; reg++) {
+		u32 val;
+		getreg32(target, reg * 4, &val);
+		membuf_store(&to, val);
+	}
 	return 0;
 }
 
@@ -1234,25 +1205,25 @@ static struct user_regset x86_64_regsets[] __ro_after_init = {
 		.core_note_type = NT_PRSTATUS,
 		.n = sizeof(struct user_regs_struct) / sizeof(long),
 		.size = sizeof(long), .align = sizeof(long),
-		.get = genregs_get, .set = genregs_set
+		.get2 = genregs_get, .set = genregs_set
 	},
 	[REGSET_FP] = {
 		.core_note_type = NT_PRFPREG,
 		.n = sizeof(struct user_i387_struct) / sizeof(long),
 		.size = sizeof(long), .align = sizeof(long),
-		.active = regset_xregset_fpregs_active, .get = xfpregs_get, .set = xfpregs_set
+		.active = regset_xregset_fpregs_active, .get2 = xfpregs_get, .set = xfpregs_set
 	},
 	[REGSET_XSTATE] = {
 		.core_note_type = NT_X86_XSTATE,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = xstateregs_active, .get = xstateregs_get,
+		.active = xstateregs_active, .get2 = xstateregs_get,
 		.set = xstateregs_set
 	},
 	[REGSET_IOPERM64] = {
 		.core_note_type = NT_386_IOPERM,
 		.n = IO_BITMAP_LONGS,
 		.size = sizeof(long), .align = sizeof(long),
-		.active = ioperm_active, .get = ioperm_get
+		.active = ioperm_active, .get2 = ioperm_get
 	},
 };
 
@@ -1275,24 +1246,24 @@ static struct user_regset x86_32_regsets[] __ro_after_init = {
 		.core_note_type = NT_PRSTATUS,
 		.n = sizeof(struct user_regs_struct32) / sizeof(u32),
 		.size = sizeof(u32), .align = sizeof(u32),
-		.get = genregs32_get, .set = genregs32_set
+		.get2 = genregs32_get, .set = genregs32_set
 	},
 	[REGSET_FP] = {
 		.core_note_type = NT_PRFPREG,
 		.n = sizeof(struct user_i387_ia32_struct) / sizeof(u32),
 		.size = sizeof(u32), .align = sizeof(u32),
-		.active = regset_fpregs_active, .get = fpregs_get, .set = fpregs_set
+		.active = regset_fpregs_active, .get2 = fpregs_get, .set = fpregs_set
 	},
 	[REGSET_XFP] = {
 		.core_note_type = NT_PRXFPREG,
 		.n = sizeof(struct user32_fxsr_struct) / sizeof(u32),
 		.size = sizeof(u32), .align = sizeof(u32),
-		.active = regset_xregset_fpregs_active, .get = xfpregs_get, .set = xfpregs_set
+		.active = regset_xregset_fpregs_active, .get2 = xfpregs_get, .set = xfpregs_set
 	},
 	[REGSET_XSTATE] = {
 		.core_note_type = NT_X86_XSTATE,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = xstateregs_active, .get = xstateregs_get,
+		.active = xstateregs_active, .get2 = xstateregs_get,
 		.set = xstateregs_set
 	},
 	[REGSET_TLS] = {
@@ -1301,13 +1272,13 @@ static struct user_regset x86_32_regsets[] __ro_after_init = {
 		.size = sizeof(struct user_desc),
 		.align = sizeof(struct user_desc),
 		.active = regset_tls_active,
-		.get = regset_tls_get, .set = regset_tls_set
+		.get2 = regset_tls_get, .set = regset_tls_set
 	},
 	[REGSET_IOPERM32] = {
 		.core_note_type = NT_386_IOPERM,
 		.n = IO_BITMAP_BYTES / sizeof(u32),
 		.size = sizeof(u32), .align = sizeof(u32),
-		.active = ioperm_active, .get = ioperm_get
+		.active = ioperm_active, .get2 = ioperm_get
 	},
 };
 
diff --git a/arch/x86/kernel/tls.c b/arch/x86/kernel/tls.c
index 71d3fef1edc9..64a496a0687f 100644
--- a/arch/x86/kernel/tls.c
+++ b/arch/x86/kernel/tls.c
@@ -256,36 +256,16 @@ int regset_tls_active(struct task_struct *target,
 }
 
 int regset_tls_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	const struct desc_struct *tls;
+	struct user_desc v;
+	int pos;
 
-	if (pos >= GDT_ENTRY_TLS_ENTRIES * sizeof(struct user_desc) ||
-	    (pos % sizeof(struct user_desc)) != 0 ||
-	    (count % sizeof(struct user_desc)) != 0)
-		return -EINVAL;
-
-	pos /= sizeof(struct user_desc);
-	count /= sizeof(struct user_desc);
-
-	tls = &target->thread.tls_array[pos];
-
-	if (kbuf) {
-		struct user_desc *info = kbuf;
-		while (count-- > 0)
-			fill_user_desc(info++, GDT_ENTRY_TLS_MIN + pos++,
-				       tls++);
-	} else {
-		struct user_desc __user *u_info = ubuf;
-		while (count-- > 0) {
-			struct user_desc info;
-			fill_user_desc(&info, GDT_ENTRY_TLS_MIN + pos++, tls++);
-			if (__copy_to_user(u_info++, &info, sizeof(info)))
-				return -EFAULT;
-		}
+	for (pos = 0, tls = target->thread.tls_array; to.left; pos++, tls++) {
+		fill_user_desc(&v, GDT_ENTRY_TLS_MIN + pos, tls);
+		membuf_write(&to, &v, sizeof(v));
 	}
-
 	return 0;
 }
 
diff --git a/arch/x86/kernel/tls.h b/arch/x86/kernel/tls.h
index 3a76e1d3535e..fc39447a0c1a 100644
--- a/arch/x86/kernel/tls.h
+++ b/arch/x86/kernel/tls.h
@@ -12,7 +12,7 @@
 #include <linux/regset.h>
 
 extern user_regset_active_fn regset_tls_active;
-extern user_regset_get_fn regset_tls_get;
+extern user_regset_get2_fn regset_tls_get;
 extern user_regset_set_fn regset_tls_set;
 
 #endif	/* _ARCH_X86_KERNEL_TLS_H */
diff --git a/arch/x86/math-emu/fpu_entry.c b/arch/x86/math-emu/fpu_entry.c
index a873da6b46d6..8679a9d6c47f 100644
--- a/arch/x86/math-emu/fpu_entry.c
+++ b/arch/x86/math-emu/fpu_entry.c
@@ -689,12 +689,10 @@ int fpregs_soft_set(struct task_struct *target,
 
 int fpregs_soft_get(struct task_struct *target,
 		    const struct user_regset *regset,
-		    unsigned int pos, unsigned int count,
-		    void *kbuf, void __user *ubuf)
+		    struct membuf to)
 {
 	struct swregs_state *s387 = &target->thread.fpu.state.soft;
 	const void *space = s387->st_space;
-	int ret;
 	int offset = (S387->ftop & 7) * 10, other = 80 - offset;
 
 	RE_ENTRANT_CHECK_OFF;
@@ -709,18 +707,11 @@ int fpregs_soft_get(struct task_struct *target,
 	S387->fos |= 0xffff0000;
 #endif /* PECULIAR_486 */
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, s387, 0,
-				  offsetof(struct swregs_state, st_space));
-
-	/* Copy all registers in stack order. */
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  space + offset, 0, other);
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  space, 0, offset);
+	membuf_write(&to, s387, offsetof(struct swregs_state, st_space));
+	membuf_write(&to, space + offset, other);
+	membuf_write(&to, space, offset);
 
 	RE_ENTRANT_CHECK_ON;
 
-	return ret;
+	return 0;
 }
-- 
2.11.0

[PATCH 20/41] powerpc: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Note: compat variant of REGSET_TM_CGPR is almost certainly wrong;
it claims to be 48*64bit, but just as compat REGSET_GPR it stores
44*32bit of (truncated) registers + 4 32bit zeros... followed by
48 more 32bit zeroes.  Might be too late to change - it's a userland
ABI, after all ;-/

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/powerpc/kernel/ptrace/ptrace-altivec.c |  37 ++----
 arch/powerpc/kernel/ptrace/ptrace-decl.h    |  44 +++----
 arch/powerpc/kernel/ptrace/ptrace-novsx.c   |   5 +-
 arch/powerpc/kernel/ptrace/ptrace-spe.c     |  16 +--
 arch/powerpc/kernel/ptrace/ptrace-tm.c      | 152 +++++++----------------
 arch/powerpc/kernel/ptrace/ptrace-view.c    | 185 ++++++++++------------------
 arch/powerpc/kernel/ptrace/ptrace-vsx.c     |  13 +-
 7 files changed, 148 insertions(+), 304 deletions(-)

diff --git a/arch/powerpc/kernel/ptrace/ptrace-altivec.c b/arch/powerpc/kernel/ptrace/ptrace-altivec.c
index dd8b75dfbd06..0d9bc4bd4972 100644
--- a/arch/powerpc/kernel/ptrace/ptrace-altivec.c
+++ b/arch/powerpc/kernel/ptrace/ptrace-altivec.c
@@ -41,38 +41,25 @@ int vr_active(struct task_struct *target, const struct user_regset *regset)
  * };
  */
 int vr_get(struct task_struct *target, const struct user_regset *regset,
-	   unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	   struct membuf to)
 {
-	int ret;
+	union {
+		elf_vrreg_t reg;
+		u32 word;
+	} vrsave;
 
 	flush_altivec_to_thread(target);
 
 	BUILD_BUG_ON(offsetof(struct thread_vr_state, vscr) !=
 		     offsetof(struct thread_vr_state, vr[32]));
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.vr_state, 0,
-				  33 * sizeof(vector128));
-	if (!ret) {
-		/*
-		 * Copy out only the low-order word of vrsave.
-		 */
-		int start, end;
-		union {
-			elf_vrreg_t reg;
-			u32 word;
-		} vrsave;
-		memset(&vrsave, 0, sizeof(vrsave));
-
-		vrsave.word = target->thread.vrsave;
-
-		start = 33 * sizeof(vector128);
-		end = start + sizeof(vrsave);
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &vrsave,
-					  start, end);
-	}
-
-	return ret;
+	membuf_write(&to, &target->thread.vr_state, 33 * sizeof(vector128));
+	/*
+	 * Copy out only the low-order word of vrsave.
+	 */
+	memset(&vrsave, 0, sizeof(vrsave));
+	vrsave.word = target->thread.vrsave;
+	return membuf_write(&to, &vrsave, sizeof(vrsave));
 }
 
 /*
diff --git a/arch/powerpc/kernel/ptrace/ptrace-decl.h b/arch/powerpc/kernel/ptrace/ptrace-decl.h
index 3c8a81999292..67447a6197eb 100644
--- a/arch/powerpc/kernel/ptrace/ptrace-decl.h
+++ b/arch/powerpc/kernel/ptrace/ptrace-decl.h
@@ -63,8 +63,7 @@ enum powerpc_regset {
 
 /* ptrace-(no)vsx */
 
-int fpr_get(struct task_struct *target, const struct user_regset *regset,
-	    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn fpr_get;
 int fpr_set(struct task_struct *target, const struct user_regset *regset,
 	    unsigned int pos, unsigned int count,
 	    const void *kbuf, const void __user *ubuf);
@@ -72,8 +71,7 @@ int fpr_set(struct task_struct *target, const struct user_regset *regset,
 /* ptrace-vsx */
 
 int vsr_active(struct task_struct *target, const struct user_regset *regset);
-int vsr_get(struct task_struct *target, const struct user_regset *regset,
-	    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn vsr_get;
 int vsr_set(struct task_struct *target, const struct user_regset *regset,
 	    unsigned int pos, unsigned int count,
 	    const void *kbuf, const void __user *ubuf);
@@ -81,8 +79,7 @@ int vsr_set(struct task_struct *target, const struct user_regset *regset,
 /* ptrace-altivec */
 
 int vr_active(struct task_struct *target, const struct user_regset *regset);
-int vr_get(struct task_struct *target, const struct user_regset *regset,
-	   unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn vr_get;
 int vr_set(struct task_struct *target, const struct user_regset *regset,
 	   unsigned int pos, unsigned int count,
 	   const void *kbuf, const void __user *ubuf);
@@ -90,8 +87,7 @@ int vr_set(struct task_struct *target, const struct user_regset *regset,
 /* ptrace-spe */
 
 int evr_active(struct task_struct *target, const struct user_regset *regset);
-int evr_get(struct task_struct *target, const struct user_regset *regset,
-	    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn evr_get;
 int evr_set(struct task_struct *target, const struct user_regset *regset,
 	    unsigned int pos, unsigned int count,
 	    const void *kbuf, const void __user *ubuf);
@@ -100,9 +96,8 @@ int evr_set(struct task_struct *target, const struct user_regset *regset,
 
 int gpr32_get_common(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-			    void *kbuf, void __user *ubuf,
-			    unsigned long *regs);
+		     struct membuf to,
+		     unsigned long *regs);
 int gpr32_set_common(struct task_struct *target,
 		     const struct user_regset *regset,
 		     unsigned int pos, unsigned int count,
@@ -118,55 +113,46 @@ static inline void flush_tmregs_to_thread(struct task_struct *tsk) { }
 #endif
 
 int tm_cgpr_active(struct task_struct *target, const struct user_regset *regset);
-int tm_cgpr_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_cgpr_get;
 int tm_cgpr_set(struct task_struct *target, const struct user_regset *regset,
 		unsigned int pos, unsigned int count,
 		const void *kbuf, const void __user *ubuf);
 int tm_cfpr_active(struct task_struct *target, const struct user_regset *regset);
-int tm_cfpr_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_cfpr_get;
 int tm_cfpr_set(struct task_struct *target, const struct user_regset *regset,
 		unsigned int pos, unsigned int count,
 		const void *kbuf, const void __user *ubuf);
 int tm_cvmx_active(struct task_struct *target, const struct user_regset *regset);
-int tm_cvmx_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_cvmx_get;
 int tm_cvmx_set(struct task_struct *target, const struct user_regset *regset,
 		unsigned int pos, unsigned int count,
 		const void *kbuf, const void __user *ubuf);
 int tm_cvsx_active(struct task_struct *target, const struct user_regset *regset);
-int tm_cvsx_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_cvsx_get;
 int tm_cvsx_set(struct task_struct *target, const struct user_regset *regset,
 		unsigned int pos, unsigned int count,
 		const void *kbuf, const void __user *ubuf);
 int tm_spr_active(struct task_struct *target, const struct user_regset *regset);
-int tm_spr_get(struct task_struct *target, const struct user_regset *regset,
-	       unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_spr_get;
 int tm_spr_set(struct task_struct *target, const struct user_regset *regset,
 	       unsigned int pos, unsigned int count,
 	       const void *kbuf, const void __user *ubuf);
 int tm_tar_active(struct task_struct *target, const struct user_regset *regset);
-int tm_tar_get(struct task_struct *target, const struct user_regset *regset,
-	       unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_tar_get;
 int tm_tar_set(struct task_struct *target, const struct user_regset *regset,
 	       unsigned int pos, unsigned int count,
 	       const void *kbuf, const void __user *ubuf);
 int tm_ppr_active(struct task_struct *target, const struct user_regset *regset);
-int tm_ppr_get(struct task_struct *target, const struct user_regset *regset,
-	       unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_ppr_get;
 int tm_ppr_set(struct task_struct *target, const struct user_regset *regset,
 	       unsigned int pos, unsigned int count,
 	       const void *kbuf, const void __user *ubuf);
 int tm_dscr_active(struct task_struct *target, const struct user_regset *regset);
-int tm_dscr_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_dscr_get;
 int tm_dscr_set(struct task_struct *target, const struct user_regset *regset,
 		unsigned int pos, unsigned int count,
 		const void *kbuf, const void __user *ubuf);
-int tm_cgpr32_get(struct task_struct *target, const struct user_regset *regset,
-		  unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf);
+user_regset_get2_fn tm_cgpr32_get;
 int tm_cgpr32_set(struct task_struct *target, const struct user_regset *regset,
 		  unsigned int pos, unsigned int count,
 		  const void *kbuf, const void __user *ubuf);
diff --git a/arch/powerpc/kernel/ptrace/ptrace-novsx.c b/arch/powerpc/kernel/ptrace/ptrace-novsx.c
index b2dc4e92d11a..b3b36835658a 100644
--- a/arch/powerpc/kernel/ptrace/ptrace-novsx.c
+++ b/arch/powerpc/kernel/ptrace/ptrace-novsx.c
@@ -19,15 +19,14 @@
  * };
  */
 int fpr_get(struct task_struct *target, const struct user_regset *regset,
-	    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	    struct membuf to)
 {
 	BUILD_BUG_ON(offsetof(struct thread_fp_state, fpscr) !=
 		     offsetof(struct thread_fp_state, fpr[32]));
 
 	flush_fp_to_thread(target);
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &target->thread.fp_state, 0, -1);
+	return membuf_write(&to, &target->thread.fp_state, 33 * sizeof(u64));
 }
 
 /*
diff --git a/arch/powerpc/kernel/ptrace/ptrace-spe.c b/arch/powerpc/kernel/ptrace/ptrace-spe.c
index 68b86b4a4be4..47034d069045 100644
--- a/arch/powerpc/kernel/ptrace/ptrace-spe.c
+++ b/arch/powerpc/kernel/ptrace/ptrace-spe.c
@@ -23,25 +23,17 @@ int evr_active(struct task_struct *target, const struct user_regset *regset)
 }
 
 int evr_get(struct task_struct *target, const struct user_regset *regset,
-	    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	    struct membuf to)
 {
-	int ret;
-
 	flush_spe_to_thread(target);
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.evr,
-				  0, sizeof(target->thread.evr));
+	membuf_write(&to, &target->thread.evr, sizeof(target->thread.evr));
 
 	BUILD_BUG_ON(offsetof(struct thread_struct, acc) + sizeof(u64) !=
 		     offsetof(struct thread_struct, spefscr));
 
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.acc,
-					  sizeof(target->thread.evr), -1);
-
-	return ret;
+	return membuf_write(&to, &target->thread.acc,
+				sizeof(u64) + sizeof(u32));
 }
 
 int evr_set(struct task_struct *target, const struct user_regset *regset,
diff --git a/arch/powerpc/kernel/ptrace/ptrace-tm.c b/arch/powerpc/kernel/ptrace/ptrace-tm.c
index 32d62c606681..54f2d076206f 100644
--- a/arch/powerpc/kernel/ptrace/ptrace-tm.c
+++ b/arch/powerpc/kernel/ptrace/ptrace-tm.c
@@ -70,10 +70,7 @@ int tm_cgpr_active(struct task_struct *target, const struct user_regset *regset)
  * tm_cgpr_get - get CGPR registers
  * @target:	The target task.
  * @regset:	The user regset structure.
- * @pos:	The buffer position.
- * @count:	Number of bytes to copy.
- * @kbuf:	Kernel buffer to copy from.
- * @ubuf:	User buffer to copy into.
+ * @to:		Destination of copy.
  *
  * This function gets transaction checkpointed GPR registers.
  *
@@ -87,10 +84,8 @@ int tm_cgpr_active(struct task_struct *target, const struct user_regset *regset)
  * };
  */
 int tm_cgpr_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
-	int ret;
-
 	if (!cpu_has_feature(CPU_FTR_TM))
 		return -ENODEV;
 
@@ -101,31 +96,18 @@ int tm_cgpr_get(struct task_struct *target, const struct user_regset *regset,
 	flush_fp_to_thread(target);
 	flush_altivec_to_thread(target);
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.ckpt_regs,
-				  0, offsetof(struct pt_regs, msr));
-	if (!ret) {
-		unsigned long msr = get_user_ckpt_msr(target);
-
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &msr,
-					  offsetof(struct pt_regs, msr),
-					  offsetof(struct pt_regs, msr) +
-					  sizeof(msr));
-	}
+	membuf_write(&to, &target->thread.ckpt_regs,
+			offsetof(struct pt_regs, msr));
+	membuf_store(&to, get_user_ckpt_msr(target));
 
 	BUILD_BUG_ON(offsetof(struct pt_regs, orig_gpr3) !=
 		     offsetof(struct pt_regs, msr) + sizeof(long));
 
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.ckpt_regs.orig_gpr3,
-					  offsetof(struct pt_regs, orig_gpr3),
-					  sizeof(struct user_pt_regs));
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       sizeof(struct user_pt_regs), -1);
-
-	return ret;
+	membuf_write(&to, &target->thread.ckpt_regs.orig_gpr3,
+			sizeof(struct user_pt_regs) -
+			offsetof(struct pt_regs, orig_gpr3));
+	return membuf_zero(&to, ELF_NGREG * sizeof(unsigned long) -
+			sizeof(struct user_pt_regs));
 }
 
 /*
@@ -229,10 +211,7 @@ int tm_cfpr_active(struct task_struct *target, const struct user_regset *regset)
  * tm_cfpr_get - get CFPR registers
  * @target:	The target task.
  * @regset:	The user regset structure.
- * @pos:	The buffer position.
- * @count:	Number of bytes to copy.
- * @kbuf:	Kernel buffer to copy from.
- * @ubuf:	User buffer to copy into.
+ * @to:		Destination of copy.
  *
  * This function gets in transaction checkpointed FPR registers.
  *
@@ -247,7 +226,7 @@ int tm_cfpr_active(struct task_struct *target, const struct user_regset *regset)
  *};
  */
 int tm_cfpr_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
 	u64 buf[33];
 	int i;
@@ -266,7 +245,7 @@ int tm_cfpr_get(struct task_struct *target, const struct user_regset *regset,
 	for (i = 0; i < 32 ; i++)
 		buf[i] = target->thread.TS_CKFPR(i);
 	buf[32] = target->thread.ckfp_state.fpscr;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, buf, 0, -1);
+	return membuf_write(&to, buf, sizeof(buf));
 }
 
 /**
@@ -344,10 +323,7 @@ int tm_cvmx_active(struct task_struct *target, const struct user_regset *regset)
  * tm_cvmx_get - get CMVX registers
  * @target:	The target task.
  * @regset:	The user regset structure.
- * @pos:	The buffer position.
- * @count:	Number of bytes to copy.
- * @kbuf:	Kernel buffer to copy from.
- * @ubuf:	User buffer to copy into.
+ * @to:		Destination of copy.
  *
  * This function gets in transaction checkpointed VMX registers.
  *
@@ -363,10 +339,12 @@ int tm_cvmx_active(struct task_struct *target, const struct user_regset *regset)
  *};
  */
 int tm_cvmx_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
-	int ret;
-
+	union {
+		elf_vrreg_t reg;
+		u32 word;
+	} vrsave;
 	BUILD_BUG_ON(TVSO(vscr) != TVSO(vr[32]));
 
 	if (!cpu_has_feature(CPU_FTR_TM))
@@ -380,23 +358,13 @@ int tm_cvmx_get(struct task_struct *target, const struct user_regset *regset,
 	flush_fp_to_thread(target);
 	flush_altivec_to_thread(target);
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &target->thread.ckvr_state,
-				  0, 33 * sizeof(vector128));
-	if (!ret) {
-		/*
-		 * Copy out only the low-order word of vrsave.
-		 */
-		union {
-			elf_vrreg_t reg;
-			u32 word;
-		} vrsave;
-		memset(&vrsave, 0, sizeof(vrsave));
-		vrsave.word = target->thread.ckvrsave;
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &vrsave,
-					  33 * sizeof(vector128), -1);
-	}
-
-	return ret;
+	membuf_write(&to, &target->thread.ckvr_state, 33 * sizeof(vector128));
+	/*
+	 * Copy out only the low-order word of vrsave.
+	 */
+	memset(&vrsave, 0, sizeof(vrsave));
+	vrsave.word = target->thread.ckvrsave;
+	return membuf_write(&to, &vrsave, sizeof(vrsave));
 }
 
 /**
@@ -484,10 +452,7 @@ int tm_cvsx_active(struct task_struct *target, const struct user_regset *regset)
  * tm_cvsx_get - get CVSX registers
  * @target:	The target task.
  * @regset:	The user regset structure.
- * @pos:	The buffer position.
- * @count:	Number of bytes to copy.
- * @kbuf:	Kernel buffer to copy from.
- * @ubuf:	User buffer to copy into.
+ * @to:		Destination of copy.
  *
  * This function gets in transaction checkpointed VSX registers.
  *
@@ -501,10 +466,10 @@ int tm_cvsx_active(struct task_struct *target, const struct user_regset *regset)
  *};
  */
 int tm_cvsx_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
 	u64 buf[32];
-	int ret, i;
+	int i;
 
 	if (!cpu_has_feature(CPU_FTR_TM))
 		return -ENODEV;
@@ -520,10 +485,7 @@ int tm_cvsx_get(struct task_struct *target, const struct user_regset *regset,
 
 	for (i = 0; i < 32 ; i++)
 		buf[i] = target->thread.ckfp_state.fpr[i][TS_VSRLOWOFFSET];
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  buf, 0, 32 * sizeof(double));
-
-	return ret;
+	return membuf_write(&to, buf, 32 * sizeof(double));
 }
 
 /**
@@ -597,10 +559,7 @@ int tm_spr_active(struct task_struct *target, const struct user_regset *regset)
  * tm_spr_get - get the TM related SPR registers
  * @target:	The target task.
  * @regset:	The user regset structure.
- * @pos:	The buffer position.
- * @count:	Number of bytes to copy.
- * @kbuf:	Kernel buffer to copy from.
- * @ubuf:	User buffer to copy into.
+ * @to:		Destination of copy.
  *
  * This function gets transactional memory related SPR registers.
  * The userspace interface buffer layout is as follows.
@@ -612,10 +571,8 @@ int tm_spr_active(struct task_struct *target, const struct user_regset *regset)
  * };
  */
 int tm_spr_get(struct task_struct *target, const struct user_regset *regset,
-	       unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	       struct membuf to)
 {
-	int ret;
-
 	/* Build tests */
 	BUILD_BUG_ON(TSO(tm_tfhar) + sizeof(u64) != TSO(tm_texasr));
 	BUILD_BUG_ON(TSO(tm_texasr) + sizeof(u64) != TSO(tm_tfiar));
@@ -630,21 +587,11 @@ int tm_spr_get(struct task_struct *target, const struct user_regset *regset,
 	flush_altivec_to_thread(target);
 
 	/* TFHAR register */
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.tm_tfhar, 0, sizeof(u64));
-
+	membuf_write(&to, &target->thread.tm_tfhar, sizeof(u64));
 	/* TEXASR register */
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.tm_texasr, sizeof(u64),
-					  2 * sizeof(u64));
-
+	membuf_write(&to, &target->thread.tm_texasr, sizeof(u64));
 	/* TFIAR register */
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.tm_tfiar,
-					  2 * sizeof(u64), 3 * sizeof(u64));
-	return ret;
+	return membuf_write(&to, &target->thread.tm_tfiar, sizeof(u64));
 }
 
 /**
@@ -714,19 +661,15 @@ int tm_tar_active(struct task_struct *target, const struct user_regset *regset)
 }
 
 int tm_tar_get(struct task_struct *target, const struct user_regset *regset,
-	       unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	       struct membuf to)
 {
-	int ret;
-
 	if (!cpu_has_feature(CPU_FTR_TM))
 		return -ENODEV;
 
 	if (!MSR_TM_ACTIVE(target->thread.regs->msr))
 		return -ENODATA;
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.tm_tar, 0, sizeof(u64));
-	return ret;
+	return membuf_write(&to, &target->thread.tm_tar, sizeof(u64));
 }
 
 int tm_tar_set(struct task_struct *target, const struct user_regset *regset,
@@ -759,19 +702,15 @@ int tm_ppr_active(struct task_struct *target, const struct user_regset *regset)
 
 
 int tm_ppr_get(struct task_struct *target, const struct user_regset *regset,
-	       unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	       struct membuf to)
 {
-	int ret;
-
 	if (!cpu_has_feature(CPU_FTR_TM))
 		return -ENODEV;
 
 	if (!MSR_TM_ACTIVE(target->thread.regs->msr))
 		return -ENODATA;
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.tm_ppr, 0, sizeof(u64));
-	return ret;
+	return membuf_write(&to, &target->thread.tm_ppr, sizeof(u64));
 }
 
 int tm_ppr_set(struct task_struct *target, const struct user_regset *regset,
@@ -803,19 +742,15 @@ int tm_dscr_active(struct task_struct *target, const struct user_regset *regset)
 }
 
 int tm_dscr_get(struct task_struct *target, const struct user_regset *regset,
-		unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
-	int ret;
-
 	if (!cpu_has_feature(CPU_FTR_TM))
 		return -ENODEV;
 
 	if (!MSR_TM_ACTIVE(target->thread.regs->msr))
 		return -ENODATA;
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.tm_dscr, 0, sizeof(u64));
-	return ret;
+	return membuf_write(&to, &target->thread.tm_dscr, sizeof(u64));
 }
 
 int tm_dscr_set(struct task_struct *target, const struct user_regset *regset,
@@ -836,10 +771,11 @@ int tm_dscr_set(struct task_struct *target, const struct user_regset *regset,
 }
 
 int tm_cgpr32_get(struct task_struct *target, const struct user_regset *regset,
-		  unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		  struct membuf to)
 {
-	return gpr32_get_common(target, regset, pos, count, kbuf, ubuf,
+	gpr32_get_common(target, regset, to,
 				&target->thread.ckpt_regs.gpr[0]);
+	return membuf_zero(&to, ELF_NGREG * sizeof(u32));
 }
 
 int tm_cgpr32_set(struct task_struct *target, const struct user_regset *regset,
diff --git a/arch/powerpc/kernel/ptrace/ptrace-view.c b/arch/powerpc/kernel/ptrace/ptrace-view.c
index caeb5822a8f4..9dbd786a1e08 100644
--- a/arch/powerpc/kernel/ptrace/ptrace-view.c
+++ b/arch/powerpc/kernel/ptrace/ptrace-view.c
@@ -215,9 +215,9 @@ int ptrace_put_reg(struct task_struct *task, int regno, unsigned long data)
 }
 
 static int gpr_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	int i, ret;
+	int i;
 
 	if (target->thread.regs == NULL)
 		return -EIO;
@@ -228,30 +228,17 @@ static int gpr_get(struct task_struct *target, const struct user_regset *regset,
 			target->thread.regs->gpr[i] = NV_REG_POISON;
 	}
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  target->thread.regs,
-				  0, offsetof(struct pt_regs, msr));
-	if (!ret) {
-		unsigned long msr = get_user_msr(target);
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &msr,
-					  offsetof(struct pt_regs, msr),
-					  offsetof(struct pt_regs, msr) +
-					  sizeof(msr));
-	}
+	membuf_write(&to, target->thread.regs, offsetof(struct pt_regs, msr));
+	membuf_store(&to, get_user_msr(target));
 
 	BUILD_BUG_ON(offsetof(struct pt_regs, orig_gpr3) !=
 		     offsetof(struct pt_regs, msr) + sizeof(long));
 
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.regs->orig_gpr3,
-					  offsetof(struct pt_regs, orig_gpr3),
-					  sizeof(struct user_pt_regs));
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       sizeof(struct user_pt_regs), -1);
-
-	return ret;
+	membuf_write(&to, &target->thread.regs->orig_gpr3,
+			sizeof(struct user_pt_regs) -
+			offsetof(struct pt_regs, orig_gpr3));
+	return membuf_zero(&to, ELF_NGREG * sizeof(unsigned long) -
+				 sizeof(struct user_pt_regs));
 }
 
 static int gpr_set(struct task_struct *target, const struct user_regset *regset,
@@ -309,10 +296,9 @@ static int gpr_set(struct task_struct *target, const struct user_regset *regset,
 
 #ifdef CONFIG_PPC64
 static int ppr_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &target->thread.regs->ppr, 0, sizeof(u64));
+	return membuf_write(&to, &target->thread.regs->ppr, sizeof(u64));
 }
 
 static int ppr_set(struct task_struct *target, const struct user_regset *regset,
@@ -324,10 +310,9 @@ static int ppr_set(struct task_struct *target, const struct user_regset *regset,
 }
 
 static int dscr_get(struct task_struct *target, const struct user_regset *regset,
-		    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		    struct membuf to)
 {
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &target->thread.dscr, 0, sizeof(u64));
+	return membuf_write(&to, &target->thread.dscr, sizeof(u64));
 }
 static int dscr_set(struct task_struct *target, const struct user_regset *regset,
 		    unsigned int pos, unsigned int count, const void *kbuf,
@@ -339,10 +324,9 @@ static int dscr_set(struct task_struct *target, const struct user_regset *regset
 #endif
 #ifdef CONFIG_PPC_BOOK3S_64
 static int tar_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &target->thread.tar, 0, sizeof(u64));
+	return membuf_write(&to, &target->thread.tar, sizeof(u64));
 }
 static int tar_set(struct task_struct *target, const struct user_regset *regset,
 		   unsigned int pos, unsigned int count, const void *kbuf,
@@ -364,7 +348,7 @@ static int ebb_active(struct task_struct *target, const struct user_regset *regs
 }
 
 static int ebb_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	/* Build tests */
 	BUILD_BUG_ON(TSO(ebbrr) + sizeof(unsigned long) != TSO(ebbhr));
@@ -376,8 +360,7 @@ static int ebb_get(struct task_struct *target, const struct user_regset *regset,
 	if (!target->thread.used_ebb)
 		return -ENODATA;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &target->thread.ebbrr,
-				   0, 3 * sizeof(unsigned long));
+	return membuf_write(&to, &target->thread.ebbrr, 3 * sizeof(unsigned long));
 }
 
 static int ebb_set(struct task_struct *target, const struct user_regset *regset,
@@ -420,7 +403,7 @@ static int pmu_active(struct task_struct *target, const struct user_regset *regs
 }
 
 static int pmu_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	/* Build tests */
 	BUILD_BUG_ON(TSO(siar) + sizeof(unsigned long) != TSO(sdar));
@@ -431,8 +414,7 @@ static int pmu_get(struct task_struct *target, const struct user_regset *regset,
 	if (!cpu_has_feature(CPU_FTR_ARCH_207S))
 		return -ENODEV;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &target->thread.siar,
-				   0, 5 * sizeof(unsigned long));
+	return membuf_write(&to, &target->thread.siar, 5 * sizeof(unsigned long));
 }
 
 static int pmu_set(struct task_struct *target, const struct user_regset *regset,
@@ -486,7 +468,7 @@ static int pkey_active(struct task_struct *target, const struct user_regset *reg
 }
 
 static int pkey_get(struct task_struct *target, const struct user_regset *regset,
-		    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+		    struct membuf to)
 {
 	BUILD_BUG_ON(TSO(amr) + sizeof(unsigned long) != TSO(iamr));
 	BUILD_BUG_ON(TSO(iamr) + sizeof(unsigned long) != TSO(uamor));
@@ -494,8 +476,7 @@ static int pkey_get(struct task_struct *target, const struct user_regset *regset
 	if (!arch_pkeys_enabled())
 		return -ENODEV;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &target->thread.amr,
-				   0, ELF_NPKEY * sizeof(unsigned long));
+	return membuf_write(&to, &target->thread.amr, ELF_NPKEY * sizeof(unsigned long));
 }
 
 static int pkey_set(struct task_struct *target, const struct user_regset *regset,
@@ -529,110 +510,110 @@ static const struct user_regset native_regsets[] = {
 	[REGSET_GPR] = {
 		.core_note_type = NT_PRSTATUS, .n = ELF_NGREG,
 		.size = sizeof(long), .align = sizeof(long),
-		.get = gpr_get, .set = gpr_set
+		.get2 = gpr_get, .set = gpr_set
 	},
 	[REGSET_FPR] = {
 		.core_note_type = NT_PRFPREG, .n = ELF_NFPREG,
 		.size = sizeof(double), .align = sizeof(double),
-		.get = fpr_get, .set = fpr_set
+		.get2 = fpr_get, .set = fpr_set
 	},
 #ifdef CONFIG_ALTIVEC
 	[REGSET_VMX] = {
 		.core_note_type = NT_PPC_VMX, .n = 34,
 		.size = sizeof(vector128), .align = sizeof(vector128),
-		.active = vr_active, .get = vr_get, .set = vr_set
+		.active = vr_active, .get2 = vr_get, .set = vr_set
 	},
 #endif
 #ifdef CONFIG_VSX
 	[REGSET_VSX] = {
 		.core_note_type = NT_PPC_VSX, .n = 32,
 		.size = sizeof(double), .align = sizeof(double),
-		.active = vsr_active, .get = vsr_get, .set = vsr_set
+		.active = vsr_active, .get2 = vsr_get, .set = vsr_set
 	},
 #endif
 #ifdef CONFIG_SPE
 	[REGSET_SPE] = {
 		.core_note_type = NT_PPC_SPE, .n = 35,
 		.size = sizeof(u32), .align = sizeof(u32),
-		.active = evr_active, .get = evr_get, .set = evr_set
+		.active = evr_active, .get2 = evr_get, .set = evr_set
 	},
 #endif
 #ifdef CONFIG_PPC_TRANSACTIONAL_MEM
 	[REGSET_TM_CGPR] = {
 		.core_note_type = NT_PPC_TM_CGPR, .n = ELF_NGREG,
 		.size = sizeof(long), .align = sizeof(long),
-		.active = tm_cgpr_active, .get = tm_cgpr_get, .set = tm_cgpr_set
+		.active = tm_cgpr_active, .get2 = tm_cgpr_get, .set = tm_cgpr_set
 	},
 	[REGSET_TM_CFPR] = {
 		.core_note_type = NT_PPC_TM_CFPR, .n = ELF_NFPREG,
 		.size = sizeof(double), .align = sizeof(double),
-		.active = tm_cfpr_active, .get = tm_cfpr_get, .set = tm_cfpr_set
+		.active = tm_cfpr_active, .get2 = tm_cfpr_get, .set = tm_cfpr_set
 	},
 	[REGSET_TM_CVMX] = {
 		.core_note_type = NT_PPC_TM_CVMX, .n = ELF_NVMX,
 		.size = sizeof(vector128), .align = sizeof(vector128),
-		.active = tm_cvmx_active, .get = tm_cvmx_get, .set = tm_cvmx_set
+		.active = tm_cvmx_active, .get2 = tm_cvmx_get, .set = tm_cvmx_set
 	},
 	[REGSET_TM_CVSX] = {
 		.core_note_type = NT_PPC_TM_CVSX, .n = ELF_NVSX,
 		.size = sizeof(double), .align = sizeof(double),
-		.active = tm_cvsx_active, .get = tm_cvsx_get, .set = tm_cvsx_set
+		.active = tm_cvsx_active, .get2 = tm_cvsx_get, .set = tm_cvsx_set
 	},
 	[REGSET_TM_SPR] = {
 		.core_note_type = NT_PPC_TM_SPR, .n = ELF_NTMSPRREG,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_spr_active, .get = tm_spr_get, .set = tm_spr_set
+		.active = tm_spr_active, .get2 = tm_spr_get, .set = tm_spr_set
 	},
 	[REGSET_TM_CTAR] = {
 		.core_note_type = NT_PPC_TM_CTAR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_tar_active, .get = tm_tar_get, .set = tm_tar_set
+		.active = tm_tar_active, .get2 = tm_tar_get, .set = tm_tar_set
 	},
 	[REGSET_TM_CPPR] = {
 		.core_note_type = NT_PPC_TM_CPPR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_ppr_active, .get = tm_ppr_get, .set = tm_ppr_set
+		.active = tm_ppr_active, .get2 = tm_ppr_get, .set = tm_ppr_set
 	},
 	[REGSET_TM_CDSCR] = {
 		.core_note_type = NT_PPC_TM_CDSCR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_dscr_active, .get = tm_dscr_get, .set = tm_dscr_set
+		.active = tm_dscr_active, .get2 = tm_dscr_get, .set = tm_dscr_set
 	},
 #endif
 #ifdef CONFIG_PPC64
 	[REGSET_PPR] = {
 		.core_note_type = NT_PPC_PPR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = ppr_get, .set = ppr_set
+		.get2 = ppr_get, .set = ppr_set
 	},
 	[REGSET_DSCR] = {
 		.core_note_type = NT_PPC_DSCR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = dscr_get, .set = dscr_set
+		.get2 = dscr_get, .set = dscr_set
 	},
 #endif
 #ifdef CONFIG_PPC_BOOK3S_64
 	[REGSET_TAR] = {
 		.core_note_type = NT_PPC_TAR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = tar_get, .set = tar_set
+		.get2 = tar_get, .set = tar_set
 	},
 	[REGSET_EBB] = {
 		.core_note_type = NT_PPC_EBB, .n = ELF_NEBB,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = ebb_active, .get = ebb_get, .set = ebb_set
+		.active = ebb_active, .get2 = ebb_get, .set = ebb_set
 	},
 	[REGSET_PMR] = {
 		.core_note_type = NT_PPC_PMU, .n = ELF_NPMU,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = pmu_active, .get = pmu_get, .set = pmu_set
+		.active = pmu_active, .get2 = pmu_get, .set = pmu_set
 	},
 #endif
 #ifdef CONFIG_PPC_MEM_KEYS
 	[REGSET_PKEY] = {
 		.core_note_type = NT_PPC_PKEY, .n = ELF_NPKEY,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = pkey_active, .get = pkey_get, .set = pkey_set
+		.active = pkey_active, .get2 = pkey_get, .set = pkey_set
 	},
 #endif
 };
@@ -646,49 +627,16 @@ const struct user_regset_view user_ppc_native_view = {
 
 int gpr32_get_common(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-			    void *kbuf, void __user *ubuf,
-			    unsigned long *regs)
+		     struct membuf to, unsigned long *regs)
 {
-	compat_ulong_t *k = kbuf;
-	compat_ulong_t __user *u = ubuf;
-	compat_ulong_t reg;
-
-	pos /= sizeof(reg);
-	count /= sizeof(reg);
-
-	if (kbuf)
-		for (; count > 0 && pos < PT_MSR; --count)
-			*k++ = regs[pos++];
-	else
-		for (; count > 0 && pos < PT_MSR; --count)
-			if (__put_user((compat_ulong_t)regs[pos++], u++))
-				return -EFAULT;
-
-	if (count > 0 && pos == PT_MSR) {
-		reg = get_user_msr(target);
-		if (kbuf)
-			*k++ = reg;
-		else if (__put_user(reg, u++))
-			return -EFAULT;
-		++pos;
-		--count;
-	}
-
-	if (kbuf)
-		for (; count > 0 && pos < PT_REGS_COUNT; --count)
-			*k++ = regs[pos++];
-	else
-		for (; count > 0 && pos < PT_REGS_COUNT; --count)
-			if (__put_user((compat_ulong_t)regs[pos++], u++))
-				return -EFAULT;
+	int i;
 
-	kbuf = k;
-	ubuf = u;
-	pos *= sizeof(reg);
-	count *= sizeof(reg);
-	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					PT_REGS_COUNT * sizeof(reg), -1);
+	for (i = 0; i < PT_MSR; i++)
+		membuf_store(&to, (u32)regs[i]);
+	membuf_store(&to, (u32)get_user_msr(target));
+	for (i++ ; i < PT_REGS_COUNT; i++)
+		membuf_store(&to, (u32)regs[i]);
+	return membuf_zero(&to, (ELF_NGREG - PT_REGS_COUNT) * sizeof(u32));
 }
 
 int gpr32_set_common(struct task_struct *target,
@@ -761,8 +709,7 @@ int gpr32_set_common(struct task_struct *target,
 
 static int gpr32_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
 	int i;
 
@@ -777,7 +724,7 @@ static int gpr32_get(struct task_struct *target,
 		for (i = 14; i < 32; i++)
 			target->thread.regs->gpr[i] = NV_REG_POISON;
 	}
-	return gpr32_get_common(target, regset, pos, count, kbuf, ubuf,
+	return gpr32_get_common(target, regset, to,
 			&target->thread.regs->gpr[0]);
 }
 
@@ -801,25 +748,25 @@ static const struct user_regset compat_regsets[] = {
 	[REGSET_GPR] = {
 		.core_note_type = NT_PRSTATUS, .n = ELF_NGREG,
 		.size = sizeof(compat_long_t), .align = sizeof(compat_long_t),
-		.get = gpr32_get, .set = gpr32_set
+		.get2 = gpr32_get, .set = gpr32_set
 	},
 	[REGSET_FPR] = {
 		.core_note_type = NT_PRFPREG, .n = ELF_NFPREG,
 		.size = sizeof(double), .align = sizeof(double),
-		.get = fpr_get, .set = fpr_set
+		.get2 = fpr_get, .set = fpr_set
 	},
 #ifdef CONFIG_ALTIVEC
 	[REGSET_VMX] = {
 		.core_note_type = NT_PPC_VMX, .n = 34,
 		.size = sizeof(vector128), .align = sizeof(vector128),
-		.active = vr_active, .get = vr_get, .set = vr_set
+		.active = vr_active, .get2 = vr_get, .set = vr_set
 	},
 #endif
 #ifdef CONFIG_SPE
 	[REGSET_SPE] = {
 		.core_note_type = NT_PPC_SPE, .n = 35,
 		.size = sizeof(u32), .align = sizeof(u32),
-		.active = evr_active, .get = evr_get, .set = evr_set
+		.active = evr_active, .get2 = evr_get, .set = evr_set
 	},
 #endif
 #ifdef CONFIG_PPC_TRANSACTIONAL_MEM
@@ -827,66 +774,66 @@ static const struct user_regset compat_regsets[] = {
 		.core_note_type = NT_PPC_TM_CGPR, .n = ELF_NGREG,
 		.size = sizeof(long), .align = sizeof(long),
 		.active = tm_cgpr_active,
-		.get = tm_cgpr32_get, .set = tm_cgpr32_set
+		.get2 = tm_cgpr32_get, .set = tm_cgpr32_set
 	},
 	[REGSET_TM_CFPR] = {
 		.core_note_type = NT_PPC_TM_CFPR, .n = ELF_NFPREG,
 		.size = sizeof(double), .align = sizeof(double),
-		.active = tm_cfpr_active, .get = tm_cfpr_get, .set = tm_cfpr_set
+		.active = tm_cfpr_active, .get2 = tm_cfpr_get, .set = tm_cfpr_set
 	},
 	[REGSET_TM_CVMX] = {
 		.core_note_type = NT_PPC_TM_CVMX, .n = ELF_NVMX,
 		.size = sizeof(vector128), .align = sizeof(vector128),
-		.active = tm_cvmx_active, .get = tm_cvmx_get, .set = tm_cvmx_set
+		.active = tm_cvmx_active, .get2 = tm_cvmx_get, .set = tm_cvmx_set
 	},
 	[REGSET_TM_CVSX] = {
 		.core_note_type = NT_PPC_TM_CVSX, .n = ELF_NVSX,
 		.size = sizeof(double), .align = sizeof(double),
-		.active = tm_cvsx_active, .get = tm_cvsx_get, .set = tm_cvsx_set
+		.active = tm_cvsx_active, .get2 = tm_cvsx_get, .set = tm_cvsx_set
 	},
 	[REGSET_TM_SPR] = {
 		.core_note_type = NT_PPC_TM_SPR, .n = ELF_NTMSPRREG,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_spr_active, .get = tm_spr_get, .set = tm_spr_set
+		.active = tm_spr_active, .get2 = tm_spr_get, .set = tm_spr_set
 	},
 	[REGSET_TM_CTAR] = {
 		.core_note_type = NT_PPC_TM_CTAR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_tar_active, .get = tm_tar_get, .set = tm_tar_set
+		.active = tm_tar_active, .get2 = tm_tar_get, .set = tm_tar_set
 	},
 	[REGSET_TM_CPPR] = {
 		.core_note_type = NT_PPC_TM_CPPR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_ppr_active, .get = tm_ppr_get, .set = tm_ppr_set
+		.active = tm_ppr_active, .get2 = tm_ppr_get, .set = tm_ppr_set
 	},
 	[REGSET_TM_CDSCR] = {
 		.core_note_type = NT_PPC_TM_CDSCR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = tm_dscr_active, .get = tm_dscr_get, .set = tm_dscr_set
+		.active = tm_dscr_active, .get2 = tm_dscr_get, .set = tm_dscr_set
 	},
 #endif
 #ifdef CONFIG_PPC64
 	[REGSET_PPR] = {
 		.core_note_type = NT_PPC_PPR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = ppr_get, .set = ppr_set
+		.get2 = ppr_get, .set = ppr_set
 	},
 	[REGSET_DSCR] = {
 		.core_note_type = NT_PPC_DSCR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = dscr_get, .set = dscr_set
+		.get2 = dscr_get, .set = dscr_set
 	},
 #endif
 #ifdef CONFIG_PPC_BOOK3S_64
 	[REGSET_TAR] = {
 		.core_note_type = NT_PPC_TAR, .n = 1,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = tar_get, .set = tar_set
+		.get2 = tar_get, .set = tar_set
 	},
 	[REGSET_EBB] = {
 		.core_note_type = NT_PPC_EBB, .n = ELF_NEBB,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.active = ebb_active, .get = ebb_get, .set = ebb_set
+		.active = ebb_active, .get2 = ebb_get, .set = ebb_set
 	},
 #endif
 };
diff --git a/arch/powerpc/kernel/ptrace/ptrace-vsx.c b/arch/powerpc/kernel/ptrace/ptrace-vsx.c
index d53466d49cc0..1da4303128ef 100644
--- a/arch/powerpc/kernel/ptrace/ptrace-vsx.c
+++ b/arch/powerpc/kernel/ptrace/ptrace-vsx.c
@@ -19,7 +19,7 @@
  * };
  */
 int fpr_get(struct task_struct *target, const struct user_regset *regset,
-	    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	    struct membuf to)
 {
 	u64 buf[33];
 	int i;
@@ -30,7 +30,7 @@ int fpr_get(struct task_struct *target, const struct user_regset *regset,
 	for (i = 0; i < 32 ; i++)
 		buf[i] = target->thread.TS_FPR(i);
 	buf[32] = target->thread.fp_state.fpscr;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, buf, 0, -1);
+	return membuf_write(&to, buf, 33 * sizeof(u64));
 }
 
 /*
@@ -95,10 +95,10 @@ int vsr_active(struct task_struct *target, const struct user_regset *regset)
  * };
  */
 int vsr_get(struct task_struct *target, const struct user_regset *regset,
-	    unsigned int pos, unsigned int count, void *kbuf, void __user *ubuf)
+	    struct membuf to)
 {
 	u64 buf[32];
-	int ret, i;
+	int i;
 
 	flush_tmregs_to_thread(target);
 	flush_fp_to_thread(target);
@@ -108,10 +108,7 @@ int vsr_get(struct task_struct *target, const struct user_regset *regset,
 	for (i = 0; i < 32 ; i++)
 		buf[i] = target->thread.fp_state.fpr[i][TS_VSRLOWOFFSET];
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  buf, 0, 32 * sizeof(double));
-
-	return ret;
+	return membuf_write(&to, buf, 32 * sizeof(double));
 }
 
 /*
-- 
2.11.0

[PATCH 21/41] s390: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

NB: compat NT_S390_LAST_BREAK might be better as compat_long_t
rather than long.  User-visible ABI, again...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/s390/kernel/ptrace.c | 199 ++++++++++++++--------------------------------
 1 file changed, 58 insertions(+), 141 deletions(-)

diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c
index ce60a459a143..421bfec0a9e9 100644
--- a/arch/s390/kernel/ptrace.c
+++ b/arch/s390/kernel/ptrace.c
@@ -890,28 +890,14 @@ asmlinkage void do_syscall_trace_exit(struct pt_regs *regs)
 
 static int s390_regs_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
+	unsigned pos;
 	if (target == current)
 		save_access_regs(target->thread.acrs);
 
-	if (kbuf) {
-		unsigned long *k = kbuf;
-		while (count > 0) {
-			*k++ = __peek_user(target, pos);
-			count -= sizeof(*k);
-			pos += sizeof(*k);
-		}
-	} else {
-		unsigned long __user *u = ubuf;
-		while (count > 0) {
-			if (__put_user(__peek_user(target, pos), u++))
-				return -EFAULT;
-			count -= sizeof(*u);
-			pos += sizeof(*u);
-		}
-	}
+	for (pos = 0; pos < sizeof(s390_regs); pos += sizeof(long))
+		membuf_store(&to, __peek_user(target, pos));
 	return 0;
 }
 
@@ -952,8 +938,8 @@ static int s390_regs_set(struct task_struct *target,
 }
 
 static int s390_fpregs_get(struct task_struct *target,
-			   const struct user_regset *regset, unsigned int pos,
-			   unsigned int count, void *kbuf, void __user *ubuf)
+			   const struct user_regset *regset,
+			   struct membuf to)
 {
 	_s390_fp_regs fp_regs;
 
@@ -963,8 +949,7 @@ static int s390_fpregs_get(struct task_struct *target,
 	fp_regs.fpc = target->thread.fpu.fpc;
 	fpregs_store(&fp_regs, &target->thread.fpu);
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &fp_regs, 0, -1);
+	return membuf_write(&to, &fp_regs, sizeof(fp_regs));
 }
 
 static int s390_fpregs_set(struct task_struct *target,
@@ -1011,20 +996,9 @@ static int s390_fpregs_set(struct task_struct *target,
 
 static int s390_last_break_get(struct task_struct *target,
 			       const struct user_regset *regset,
-			       unsigned int pos, unsigned int count,
-			       void *kbuf, void __user *ubuf)
+			       struct membuf to)
 {
-	if (count > 0) {
-		if (kbuf) {
-			unsigned long *k = kbuf;
-			*k = target->thread.last_break;
-		} else {
-			unsigned long  __user *u = ubuf;
-			if (__put_user(target->thread.last_break, u))
-				return -EFAULT;
-		}
-	}
-	return 0;
+	return membuf_store(&to, target->thread.last_break);
 }
 
 static int s390_last_break_set(struct task_struct *target,
@@ -1037,16 +1011,13 @@ static int s390_last_break_set(struct task_struct *target,
 
 static int s390_tdb_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
 	struct pt_regs *regs = task_pt_regs(target);
-	unsigned char *data;
 
 	if (!(regs->int_code & 0x200))
 		return -ENODATA;
-	data = target->thread.trap_tdb;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, data, 0, 256);
+	return membuf_write(&to, target->thread.trap_tdb, 256);
 }
 
 static int s390_tdb_set(struct task_struct *target,
@@ -1059,8 +1030,7 @@ static int s390_tdb_set(struct task_struct *target,
 
 static int s390_vxrs_low_get(struct task_struct *target,
 			     const struct user_regset *regset,
-			     unsigned int pos, unsigned int count,
-			     void *kbuf, void __user *ubuf)
+			     struct membuf to)
 {
 	__u64 vxrs[__NUM_VXRS_LOW];
 	int i;
@@ -1071,7 +1041,7 @@ static int s390_vxrs_low_get(struct task_struct *target,
 		save_fpu_regs();
 	for (i = 0; i < __NUM_VXRS_LOW; i++)
 		vxrs[i] = *((__u64 *)(target->thread.fpu.vxrs + i) + 1);
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, vxrs, 0, -1);
+	return membuf_write(&to, vxrs, sizeof(vxrs));
 }
 
 static int s390_vxrs_low_set(struct task_struct *target,
@@ -1100,18 +1070,14 @@ static int s390_vxrs_low_set(struct task_struct *target,
 
 static int s390_vxrs_high_get(struct task_struct *target,
 			      const struct user_regset *regset,
-			      unsigned int pos, unsigned int count,
-			      void *kbuf, void __user *ubuf)
+			      struct membuf to)
 {
-	__vector128 vxrs[__NUM_VXRS_HIGH];
-
 	if (!MACHINE_HAS_VX)
 		return -ENODEV;
 	if (target == current)
 		save_fpu_regs();
-	memcpy(vxrs, target->thread.fpu.vxrs + __NUM_VXRS_LOW, sizeof(vxrs));
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, vxrs, 0, -1);
+	return membuf_write(&to, target->thread.fpu.vxrs + __NUM_VXRS_LOW,
+			    __NUM_VXRS_HIGH * sizeof(__vector128));
 }
 
 static int s390_vxrs_high_set(struct task_struct *target,
@@ -1133,12 +1099,9 @@ static int s390_vxrs_high_set(struct task_struct *target,
 
 static int s390_system_call_get(struct task_struct *target,
 				const struct user_regset *regset,
-				unsigned int pos, unsigned int count,
-				void *kbuf, void __user *ubuf)
+				struct membuf to)
 {
-	unsigned int *data = &target->thread.system_call;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   data, 0, sizeof(unsigned int));
+	return membuf_store(&to, target->thread.system_call);
 }
 
 static int s390_system_call_set(struct task_struct *target,
@@ -1153,8 +1116,7 @@ static int s390_system_call_set(struct task_struct *target,
 
 static int s390_gs_cb_get(struct task_struct *target,
 			  const struct user_regset *regset,
-			  unsigned int pos, unsigned int count,
-			  void *kbuf, void __user *ubuf)
+			  struct membuf to)
 {
 	struct gs_cb *data = target->thread.gs_cb;
 
@@ -1164,8 +1126,7 @@ static int s390_gs_cb_get(struct task_struct *target,
 		return -ENODATA;
 	if (target == current)
 		save_gs_cb(data);
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   data, 0, sizeof(struct gs_cb));
+	return membuf_write(&to, data, sizeof(struct gs_cb));
 }
 
 static int s390_gs_cb_set(struct task_struct *target,
@@ -1209,8 +1170,7 @@ static int s390_gs_cb_set(struct task_struct *target,
 
 static int s390_gs_bc_get(struct task_struct *target,
 			  const struct user_regset *regset,
-			  unsigned int pos, unsigned int count,
-			  void *kbuf, void __user *ubuf)
+			  struct membuf to)
 {
 	struct gs_cb *data = target->thread.gs_bc_cb;
 
@@ -1218,8 +1178,7 @@ static int s390_gs_bc_get(struct task_struct *target,
 		return -ENODEV;
 	if (!data)
 		return -ENODATA;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   data, 0, sizeof(struct gs_cb));
+	return membuf_write(&to, data, sizeof(struct gs_cb));
 }
 
 static int s390_gs_bc_set(struct task_struct *target,
@@ -1270,8 +1229,7 @@ static bool is_ri_cb_valid(struct runtime_instr_cb *cb)
 
 static int s390_runtime_instr_get(struct task_struct *target,
 				const struct user_regset *regset,
-				unsigned int pos, unsigned int count,
-				void *kbuf, void __user *ubuf)
+				struct membuf to)
 {
 	struct runtime_instr_cb *data = target->thread.ri_cb;
 
@@ -1280,8 +1238,7 @@ static int s390_runtime_instr_get(struct task_struct *target,
 	if (!data)
 		return -ENODATA;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   data, 0, sizeof(struct runtime_instr_cb));
+	return membuf_write(&to, data, sizeof(struct runtime_instr_cb));
 }
 
 static int s390_runtime_instr_set(struct task_struct *target,
@@ -1337,7 +1294,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = sizeof(s390_regs) / sizeof(long),
 		.size = sizeof(long),
 		.align = sizeof(long),
-		.get = s390_regs_get,
+		.get2 = s390_regs_get,
 		.set = s390_regs_set,
 	},
 	{
@@ -1345,7 +1302,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = sizeof(s390_fp_regs) / sizeof(long),
 		.size = sizeof(long),
 		.align = sizeof(long),
-		.get = s390_fpregs_get,
+		.get2 = s390_fpregs_get,
 		.set = s390_fpregs_set,
 	},
 	{
@@ -1353,7 +1310,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = 1,
 		.size = sizeof(unsigned int),
 		.align = sizeof(unsigned int),
-		.get = s390_system_call_get,
+		.get2 = s390_system_call_get,
 		.set = s390_system_call_set,
 	},
 	{
@@ -1361,7 +1318,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = 1,
 		.size = sizeof(long),
 		.align = sizeof(long),
-		.get = s390_last_break_get,
+		.get2 = s390_last_break_get,
 		.set = s390_last_break_set,
 	},
 	{
@@ -1369,7 +1326,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = 1,
 		.size = 256,
 		.align = 1,
-		.get = s390_tdb_get,
+		.get2 = s390_tdb_get,
 		.set = s390_tdb_set,
 	},
 	{
@@ -1377,7 +1334,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = __NUM_VXRS_LOW,
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_vxrs_low_get,
+		.get2 = s390_vxrs_low_get,
 		.set = s390_vxrs_low_set,
 	},
 	{
@@ -1385,7 +1342,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = __NUM_VXRS_HIGH,
 		.size = sizeof(__vector128),
 		.align = sizeof(__vector128),
-		.get = s390_vxrs_high_get,
+		.get2 = s390_vxrs_high_get,
 		.set = s390_vxrs_high_set,
 	},
 	{
@@ -1393,7 +1350,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = sizeof(struct gs_cb) / sizeof(__u64),
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_gs_cb_get,
+		.get2 = s390_gs_cb_get,
 		.set = s390_gs_cb_set,
 	},
 	{
@@ -1401,7 +1358,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = sizeof(struct gs_cb) / sizeof(__u64),
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_gs_bc_get,
+		.get2 = s390_gs_bc_get,
 		.set = s390_gs_bc_set,
 	},
 	{
@@ -1409,7 +1366,7 @@ static const struct user_regset s390_regsets[] = {
 		.n = sizeof(struct runtime_instr_cb) / sizeof(__u64),
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_runtime_instr_get,
+		.get2 = s390_runtime_instr_get,
 		.set = s390_runtime_instr_set,
 	},
 };
@@ -1424,28 +1381,15 @@ static const struct user_regset_view user_s390_view = {
 #ifdef CONFIG_COMPAT
 static int s390_compat_regs_get(struct task_struct *target,
 				const struct user_regset *regset,
-				unsigned int pos, unsigned int count,
-				void *kbuf, void __user *ubuf)
+				struct membuf to)
 {
+	unsigned n;
+
 	if (target == current)
 		save_access_regs(target->thread.acrs);
 
-	if (kbuf) {
-		compat_ulong_t *k = kbuf;
-		while (count > 0) {
-			*k++ = __peek_user_compat(target, pos);
-			count -= sizeof(*k);
-			pos += sizeof(*k);
-		}
-	} else {
-		compat_ulong_t __user *u = ubuf;
-		while (count > 0) {
-			if (__put_user(__peek_user_compat(target, pos), u++))
-				return -EFAULT;
-			count -= sizeof(*u);
-			pos += sizeof(*u);
-		}
-	}
+	for (n = 0; n < sizeof(s390_compat_regs); n += sizeof(compat_ulong_t))
+		membuf_store(&to, __peek_user_compat(target, n));
 	return 0;
 }
 
@@ -1487,29 +1431,14 @@ static int s390_compat_regs_set(struct task_struct *target,
 
 static int s390_compat_regs_high_get(struct task_struct *target,
 				     const struct user_regset *regset,
-				     unsigned int pos, unsigned int count,
-				     void *kbuf, void __user *ubuf)
+				     struct membuf to)
 {
 	compat_ulong_t *gprs_high;
+	int i;
 
-	gprs_high = (compat_ulong_t *)
-		&task_pt_regs(target)->gprs[pos / sizeof(compat_ulong_t)];
-	if (kbuf) {
-		compat_ulong_t *k = kbuf;
-		while (count > 0) {
-			*k++ = *gprs_high;
-			gprs_high += 2;
-			count -= sizeof(*k);
-		}
-	} else {
-		compat_ulong_t __user *u = ubuf;
-		while (count > 0) {
-			if (__put_user(*gprs_high, u++))
-				return -EFAULT;
-			gprs_high += 2;
-			count -= sizeof(*u);
-		}
-	}
+	gprs_high = (compat_ulong_t *)task_pt_regs(target)->gprs;
+	for (i = 0; i < NUM_GPRS; i++, gprs_high += 2)
+		membuf_store(&to, *gprs_high);
 	return 0;
 }
 
@@ -1548,23 +1477,11 @@ static int s390_compat_regs_high_set(struct task_struct *target,
 
 static int s390_compat_last_break_get(struct task_struct *target,
 				      const struct user_regset *regset,
-				      unsigned int pos, unsigned int count,
-				      void *kbuf, void __user *ubuf)
+				      struct membuf to)
 {
-	compat_ulong_t last_break;
+	compat_ulong_t last_break = target->thread.last_break;
 
-	if (count > 0) {
-		last_break = target->thread.last_break;
-		if (kbuf) {
-			unsigned long *k = kbuf;
-			*k = last_break;
-		} else {
-			unsigned long  __user *u = ubuf;
-			if (__put_user(last_break, u))
-				return -EFAULT;
-		}
-	}
-	return 0;
+	return membuf_store(&to, (unsigned long)last_break);
 }
 
 static int s390_compat_last_break_set(struct task_struct *target,
@@ -1581,7 +1498,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = sizeof(s390_compat_regs) / sizeof(compat_long_t),
 		.size = sizeof(compat_long_t),
 		.align = sizeof(compat_long_t),
-		.get = s390_compat_regs_get,
+		.get2 = s390_compat_regs_get,
 		.set = s390_compat_regs_set,
 	},
 	{
@@ -1589,7 +1506,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = sizeof(s390_fp_regs) / sizeof(compat_long_t),
 		.size = sizeof(compat_long_t),
 		.align = sizeof(compat_long_t),
-		.get = s390_fpregs_get,
+		.get2 = s390_fpregs_get,
 		.set = s390_fpregs_set,
 	},
 	{
@@ -1597,7 +1514,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = 1,
 		.size = sizeof(compat_uint_t),
 		.align = sizeof(compat_uint_t),
-		.get = s390_system_call_get,
+		.get2 = s390_system_call_get,
 		.set = s390_system_call_set,
 	},
 	{
@@ -1605,7 +1522,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = 1,
 		.size = sizeof(long),
 		.align = sizeof(long),
-		.get = s390_compat_last_break_get,
+		.get2 = s390_compat_last_break_get,
 		.set = s390_compat_last_break_set,
 	},
 	{
@@ -1613,7 +1530,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = 1,
 		.size = 256,
 		.align = 1,
-		.get = s390_tdb_get,
+		.get2 = s390_tdb_get,
 		.set = s390_tdb_set,
 	},
 	{
@@ -1621,7 +1538,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = __NUM_VXRS_LOW,
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_vxrs_low_get,
+		.get2 = s390_vxrs_low_get,
 		.set = s390_vxrs_low_set,
 	},
 	{
@@ -1629,7 +1546,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = __NUM_VXRS_HIGH,
 		.size = sizeof(__vector128),
 		.align = sizeof(__vector128),
-		.get = s390_vxrs_high_get,
+		.get2 = s390_vxrs_high_get,
 		.set = s390_vxrs_high_set,
 	},
 	{
@@ -1637,7 +1554,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = sizeof(s390_compat_regs_high) / sizeof(compat_long_t),
 		.size = sizeof(compat_long_t),
 		.align = sizeof(compat_long_t),
-		.get = s390_compat_regs_high_get,
+		.get2 = s390_compat_regs_high_get,
 		.set = s390_compat_regs_high_set,
 	},
 	{
@@ -1645,7 +1562,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = sizeof(struct gs_cb) / sizeof(__u64),
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_gs_cb_get,
+		.get2 = s390_gs_cb_get,
 		.set = s390_gs_cb_set,
 	},
 	{
@@ -1653,7 +1570,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = sizeof(struct gs_cb) / sizeof(__u64),
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_gs_bc_get,
+		.get2 = s390_gs_bc_get,
 		.set = s390_gs_bc_set,
 	},
 	{
@@ -1661,7 +1578,7 @@ static const struct user_regset s390_compat_regsets[] = {
 		.n = sizeof(struct runtime_instr_cb) / sizeof(__u64),
 		.size = sizeof(__u64),
 		.align = sizeof(__u64),
-		.get = s390_runtime_instr_get,
+		.get2 = s390_runtime_instr_get,
 		.set = s390_runtime_instr_set,
 	},
 };
-- 
2.11.0

[PATCH 22/41] sparc: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/sparc/kernel/ptrace_32.c | 127 ++++-----------
 arch/sparc/kernel/ptrace_64.c | 352 ++++++++++--------------------------------
 2 files changed, 110 insertions(+), 369 deletions(-)

diff --git a/arch/sparc/kernel/ptrace_32.c b/arch/sparc/kernel/ptrace_32.c
index 0856e0104539..948a1dcd8e98 100644
--- a/arch/sparc/kernel/ptrace_32.c
+++ b/arch/sparc/kernel/ptrace_32.c
@@ -83,39 +83,25 @@ static int regwindow32_set(struct task_struct *target,
 
 static int genregs32_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
 	const struct pt_regs *regs = target->thread.kregs;
 	u32 uregs[16];
-	int ret;
 
 	if (target == current)
 		flush_user_windows();
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  regs->u_regs,
-				  0, 16 * sizeof(u32));
-	if (ret || !count)
-		return ret;
-
+	membuf_write(&to, regs->u_regs, 16 * sizeof(u32));
+	if (!to.left)
+		return 0;
 	if (regwindow32_get(target, regs, uregs))
 		return -EFAULT;
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  uregs,
-				  16 * sizeof(u32), 32 * sizeof(u32));
-	if (ret)
-		return ret;
-
-	uregs[0] = regs->psr;
-	uregs[1] = regs->pc;
-	uregs[2] = regs->npc;
-	uregs[3] = regs->y;
-	uregs[4] = 0;	/* WIM */
-	uregs[5] = 0;	/* TBR */
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  uregs,
-				  32 * sizeof(u32), 38 * sizeof(u32));
+	membuf_write(&to, uregs, 16 * sizeof(u32));
+	membuf_store(&to, regs->psr);
+	membuf_store(&to, regs->pc);
+	membuf_store(&to, regs->npc);
+	membuf_store(&to, regs->y);
+	return membuf_zero(&to, 2 * sizeof(u32));
 }
 
 static int genregs32_set(struct task_struct *target,
@@ -179,46 +165,18 @@ static int genregs32_set(struct task_struct *target,
 
 static int fpregs32_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
-	const unsigned long *fpregs = target->thread.float_regs;
-	int ret = 0;
-
 #if 0
 	if (target == current)
 		save_and_clear_fpu();
 #endif
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  fpregs,
-				  0, 32 * sizeof(u32));
-
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       32 * sizeof(u32),
-					       33 * sizeof(u32));
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.fsr,
-					  33 * sizeof(u32),
-					  34 * sizeof(u32));
-
-	if (!ret) {
-		unsigned long val;
-
-		val = (1 << 8) | (8 << 16);
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &val,
-					  34 * sizeof(u32),
-					  35 * sizeof(u32));
-	}
-
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       35 * sizeof(u32), -1);
-
-	return ret;
+	membuf_write(&to, target->thread.float_regs, 32 * sizeof(u32));
+	membuf_zero(&to, sizeof(u32));
+	membuf_write(&to, &target->thread.fsr, sizeof(u32));
+	membuf_store(&to, (u32)((1 << 8) | (8 << 16)));
+	return membuf_zero(&to, 64 * sizeof(u32));
 }
 
 static int fpregs32_set(struct task_struct *target,
@@ -263,7 +221,7 @@ static const struct user_regset sparc32_regsets[] = {
 		.core_note_type = NT_PRSTATUS,
 		.n = 38,
 		.size = sizeof(u32), .align = sizeof(u32),
-		.get = genregs32_get, .set = genregs32_set
+		.get2 = genregs32_get, .set = genregs32_set
 	},
 	/* Format is:
 	 *	F0 --> F31
@@ -279,35 +237,24 @@ static const struct user_regset sparc32_regsets[] = {
 		.core_note_type = NT_PRFPREG,
 		.n = 99,
 		.size = sizeof(u32), .align = sizeof(u32),
-		.get = fpregs32_get, .set = fpregs32_set
+		.get2 = fpregs32_get, .set = fpregs32_set
 	},
 };
 
 static int getregs_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
 	const struct pt_regs *regs = target->thread.kregs;
-	u32 v[4];
-	int ret;
 
 	if (target == current)
 		flush_user_windows();
 
-	v[0] = regs->psr;
-	v[1] = regs->pc;
-	v[2] = regs->npc;
-	v[3] = regs->y;
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  v,
-				  0 * sizeof(u32), 4 * sizeof(u32));
-	if (ret)
-		return ret;
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  regs->u_regs + 1,
-				  4 * sizeof(u32), 19 * sizeof(u32));
+	membuf_store(&to, regs->psr);
+	membuf_store(&to, regs->pc);
+	membuf_store(&to, regs->npc);
+	membuf_store(&to, regs->y);
+	return membuf_write(&to, regs->u_regs + 1, 15 * sizeof(u32));
 }
 
 static int setregs_set(struct task_struct *target,
@@ -337,29 +284,15 @@ static int setregs_set(struct task_struct *target,
 
 static int getfpregs_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
-	const unsigned long *fpregs = target->thread.float_regs;
-	int ret = 0;
-
 #if 0
 	if (target == current)
 		save_and_clear_fpu();
 #endif
-
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  fpregs,
-				  0, 32 * sizeof(u32));
-	if (ret)
-		return ret;
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.fsr,
-				  32 * sizeof(u32), 33 * sizeof(u32));
-	if (ret)
-		return ret;
-	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-				  33 * sizeof(u32), 68 * sizeof(u32));
+	membuf_write(&to, &target->thread.float_regs, 32 * sizeof(u32));
+	membuf_write(&to, &target->thread.fsr, sizeof(u32));
+	return membuf_zero(&to, 35 * sizeof(u32));
 }
 
 static int setfpregs_set(struct task_struct *target,
@@ -388,11 +321,11 @@ static int setfpregs_set(struct task_struct *target,
 static const struct user_regset ptrace32_regsets[] = {
 	[REGSET_GENERAL] = {
 		.n = 19, .size = sizeof(u32),
-		.get = getregs_get, .set = setregs_set,
+		.get2 = getregs_get, .set = setregs_set,
 	},
 	[REGSET_FP] = {
 		.n = 68, .size = sizeof(u32),
-		.get = getfpregs_get, .set = setfpregs_set,
+		.get2 = getfpregs_get, .set = setfpregs_set,
 	},
 };
 
diff --git a/arch/sparc/kernel/ptrace_64.c b/arch/sparc/kernel/ptrace_64.c
index 3c9eee12102a..8ac5a3e5b46c 100644
--- a/arch/sparc/kernel/ptrace_64.c
+++ b/arch/sparc/kernel/ptrace_64.c
@@ -246,52 +246,23 @@ enum sparc_regset {
 
 static int genregs64_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	int ret;
+	struct reg_window window;
 
 	if (target == current)
 		flushw_user();
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  regs->u_regs,
-				  0, 16 * sizeof(u64));
-	if (!ret && count) {
-		struct reg_window window;
-
-		if (regwindow64_get(target, regs, &window))
-			return -EFAULT;
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &window,
-					  16 * sizeof(u64),
-					  32 * sizeof(u64));
-	}
-
-	if (!ret) {
-		/* TSTATE, TPC, TNPC */
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &regs->tstate,
-					  32 * sizeof(u64),
-					  35 * sizeof(u64));
-	}
-
-	if (!ret) {
-		unsigned long y = regs->y;
-
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &y,
-					  35 * sizeof(u64),
-					  36 * sizeof(u64));
-	}
-
-	if (!ret) {
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       36 * sizeof(u64), -1);
-
-	}
-	return ret;
+	membuf_write(&to, regs->u_regs, 16 * sizeof(u64));
+	if (!to.left)
+		return 0;
+	if (regwindow64_get(target, regs, &window))
+		return -EFAULT;
+	membuf_write(&to, &window, 16 * sizeof(u64));
+	/* TSTATE, TPC, TNPC */
+	membuf_write(&to, &regs->tstate, 3 * sizeof(u64));
+	return membuf_store(&to, (u64)regs->y);
 }
 
 static int genregs64_set(struct task_struct *target,
@@ -370,69 +341,32 @@ static int genregs64_set(struct task_struct *target,
 
 static int fpregs64_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
-	const unsigned long *fpregs = task_thread_info(target)->fpregs;
-	unsigned long fprs, fsr, gsr;
-	int ret;
+	struct thread_info *t = task_thread_info(target);
+	unsigned long fprs;
 
 	if (target == current)
 		save_and_clear_fpu();
 
-	fprs = task_thread_info(target)->fpsaved[0];
+	fprs = t->fpsaved[0];
 
 	if (fprs & FPRS_DL)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  fpregs,
-					  0, 16 * sizeof(u64));
+		membuf_write(&to, t->fpregs, 16 * sizeof(u64));
 	else
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       0,
-					       16 * sizeof(u64));
-
-	if (!ret) {
-		if (fprs & FPRS_DU)
-			ret = user_regset_copyout(&pos, &count,
-						  &kbuf, &ubuf,
-						  fpregs + 16,
-						  16 * sizeof(u64),
-						  32 * sizeof(u64));
-		else
-			ret = user_regset_copyout_zero(&pos, &count,
-						       &kbuf, &ubuf,
-						       16 * sizeof(u64),
-						       32 * sizeof(u64));
-	}
+		membuf_zero(&to, 16 * sizeof(u64));
 
+	if (fprs & FPRS_DU)
+		membuf_write(&to, t->fpregs + 16, 16 * sizeof(u64));
+	else
+		membuf_zero(&to, 16 * sizeof(u64));
 	if (fprs & FPRS_FEF) {
-		fsr = task_thread_info(target)->xfsr[0];
-		gsr = task_thread_info(target)->gsr[0];
+		membuf_store(&to, t->xfsr[0]);
+		membuf_store(&to, t->gsr[0]);
 	} else {
-		fsr = gsr = 0;
+		membuf_zero(&to, 2 * sizeof(u64));
 	}
-
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &fsr,
-					  32 * sizeof(u64),
-					  33 * sizeof(u64));
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &gsr,
-					  33 * sizeof(u64),
-					  34 * sizeof(u64));
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &fprs,
-					  34 * sizeof(u64),
-					  35 * sizeof(u64));
-
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       35 * sizeof(u64), -1);
-
-	return ret;
+	return membuf_store(&to, fprs);
 }
 
 static int fpregs64_set(struct task_struct *target,
@@ -490,7 +424,7 @@ static const struct user_regset sparc64_regsets[] = {
 		.core_note_type = NT_PRSTATUS,
 		.n = 36,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = genregs64_get, .set = genregs64_set
+		.get2 = genregs64_get, .set = genregs64_set
 	},
 	/* Format is:
 	 *	F0 --> F63
@@ -502,43 +436,23 @@ static const struct user_regset sparc64_regsets[] = {
 		.core_note_type = NT_PRFPREG,
 		.n = 35,
 		.size = sizeof(u64), .align = sizeof(u64),
-		.get = fpregs64_get, .set = fpregs64_set
+		.get2 = fpregs64_get, .set = fpregs64_set
 	},
 };
 
 static int getregs64_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	int ret;
 
 	if (target == current)
 		flushw_user();
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  regs->u_regs + 1,
-				  0, 15 * sizeof(u64));
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-				  15 * sizeof(u64), 16 * sizeof(u64));
-	if (!ret) {
-		/* TSTATE, TPC, TNPC */
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &regs->tstate,
-					  16 * sizeof(u64),
-					  19 * sizeof(u64));
-	}
-	if (!ret) {
-		unsigned long y = regs->y;
-
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &y,
-					  19 * sizeof(u64),
-					  20 * sizeof(u64));
-	}
-	return ret;
+	membuf_write(&to, regs->u_regs + 1, 15 * sizeof(u64));
+	return membuf_store(&to, (u64)0);
+	membuf_write(&to, &regs->tstate, 3 * sizeof(u64));
+	return membuf_store(&to, (u64)regs->y);
 }
 
 static int setregs64_set(struct task_struct *target,
@@ -604,7 +518,7 @@ static const struct user_regset ptrace64_regsets[] = {
 	 */
 	[REGSET_GENERAL] = {
 		.n = 20, .size = sizeof(u64),
-		.get = getregs64_get, .set = setregs64_set,
+		.get2 = getregs64_get, .set = setregs64_set,
 	},
 };
 
@@ -620,81 +534,28 @@ static const struct user_regset_view user_sparc64_view = {
 #ifdef CONFIG_COMPAT
 static int genregs32_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	compat_ulong_t *k = kbuf;
-	compat_ulong_t __user *u = ubuf;
 	u32 uregs[16];
-	u32 reg;
+	int i;
 
 	if (target == current)
 		flushw_user();
 
-	pos /= sizeof(reg);
-	count /= sizeof(reg);
-
-	if (kbuf) {
-		for (; count > 0 && pos < 16; count--)
-			*k++ = regs->u_regs[pos++];
-
-		if (count) {
-			if (get_from_target(target, regs->u_regs[UREG_I6],
-					uregs, sizeof(uregs)))
-				return -EFAULT;
-			for (; count > 0 && pos < 32; count--)
-				*k++ = uregs[pos++ - 16];
-
-		}
-	} else {
-		for (; count > 0 && pos < 16; count--)
-			if (put_user((compat_ulong_t) regs->u_regs[pos++], u++))
-				return -EFAULT;
-		if (count) {
-			if (get_from_target(target, regs->u_regs[UREG_I6],
-					uregs, sizeof(uregs)))
-				return -EFAULT;
-			for (; count > 0 && pos < 32; count--)
-				if (put_user(uregs[pos++ - 16], u++))
-					return -EFAULT;
-		}
-	}
-	while (count > 0) {
-		switch (pos) {
-		case 32: /* PSR */
-			reg = tstate_to_psr(regs->tstate);
-			break;
-		case 33: /* PC */
-			reg = regs->tpc;
-			break;
-		case 34: /* NPC */
-			reg = regs->tnpc;
-			break;
-		case 35: /* Y */
-			reg = regs->y;
-			break;
-		case 36: /* WIM */
-		case 37: /* TBR */
-			reg = 0;
-			break;
-		default:
-			goto finish;
-		}
-
-		if (kbuf)
-			*k++ = reg;
-		else if (put_user(reg, u++))
-			return -EFAULT;
-		pos++;
-		count--;
-	}
-finish:
-	pos *= sizeof(reg);
-	count *= sizeof(reg);
-
-	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					38 * sizeof(reg), -1);
+	for (i = 0; i < 16; i++)
+		membuf_store(&to, (u32)regs->u_regs[i]);
+	if (!to.left)
+		return 0;
+	if (get_from_target(target, regs->u_regs[UREG_I6],
+			    uregs, sizeof(uregs)))
+		return -EFAULT;
+	membuf_write(&to, uregs, 16 * sizeof(u32));
+	membuf_store(&to, (u32)tstate_to_psr(regs->tstate));
+	membuf_store(&to, (u32)(regs->tpc));
+	membuf_store(&to, (u32)(regs->tnpc));
+	membuf_store(&to, (u32)(regs->y));
+	return membuf_zero(&to, 2 * sizeof(u32));
 }
 
 static int genregs32_set(struct task_struct *target,
@@ -816,56 +677,24 @@ static int genregs32_set(struct task_struct *target,
 
 static int fpregs32_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
-	const unsigned long *fpregs = task_thread_info(target)->fpregs;
-	compat_ulong_t enabled;
-	unsigned long fprs;
-	compat_ulong_t fsr;
-	int ret = 0;
+	struct thread_info *t = task_thread_info(target);
+	bool enabled;
 
 	if (target == current)
 		save_and_clear_fpu();
 
-	fprs = task_thread_info(target)->fpsaved[0];
-	if (fprs & FPRS_FEF) {
-		fsr = task_thread_info(target)->xfsr[0];
-		enabled = 1;
-	} else {
-		fsr = 0;
-		enabled = 0;
-	}
-
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  fpregs,
-				  0, 32 * sizeof(u32));
-
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       32 * sizeof(u32),
-					       33 * sizeof(u32));
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &fsr,
-					  33 * sizeof(u32),
-					  34 * sizeof(u32));
+	enabled = t->fpsaved[0] & FPRS_FEF;
 
-	if (!ret) {
-		compat_ulong_t val;
-
-		val = (enabled << 8) | (8 << 16);
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &val,
-					  34 * sizeof(u32),
-					  35 * sizeof(u32));
-	}
-
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       35 * sizeof(u32), -1);
-
-	return ret;
+	membuf_write(&to, t->fpregs, 32 * sizeof(u32));
+	membuf_zero(&to, sizeof(u32));
+	if (enabled)
+		membuf_store(&to, (u32)t->xfsr[0]);
+	else
+		membuf_zero(&to, sizeof(u32));
+	membuf_store(&to, (u32)((enabled << 8) | (8 << 16)));
+	return membuf_zero(&to, 64 * sizeof(u32));
 }
 
 static int fpregs32_set(struct task_struct *target,
@@ -926,7 +755,7 @@ static const struct user_regset sparc32_regsets[] = {
 		.core_note_type = NT_PRSTATUS,
 		.n = 38,
 		.size = sizeof(u32), .align = sizeof(u32),
-		.get = genregs32_get, .set = genregs32_set
+		.get2 = genregs32_get, .set = genregs32_set
 	},
 	/* Format is:
 	 *	F0 --> F31
@@ -942,31 +771,27 @@ static const struct user_regset sparc32_regsets[] = {
 		.core_note_type = NT_PRFPREG,
 		.n = 99,
 		.size = sizeof(u32), .align = sizeof(u32),
-		.get = fpregs32_get, .set = fpregs32_set
+		.get2 = fpregs32_get, .set = fpregs32_set
 	},
 };
 
 static int getregs_get(struct task_struct *target,
-		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+			 const struct user_regset *regset,
+			 struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	u32 uregs[19];
 	int i;
 
 	if (target == current)
 		flushw_user();
 
-	uregs[0] = tstate_to_psr(regs->tstate);
-	uregs[1] = regs->tpc;
-	uregs[2] = regs->tnpc;
-	uregs[3] = regs->y;
+	membuf_store(&to, (u32)tstate_to_psr(regs->tstate));
+	membuf_store(&to, (u32)(regs->tpc));
+	membuf_store(&to, (u32)(regs->tnpc));
+	membuf_store(&to, (u32)(regs->y));
 	for (i = 1; i < 16; i++)
-		uregs[3 + i] = regs->u_regs[i];
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   uregs,
-				   0, 19 * sizeof(u32));
+		membuf_store(&to, (u32)regs->u_regs[i]);
+	return to.left;
 }
 
 static int setregs_set(struct task_struct *target,
@@ -1005,36 +830,19 @@ static int setregs_set(struct task_struct *target,
 
 static int getfpregs_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
-	const unsigned long *fpregs = task_thread_info(target)->fpregs;
-	unsigned long fprs;
-	compat_ulong_t fsr;
-	int ret = 0;
+	struct thread_info *t = task_thread_info(target);
 
 	if (target == current)
 		save_and_clear_fpu();
 
-	fprs = task_thread_info(target)->fpsaved[0];
-	if (fprs & FPRS_FEF) {
-		fsr = task_thread_info(target)->xfsr[0];
-	} else {
-		fsr = 0;
-	}
-
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  fpregs,
-				  0, 32 * sizeof(u32));
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &fsr,
-					  32 * sizeof(u32),
-					  33 * sizeof(u32));
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					  33 * sizeof(u32), 68 * sizeof(u32));
-	return ret;
+	membuf_write(&to, t->fpregs, 32 * sizeof(u32));
+	if (t->fpsaved[0] & FPRS_FEF)
+		membuf_store(&to, (u32)t->xfsr[0]);
+	else
+		membuf_zero(&to, sizeof(u32));
+	return membuf_zero(&to, 35 * sizeof(u32));
 }
 
 static int setfpregs_set(struct task_struct *target,
@@ -1078,11 +886,11 @@ static int setfpregs_set(struct task_struct *target,
 static const struct user_regset ptrace32_regsets[] = {
 	[REGSET_GENERAL] = {
 		.n = 19, .size = sizeof(u32),
-		.get = getregs_get, .set = setregs_set,
+		.get2 = getregs_get, .set = setregs_set,
 	},
 	[REGSET_FP] = {
 		.n = 68, .size = sizeof(u32),
-		.get = getfpregs_get, .set = setfpregs_set,
+		.get2 = getfpregs_get, .set = setfpregs_set,
 	},
 };
 
-- 
2.11.0

[PATCH 23/41] mips: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/mips/kernel/ptrace.c | 204 +++++++++++++---------------------------------
 1 file changed, 58 insertions(+), 146 deletions(-)

diff --git a/arch/mips/kernel/ptrace.c b/arch/mips/kernel/ptrace.c
index 2a61641c680b..9622c059f54d 100644
--- a/arch/mips/kernel/ptrace.c
+++ b/arch/mips/kernel/ptrace.c
@@ -210,15 +210,13 @@ int ptrace_set_watch_regs(struct task_struct *child,
 
 static int gpr32_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
 	struct pt_regs *regs = task_pt_regs(target);
 	u32 uregs[ELF_NGREG] = {};
 
 	mips_dump_regs32(uregs, regs);
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, uregs, 0,
-				   sizeof(uregs));
+	return membuf_write(&to, uregs, sizeof(uregs));
 }
 
 static int gpr32_set(struct task_struct *target,
@@ -277,15 +275,13 @@ static int gpr32_set(struct task_struct *target,
 
 static int gpr64_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
 	struct pt_regs *regs = task_pt_regs(target);
 	u64 uregs[ELF_NGREG] = {};
 
 	mips_dump_regs64(uregs, regs);
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, uregs, 0,
-				   sizeof(uregs));
+	return membuf_write(&to, uregs, sizeof(uregs));
 }
 
 static int gpr64_set(struct task_struct *target,
@@ -408,13 +404,11 @@ int ptrace_setfpregs(struct task_struct *child, __u32 __user *data)
  * !CONFIG_CPU_HAS_MSA variant.  FP context's general register slots
  * correspond 1:1 to buffer slots.  Only general registers are copied.
  */
-static int fpr_get_fpa(struct task_struct *target,
-		       unsigned int *pos, unsigned int *count,
-		       void **kbuf, void __user **ubuf)
+static void fpr_get_fpa(struct task_struct *target,
+		       struct membuf *to)
 {
-	return user_regset_copyout(pos, count, kbuf, ubuf,
-				   &target->thread.fpu,
-				   0, NUM_FPU_REGS * sizeof(elf_fpreg_t));
+	membuf_write(to, &target->thread.fpu,
+			NUM_FPU_REGS * sizeof(elf_fpreg_t));
 }
 
 /*
@@ -423,25 +417,13 @@ static int fpr_get_fpa(struct task_struct *target,
  * general register slots are copied to buffer slots.  Only general
  * registers are copied.
  */
-static int fpr_get_msa(struct task_struct *target,
-		       unsigned int *pos, unsigned int *count,
-		       void **kbuf, void __user **ubuf)
+static void fpr_get_msa(struct task_struct *target, struct membuf *to)
 {
 	unsigned int i;
-	u64 fpr_val;
-	int err;
-
-	BUILD_BUG_ON(sizeof(fpr_val) != sizeof(elf_fpreg_t));
-	for (i = 0; i < NUM_FPU_REGS; i++) {
-		fpr_val = get_fpr64(&target->thread.fpu.fpr[i], 0);
-		err = user_regset_copyout(pos, count, kbuf, ubuf,
-					  &fpr_val, i * sizeof(elf_fpreg_t),
-					  (i + 1) * sizeof(elf_fpreg_t));
-		if (err)
-			return err;
-	}
 
-	return 0;
+	BUILD_BUG_ON(sizeof(u64) != sizeof(elf_fpreg_t));
+	for (i = 0; i < NUM_FPU_REGS; i++)
+		membuf_store(to, get_fpr64(&target->thread.fpu.fpr[i], 0));
 }
 
 /*
@@ -451,31 +433,16 @@ static int fpr_get_msa(struct task_struct *target,
  */
 static int fpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	const int fcr31_pos = NUM_FPU_REGS * sizeof(elf_fpreg_t);
-	const int fir_pos = fcr31_pos + sizeof(u32);
-	int err;
-
 	if (sizeof(target->thread.fpu.fpr[0]) == sizeof(elf_fpreg_t))
-		err = fpr_get_fpa(target, &pos, &count, &kbuf, &ubuf);
+		fpr_get_fpa(target, &to);
 	else
-		err = fpr_get_msa(target, &pos, &count, &kbuf, &ubuf);
-	if (err)
-		return err;
-
-	err = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.fpu.fcr31,
-				  fcr31_pos, fcr31_pos + sizeof(u32));
-	if (err)
-		return err;
+		fpr_get_msa(target, &to);
 
-	err = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &boot_cpu_data.fpu_id,
-				  fir_pos, fir_pos + sizeof(u32));
-
-	return err;
+	membuf_write(&to, &target->thread.fpu.fcr31, sizeof(u32));
+	membuf_write(&to, &boot_cpu_data.fpu_id, sizeof(u32));
+	return 0;
 }
 
 /*
@@ -576,14 +543,9 @@ static int fpr_set(struct task_struct *target,
 /* Copy the FP mode setting to the supplied NT_MIPS_FP_MODE buffer.  */
 static int fp_mode_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+		       struct membuf to)
 {
-	int fp_mode;
-
-	fp_mode = mips_get_process_fp_mode(target);
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &fp_mode, 0,
-				   sizeof(fp_mode));
+	return membuf_store(&to, (int)mips_get_process_fp_mode(target));
 }
 
 /*
@@ -630,13 +592,12 @@ struct msa_control_regs {
 	unsigned int msacsr;
 };
 
-static int copy_pad_fprs(struct task_struct *target,
+static void copy_pad_fprs(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int *ppos, unsigned int *pcount,
-			 void **pkbuf, void __user **pubuf,
+			 struct membuf *to,
 			 unsigned int live_sz)
 {
-	int i, j, start, start_pad, err;
+	int i, j;
 	unsigned long long fill = ~0ull;
 	unsigned int cp_sz, pad_sz;
 
@@ -644,28 +605,16 @@ static int copy_pad_fprs(struct task_struct *target,
 	pad_sz = regset->size - cp_sz;
 	WARN_ON(pad_sz % sizeof(fill));
 
-	i = start = err = 0;
-	for (; i < NUM_FPU_REGS; i++, start += regset->size) {
-		err |= user_regset_copyout(ppos, pcount, pkbuf, pubuf,
-					   &target->thread.fpu.fpr[i],
-					   start, start + cp_sz);
-
-		start_pad = start + cp_sz;
-		for (j = 0; j < (pad_sz / sizeof(fill)); j++) {
-			err |= user_regset_copyout(ppos, pcount, pkbuf, pubuf,
-						   &fill, start_pad,
-						   start_pad + sizeof(fill));
-			start_pad += sizeof(fill);
-		}
+	for (i = 0; i < NUM_FPU_REGS; i++) {
+		membuf_write(to, &target->thread.fpu.fpr[i], cp_sz);
+		for (j = 0; j < (pad_sz / sizeof(fill)); j++)
+			membuf_store(to, fill);
 	}
-
-	return err;
 }
 
 static int msa_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	const unsigned int wr_size = NUM_FPU_REGS * regset->size;
 	const struct msa_control_regs ctrl_regs = {
@@ -674,32 +623,23 @@ static int msa_get(struct task_struct *target,
 		.msair = boot_cpu_data.msa_id,
 		.msacsr = target->thread.fpu.msacsr,
 	};
-	int err;
 
 	if (!tsk_used_math(target)) {
 		/* The task hasn't used FP or MSA, fill with 0xff */
-		err = copy_pad_fprs(target, regset, &pos, &count,
-				    &kbuf, &ubuf, 0);
+		copy_pad_fprs(target, regset, &to, 0);
 	} else if (!test_tsk_thread_flag(target, TIF_MSA_CTX_LIVE)) {
 		/* Copy scalar FP context, fill the rest with 0xff */
-		err = copy_pad_fprs(target, regset, &pos, &count,
-				    &kbuf, &ubuf, 8);
+		copy_pad_fprs(target, regset, &to, 8);
 	} else if (sizeof(target->thread.fpu.fpr[0]) == regset->size) {
 		/* Trivially copy the vector registers */
-		err = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &target->thread.fpu.fpr,
-					  0, wr_size);
+		membuf_write(&to, &target->thread.fpu.fpr, wr_size);
 	} else {
 		/* Copy as much context as possible, fill the rest with 0xff */
-		err = copy_pad_fprs(target, regset, &pos, &count,
-				    &kbuf, &ubuf,
-				    sizeof(target->thread.fpu.fpr[0]));
+		copy_pad_fprs(target, regset, &to,
+				sizeof(target->thread.fpu.fpr[0]));
 	}
 
-	err |= user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &ctrl_regs, wr_size,
-				   wr_size + sizeof(ctrl_regs));
-	return err;
+	return membuf_write(&to, &ctrl_regs, sizeof(ctrl_regs));
 }
 
 static int msa_set(struct task_struct *target,
@@ -752,34 +692,20 @@ static int msa_set(struct task_struct *target,
  */
 static int dsp32_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
-	unsigned int start, num_regs, i;
 	u32 dspregs[NUM_DSP_REGS + 1];
+	unsigned int i;
 
-	BUG_ON(count % sizeof(u32));
+	BUG_ON(to.left % sizeof(u32));
 
 	if (!cpu_has_dsp)
 		return -EIO;
 
-	start = pos / sizeof(u32);
-	num_regs = count / sizeof(u32);
-
-	if (start + num_regs > NUM_DSP_REGS + 1)
-		return -EIO;
-
-	for (i = start; i < num_regs; i++)
-		switch (i) {
-		case 0 ... NUM_DSP_REGS - 1:
-			dspregs[i] = target->thread.dsp.dspr[i];
-			break;
-		case NUM_DSP_REGS:
-			dspregs[i] = target->thread.dsp.dspcontrol;
-			break;
-		}
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, dspregs, 0,
-				   sizeof(dspregs));
+	for (i = 0; i < NUM_DSP_REGS; i++)
+		dspregs[i] = target->thread.dsp.dspr[i];
+	dspregs[NUM_DSP_REGS] = target->thread.dsp.dspcontrol;
+	return membuf_write(&to, dspregs, sizeof(dspregs));
 }
 
 /*
@@ -832,34 +758,20 @@ static int dsp32_set(struct task_struct *target,
  */
 static int dsp64_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
-	unsigned int start, num_regs, i;
 	u64 dspregs[NUM_DSP_REGS + 1];
+	unsigned int i;
 
-	BUG_ON(count % sizeof(u64));
+	BUG_ON(to.left % sizeof(u64));
 
 	if (!cpu_has_dsp)
 		return -EIO;
 
-	start = pos / sizeof(u64);
-	num_regs = count / sizeof(u64);
-
-	if (start + num_regs > NUM_DSP_REGS + 1)
-		return -EIO;
-
-	for (i = start; i < num_regs; i++)
-		switch (i) {
-		case 0 ... NUM_DSP_REGS - 1:
-			dspregs[i] = target->thread.dsp.dspr[i];
-			break;
-		case NUM_DSP_REGS:
-			dspregs[i] = target->thread.dsp.dspcontrol;
-			break;
-		}
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, dspregs, 0,
-				   sizeof(dspregs));
+	for (i = 0; i < NUM_DSP_REGS; i++)
+		dspregs[i] = target->thread.dsp.dspr[i];
+	dspregs[NUM_DSP_REGS] = target->thread.dsp.dspcontrol;
+	return membuf_write(&to, dspregs, sizeof(dspregs));
 }
 
 /*
@@ -1018,7 +930,7 @@ static const struct user_regset mips_regsets[] = {
 		.n		= ELF_NGREG,
 		.size		= sizeof(unsigned int),
 		.align		= sizeof(unsigned int),
-		.get		= gpr32_get,
+		.get2		= gpr32_get,
 		.set		= gpr32_set,
 	},
 	[REGSET_DSP] = {
@@ -1026,7 +938,7 @@ static const struct user_regset mips_regsets[] = {
 		.n		= NUM_DSP_REGS + 1,
 		.size		= sizeof(u32),
 		.align		= sizeof(u32),
-		.get		= dsp32_get,
+		.get2		= dsp32_get,
 		.set		= dsp32_set,
 		.active		= dsp_active,
 	},
@@ -1036,7 +948,7 @@ static const struct user_regset mips_regsets[] = {
 		.n		= ELF_NFPREG,
 		.size		= sizeof(elf_fpreg_t),
 		.align		= sizeof(elf_fpreg_t),
-		.get		= fpr_get,
+		.get2		= fpr_get,
 		.set		= fpr_set,
 	},
 	[REGSET_FP_MODE] = {
@@ -1044,7 +956,7 @@ static const struct user_regset mips_regsets[] = {
 		.n		= 1,
 		.size		= sizeof(int),
 		.align		= sizeof(int),
-		.get		= fp_mode_get,
+		.get2		= fp_mode_get,
 		.set		= fp_mode_set,
 	},
 #endif
@@ -1054,7 +966,7 @@ static const struct user_regset mips_regsets[] = {
 		.n		= NUM_FPU_REGS + 1,
 		.size		= 16,
 		.align		= 16,
-		.get		= msa_get,
+		.get2		= msa_get,
 		.set		= msa_set,
 	},
 #endif
@@ -1078,7 +990,7 @@ static const struct user_regset mips64_regsets[] = {
 		.n		= ELF_NGREG,
 		.size		= sizeof(unsigned long),
 		.align		= sizeof(unsigned long),
-		.get		= gpr64_get,
+		.get2		= gpr64_get,
 		.set		= gpr64_set,
 	},
 	[REGSET_DSP] = {
@@ -1086,7 +998,7 @@ static const struct user_regset mips64_regsets[] = {
 		.n		= NUM_DSP_REGS + 1,
 		.size		= sizeof(u64),
 		.align		= sizeof(u64),
-		.get		= dsp64_get,
+		.get2		= dsp64_get,
 		.set		= dsp64_set,
 		.active		= dsp_active,
 	},
@@ -1096,7 +1008,7 @@ static const struct user_regset mips64_regsets[] = {
 		.n		= 1,
 		.size		= sizeof(int),
 		.align		= sizeof(int),
-		.get		= fp_mode_get,
+		.get2		= fp_mode_get,
 		.set		= fp_mode_set,
 	},
 	[REGSET_FPR] = {
@@ -1104,7 +1016,7 @@ static const struct user_regset mips64_regsets[] = {
 		.n		= ELF_NFPREG,
 		.size		= sizeof(elf_fpreg_t),
 		.align		= sizeof(elf_fpreg_t),
-		.get		= fpr_get,
+		.get2		= fpr_get,
 		.set		= fpr_set,
 	},
 #endif
@@ -1114,7 +1026,7 @@ static const struct user_regset mips64_regsets[] = {
 		.n		= NUM_FPU_REGS + 1,
 		.size		= 16,
 		.align		= 16,
-		.get		= msa_get,
+		.get2		= msa_get,
 		.set		= msa_set,
 	},
 #endif
-- 
2.11.0

[PATCH 24/41] arm64: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/arm64/kernel/ptrace.c | 225 +++++++++++++--------------------------------
 1 file changed, 62 insertions(+), 163 deletions(-)

diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index 9f769e862f68..8745aecffcae 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -474,11 +474,10 @@ static int ptrace_hbp_set_addr(unsigned int note_type,
 
 static int hw_break_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
 	unsigned int note_type = regset->core_note_type;
-	int ret, idx = 0, offset, limit;
+	int ret, idx = 0;
 	u32 info, ctrl;
 	u64 addr;
 
@@ -487,49 +486,21 @@ static int hw_break_get(struct task_struct *target,
 	if (ret)
 		return ret;
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &info, 0,
-				  sizeof(info));
-	if (ret)
-		return ret;
-
-	/* Pad */
-	offset = offsetof(struct user_hwdebug_state, pad);
-	ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf, offset,
-				       offset + PTRACE_HBP_PAD_SZ);
-	if (ret)
-		return ret;
-
+	membuf_write(&to, &info, sizeof(info));
+	membuf_zero(&to, sizeof(u32));
 	/* (address, ctrl) registers */
-	offset = offsetof(struct user_hwdebug_state, dbg_regs);
-	limit = regset->n * regset->size;
-	while (count && offset < limit) {
+	while (to.left) {
 		ret = ptrace_hbp_get_addr(note_type, target, idx, &addr);
 		if (ret)
 			return ret;
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &addr,
-					  offset, offset + PTRACE_HBP_ADDR_SZ);
-		if (ret)
-			return ret;
-		offset += PTRACE_HBP_ADDR_SZ;
-
 		ret = ptrace_hbp_get_ctrl(note_type, target, idx, &ctrl);
 		if (ret)
 			return ret;
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &ctrl,
-					  offset, offset + PTRACE_HBP_CTRL_SZ);
-		if (ret)
-			return ret;
-		offset += PTRACE_HBP_CTRL_SZ;
-
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       offset,
-					       offset + PTRACE_HBP_PAD_SZ);
-		if (ret)
-			return ret;
-		offset += PTRACE_HBP_PAD_SZ;
+		membuf_store(&to, addr);
+		membuf_store(&to, ctrl);
+		membuf_zero(&to, sizeof(u32));
 		idx++;
 	}
-
 	return 0;
 }
 
@@ -589,11 +560,10 @@ static int hw_break_set(struct task_struct *target,
 
 static int gpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	struct user_pt_regs *uregs = &task_pt_regs(target)->user_regs;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, uregs, 0, -1);
+	return membuf_write(&to, uregs, sizeof(*uregs));
 }
 
 static int gpr_set(struct task_struct *target, const struct user_regset *regset,
@@ -626,8 +596,7 @@ static int fpr_active(struct task_struct *target, const struct user_regset *regs
  */
 static int __fpr_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf, unsigned int start_pos)
+		     struct membuf to)
 {
 	struct user_fpsimd_state *uregs;
 
@@ -635,13 +604,11 @@ static int __fpr_get(struct task_struct *target,
 
 	uregs = &target->thread.uw.fpsimd_state;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, uregs,
-				   start_pos, start_pos + sizeof(*uregs));
+	return membuf_write(&to, uregs, sizeof(*uregs));
 }
 
 static int fpr_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	if (!system_supports_fpsimd())
 		return -EINVAL;
@@ -649,7 +616,7 @@ static int fpr_get(struct task_struct *target, const struct user_regset *regset,
 	if (target == current)
 		fpsimd_preserve_current_state();
 
-	return __fpr_get(target, regset, pos, count, kbuf, ubuf, 0);
+	return __fpr_get(target, regset, to);
 }
 
 static int __fpr_set(struct task_struct *target,
@@ -699,15 +666,12 @@ static int fpr_set(struct task_struct *target, const struct user_regset *regset,
 }
 
 static int tls_get(struct task_struct *target, const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	unsigned long *tls = &target->thread.uw.tp_value;
-
 	if (target == current)
 		tls_preserve_current_state();
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, tls, 0, -1);
+	return membuf_store(&to, target->thread.uw.tp_value);
 }
 
 static int tls_set(struct task_struct *target, const struct user_regset *regset,
@@ -727,13 +691,9 @@ static int tls_set(struct task_struct *target, const struct user_regset *regset,
 
 static int system_call_get(struct task_struct *target,
 			   const struct user_regset *regset,
-			   unsigned int pos, unsigned int count,
-			   void *kbuf, void __user *ubuf)
+			   struct membuf to)
 {
-	int syscallno = task_pt_regs(target)->syscallno;
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &syscallno, 0, -1);
+	return membuf_store(&to, task_pt_regs(target)->syscallno);
 }
 
 static int system_call_set(struct task_struct *target,
@@ -794,10 +754,8 @@ static unsigned int sve_get_size(struct task_struct *target,
 
 static int sve_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	int ret;
 	struct user_sve_header header;
 	unsigned int vq;
 	unsigned long start, end;
@@ -809,10 +767,7 @@ static int sve_get(struct task_struct *target,
 	sve_init_header_from_task(&header, target);
 	vq = sve_vq_from_vl(header.vl);
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &header,
-				  0, sizeof(header));
-	if (ret)
-		return ret;
+	membuf_write(&to, &header, sizeof(header));
 
 	if (target == current)
 		fpsimd_preserve_current_state();
@@ -821,26 +776,18 @@ static int sve_get(struct task_struct *target,
 
 	BUILD_BUG_ON(SVE_PT_FPSIMD_OFFSET != sizeof(header));
 	if ((header.flags & SVE_PT_REGS_MASK) == SVE_PT_REGS_FPSIMD)
-		return __fpr_get(target, regset, pos, count, kbuf, ubuf,
-				 SVE_PT_FPSIMD_OFFSET);
+		return __fpr_get(target, regset, to);
 
 	/* Otherwise: full SVE case */
 
 	BUILD_BUG_ON(SVE_PT_SVE_OFFSET != sizeof(header));
 	start = SVE_PT_SVE_OFFSET;
 	end = SVE_PT_SVE_FFR_OFFSET(vq) + SVE_PT_SVE_FFR_SIZE(vq);
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  target->thread.sve_state,
-				  start, end);
-	if (ret)
-		return ret;
+	membuf_write(&to, target->thread.sve_state, end - start);
 
 	start = end;
 	end = SVE_PT_SVE_FPSR_OFFSET(vq);
-	ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-				       start, end);
-	if (ret)
-		return ret;
+	membuf_zero(&to, end - start);
 
 	/*
 	 * Copy fpsr, and fpcr which must follow contiguously in
@@ -848,16 +795,11 @@ static int sve_get(struct task_struct *target,
 	 */
 	start = end;
 	end = SVE_PT_SVE_FPCR_OFFSET(vq) + SVE_PT_SVE_FPCR_SIZE;
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &target->thread.uw.fpsimd_state.fpsr,
-				  start, end);
-	if (ret)
-		return ret;
+	membuf_write(&to, &target->thread.uw.fpsimd_state.fpsr, end - start);
 
 	start = end;
 	end = sve_size_from_header(&header);
-	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					start, end);
+	return membuf_zero(&to, end - start);
 }
 
 static int sve_set(struct task_struct *target,
@@ -961,8 +903,7 @@ static int sve_set(struct task_struct *target,
 #ifdef CONFIG_ARM64_PTR_AUTH
 static int pac_mask_get(struct task_struct *target,
 			const struct user_regset *regset,
-			unsigned int pos, unsigned int count,
-			void *kbuf, void __user *ubuf)
+			struct membuf to)
 {
 	/*
 	 * The PAC bits can differ across data and instruction pointers
@@ -978,7 +919,7 @@ static int pac_mask_get(struct task_struct *target,
 	if (!system_supports_address_auth())
 		return -EINVAL;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &uregs, 0, -1);
+	return membuf_write(&to, &uregs, sizeof(uregs));
 }
 
 #ifdef CONFIG_CHECKPOINT_RESTORE
@@ -1017,8 +958,7 @@ static void pac_address_keys_from_user(struct ptrauth_keys_user *keys,
 
 static int pac_address_keys_get(struct task_struct *target,
 				const struct user_regset *regset,
-				unsigned int pos, unsigned int count,
-				void *kbuf, void __user *ubuf)
+				struct membuf to)
 {
 	struct ptrauth_keys_user *keys = &target->thread.keys_user;
 	struct user_pac_address_keys user_keys;
@@ -1028,8 +968,7 @@ static int pac_address_keys_get(struct task_struct *target,
 
 	pac_address_keys_to_user(&user_keys, keys);
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &user_keys, 0, -1);
+	return membuf_write(&to, &user_keys, sizeof(user_keys));
 }
 
 static int pac_address_keys_set(struct task_struct *target,
@@ -1068,8 +1007,7 @@ static void pac_generic_keys_from_user(struct ptrauth_keys_user *keys,
 
 static int pac_generic_keys_get(struct task_struct *target,
 				const struct user_regset *regset,
-				unsigned int pos, unsigned int count,
-				void *kbuf, void __user *ubuf)
+				struct membuf to)
 {
 	struct ptrauth_keys_user *keys = &target->thread.keys_user;
 	struct user_pac_generic_keys user_keys;
@@ -1079,8 +1017,7 @@ static int pac_generic_keys_get(struct task_struct *target,
 
 	pac_generic_keys_to_user(&user_keys, keys);
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &user_keys, 0, -1);
+	return membuf_write(&to, &user_keys, sizeof(user_keys));
 }
 
 static int pac_generic_keys_set(struct task_struct *target,
@@ -1134,7 +1071,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = sizeof(struct user_pt_regs) / sizeof(u64),
 		.size = sizeof(u64),
 		.align = sizeof(u64),
-		.get = gpr_get,
+		.get2 = gpr_get,
 		.set = gpr_set
 	},
 	[REGSET_FPR] = {
@@ -1147,7 +1084,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.size = sizeof(u32),
 		.align = sizeof(u32),
 		.active = fpr_active,
-		.get = fpr_get,
+		.get2 = fpr_get,
 		.set = fpr_set
 	},
 	[REGSET_TLS] = {
@@ -1155,7 +1092,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = 1,
 		.size = sizeof(void *),
 		.align = sizeof(void *),
-		.get = tls_get,
+		.get2 = tls_get,
 		.set = tls_set,
 	},
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
@@ -1164,7 +1101,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = sizeof(struct user_hwdebug_state) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = hw_break_get,
+		.get2 = hw_break_get,
 		.set = hw_break_set,
 	},
 	[REGSET_HW_WATCH] = {
@@ -1172,7 +1109,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = sizeof(struct user_hwdebug_state) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = hw_break_get,
+		.get2 = hw_break_get,
 		.set = hw_break_set,
 	},
 #endif
@@ -1181,7 +1118,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = 1,
 		.size = sizeof(int),
 		.align = sizeof(int),
-		.get = system_call_get,
+		.get2 = system_call_get,
 		.set = system_call_set,
 	},
 #ifdef CONFIG_ARM64_SVE
@@ -1191,7 +1128,7 @@ static const struct user_regset aarch64_regsets[] = {
 				  SVE_VQ_BYTES),
 		.size = SVE_VQ_BYTES,
 		.align = SVE_VQ_BYTES,
-		.get = sve_get,
+		.get2 = sve_get,
 		.set = sve_set,
 		.get_size = sve_get_size,
 	},
@@ -1202,7 +1139,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = sizeof(struct user_pac_mask) / sizeof(u64),
 		.size = sizeof(u64),
 		.align = sizeof(u64),
-		.get = pac_mask_get,
+		.get2 = pac_mask_get,
 		/* this cannot be set dynamically */
 	},
 #ifdef CONFIG_CHECKPOINT_RESTORE
@@ -1211,7 +1148,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = sizeof(struct user_pac_address_keys) / sizeof(__uint128_t),
 		.size = sizeof(__uint128_t),
 		.align = sizeof(__uint128_t),
-		.get = pac_address_keys_get,
+		.get2 = pac_address_keys_get,
 		.set = pac_address_keys_set,
 	},
 	[REGSET_PACG_KEYS] = {
@@ -1219,7 +1156,7 @@ static const struct user_regset aarch64_regsets[] = {
 		.n = sizeof(struct user_pac_generic_keys) / sizeof(__uint128_t),
 		.size = sizeof(__uint128_t),
 		.align = sizeof(__uint128_t),
-		.get = pac_generic_keys_get,
+		.get2 = pac_generic_keys_get,
 		.set = pac_generic_keys_set,
 	},
 #endif
@@ -1255,39 +1192,13 @@ static inline compat_ulong_t compat_get_user_reg(struct task_struct *task, int i
 
 static int compat_gpr_get(struct task_struct *target,
 			  const struct user_regset *regset,
-			  unsigned int pos, unsigned int count,
-			  void *kbuf, void __user *ubuf)
+			  struct membuf to)
 {
-	int ret = 0;
-	unsigned int i, start, num_regs;
-
-	/* Calculate the number of AArch32 registers contained in count */
-	num_regs = count / regset->size;
-
-	/* Convert pos into an register number */
-	start = pos / regset->size;
-
-	if (start + num_regs > regset->n)
-		return -EIO;
-
-	for (i = 0; i < num_regs; ++i) {
-		compat_ulong_t reg = compat_get_user_reg(target, start + i);
-
-		if (kbuf) {
-			memcpy(kbuf, &reg, sizeof(reg));
-			kbuf += sizeof(reg);
-		} else {
-			ret = copy_to_user(ubuf, &reg, sizeof(reg));
-			if (ret) {
-				ret = -EFAULT;
-				break;
-			}
+	int i = 0;
 
-			ubuf += sizeof(reg);
-		}
-	}
-
-	return ret;
+	while (to.left)
+		membuf_store(&to, compat_get_user_reg(target, i++));
+	return 0;
 }
 
 static int compat_gpr_set(struct task_struct *target,
@@ -1354,12 +1265,10 @@ static int compat_gpr_set(struct task_struct *target,
 
 static int compat_vfp_get(struct task_struct *target,
 			  const struct user_regset *regset,
-			  unsigned int pos, unsigned int count,
-			  void *kbuf, void __user *ubuf)
+			  struct membuf to)
 {
 	struct user_fpsimd_state *uregs;
 	compat_ulong_t fpscr;
-	int ret, vregs_end_pos;
 
 	if (!system_supports_fpsimd())
 		return -EINVAL;
@@ -1373,19 +1282,10 @@ static int compat_vfp_get(struct task_struct *target,
 	 * The VFP registers are packed into the fpsimd_state, so they all sit
 	 * nicely together for us. We just need to create the fpscr separately.
 	 */
-	vregs_end_pos = VFP_STATE_SIZE - sizeof(compat_ulong_t);
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, uregs,
-				  0, vregs_end_pos);
-
-	if (count && !ret) {
-		fpscr = (uregs->fpsr & VFP_FPSCR_STAT_MASK) |
-			(uregs->fpcr & VFP_FPSCR_CTRL_MASK);
-
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &fpscr,
-					  vregs_end_pos, VFP_STATE_SIZE);
-	}
-
-	return ret;
+	membuf_write(&to, uregs, VFP_STATE_SIZE - sizeof(compat_ulong_t));
+	fpscr = (uregs->fpsr & VFP_FPSCR_STAT_MASK) |
+		(uregs->fpcr & VFP_FPSCR_CTRL_MASK);
+	return membuf_store(&to, fpscr);
 }
 
 static int compat_vfp_set(struct task_struct *target,
@@ -1420,11 +1320,10 @@ static int compat_vfp_set(struct task_struct *target,
 }
 
 static int compat_tls_get(struct task_struct *target,
-			  const struct user_regset *regset, unsigned int pos,
-			  unsigned int count, void *kbuf, void __user *ubuf)
+			  const struct user_regset *regset,
+			  struct membuf to)
 {
-	compat_ulong_t tls = (compat_ulong_t)target->thread.uw.tp_value;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &tls, 0, -1);
+	return membuf_store(&to, (compat_ulong_t)target->thread.uw.tp_value);
 }
 
 static int compat_tls_set(struct task_struct *target,
@@ -1449,7 +1348,7 @@ static const struct user_regset aarch32_regsets[] = {
 		.n = COMPAT_ELF_NGREG,
 		.size = sizeof(compat_elf_greg_t),
 		.align = sizeof(compat_elf_greg_t),
-		.get = compat_gpr_get,
+		.get2 = compat_gpr_get,
 		.set = compat_gpr_set
 	},
 	[REGSET_COMPAT_VFP] = {
@@ -1458,7 +1357,7 @@ static const struct user_regset aarch32_regsets[] = {
 		.size = sizeof(compat_ulong_t),
 		.align = sizeof(compat_ulong_t),
 		.active = fpr_active,
-		.get = compat_vfp_get,
+		.get2 = compat_vfp_get,
 		.set = compat_vfp_set
 	},
 };
@@ -1474,7 +1373,7 @@ static const struct user_regset aarch32_ptrace_regsets[] = {
 		.n = COMPAT_ELF_NGREG,
 		.size = sizeof(compat_elf_greg_t),
 		.align = sizeof(compat_elf_greg_t),
-		.get = compat_gpr_get,
+		.get2 = compat_gpr_get,
 		.set = compat_gpr_set
 	},
 	[REGSET_FPR] = {
@@ -1482,7 +1381,7 @@ static const struct user_regset aarch32_ptrace_regsets[] = {
 		.n = VFP_STATE_SIZE / sizeof(compat_ulong_t),
 		.size = sizeof(compat_ulong_t),
 		.align = sizeof(compat_ulong_t),
-		.get = compat_vfp_get,
+		.get2 = compat_vfp_get,
 		.set = compat_vfp_set
 	},
 	[REGSET_TLS] = {
@@ -1490,7 +1389,7 @@ static const struct user_regset aarch32_ptrace_regsets[] = {
 		.n = 1,
 		.size = sizeof(compat_ulong_t),
 		.align = sizeof(compat_ulong_t),
-		.get = compat_tls_get,
+		.get2 = compat_tls_get,
 		.set = compat_tls_set,
 	},
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
@@ -1499,7 +1398,7 @@ static const struct user_regset aarch32_ptrace_regsets[] = {
 		.n = sizeof(struct user_hwdebug_state) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = hw_break_get,
+		.get2 = hw_break_get,
 		.set = hw_break_set,
 	},
 	[REGSET_HW_WATCH] = {
@@ -1507,7 +1406,7 @@ static const struct user_regset aarch32_ptrace_regsets[] = {
 		.n = sizeof(struct user_hwdebug_state) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = hw_break_get,
+		.get2 = hw_break_get,
 		.set = hw_break_set,
 	},
 #endif
@@ -1516,7 +1415,7 @@ static const struct user_regset aarch32_ptrace_regsets[] = {
 		.n = 1,
 		.size = sizeof(int),
 		.align = sizeof(int),
-		.get = system_call_get,
+		.get2 = system_call_get,
 		.set = system_call_set,
 	},
 };
-- 
2.11.0

[PATCH 25/41] sh: convert to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

NB: there's a direct call of fpregs_get() left in dump_fpu().
To be taken out once we convert ELF_FDPIC to use of regset.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/sh/kernel/process_32.c |  5 ++---
 arch/sh/kernel/ptrace_32.c  | 48 ++++++++++-----------------------------------
 2 files changed, 12 insertions(+), 41 deletions(-)

diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c
index 456cc8d171f7..6ab397bc47ed 100644
--- a/arch/sh/kernel/process_32.c
+++ b/arch/sh/kernel/process_32.c
@@ -103,9 +103,8 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
 
 	fpvalid = !!tsk_used_math(tsk);
 	if (fpvalid)
-		fpvalid = !fpregs_get(tsk, NULL, 0,
-				      sizeof(struct user_fpu_struct),
-				      fpu, NULL);
+		fpvalid = !fpregs_get(tsk, NULL,
+				      (struct membuf){fpu, sizeof(*fpu)});
 #endif
 
 	return fpvalid;
diff --git a/arch/sh/kernel/ptrace_32.c b/arch/sh/kernel/ptrace_32.c
index 64bfb714943e..7aa5e61ccb7b 100644
--- a/arch/sh/kernel/ptrace_32.c
+++ b/arch/sh/kernel/ptrace_32.c
@@ -134,26 +134,11 @@ void ptrace_disable(struct task_struct *child)
 
 static int genregs_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+		       struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	int ret;
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  regs->regs,
-				  0, 16 * sizeof(unsigned long));
-	if (!ret)
-		/* PC, PR, SR, GBR, MACH, MACL, TRA */
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &regs->pc,
-					  offsetof(struct pt_regs, pc),
-					  sizeof(struct pt_regs));
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       sizeof(struct pt_regs), -1);
-
-	return ret;
+	return membuf_write(&to, regs, sizeof(struct pt_regs));
 }
 
 static int genregs_set(struct task_struct *target,
@@ -182,8 +167,7 @@ static int genregs_set(struct task_struct *target,
 #ifdef CONFIG_SH_FPU
 int fpregs_get(struct task_struct *target,
 	       const struct user_regset *regset,
-	       unsigned int pos, unsigned int count,
-	       void *kbuf, void __user *ubuf)
+	       struct membuf to)
 {
 	int ret;
 
@@ -191,12 +175,8 @@ int fpregs_get(struct task_struct *target,
 	if (ret)
 		return ret;
 
-	if ((boot_cpu_data.flags & CPU_HAS_FPU))
-		return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					   &target->thread.xstate->hardfpu, 0, -1);
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &target->thread.xstate->softfpu, 0, -1);
+	return membuf_write(&to, target->thread.xstate,
+			    sizeof(struct user_fpu_struct));
 }
 
 static int fpregs_set(struct task_struct *target,
@@ -230,20 +210,12 @@ static int fpregs_active(struct task_struct *target,
 #ifdef CONFIG_SH_DSP
 static int dspregs_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+		       struct membuf to)
 {
 	const struct pt_dspregs *regs =
 		(struct pt_dspregs *)&target->thread.dsp_status.dsp_regs;
-	int ret;
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, regs,
-				  0, sizeof(struct pt_dspregs));
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       sizeof(struct pt_dspregs), -1);
-
-	return ret;
+	return membuf_write(&to, regs, sizeof(struct pt_dspregs));
 }
 
 static int dspregs_set(struct task_struct *target,
@@ -324,7 +296,7 @@ static const struct user_regset sh_regsets[] = {
 		.n		= ELF_NGREG,
 		.size		= sizeof(long),
 		.align		= sizeof(long),
-		.get		= genregs_get,
+		.get2		= genregs_get,
 		.set		= genregs_set,
 	},
 
@@ -334,7 +306,7 @@ static const struct user_regset sh_regsets[] = {
 		.n		= sizeof(struct user_fpu_struct) / sizeof(long),
 		.size		= sizeof(long),
 		.align		= sizeof(long),
-		.get		= fpregs_get,
+		.get2		= fpregs_get,
 		.set		= fpregs_set,
 		.active		= fpregs_active,
 	},
@@ -345,7 +317,7 @@ static const struct user_regset sh_regsets[] = {
 		.n		= sizeof(struct pt_dspregs) / sizeof(long),
 		.size		= sizeof(long),
 		.align		= sizeof(long),
-		.get		= dspregs_get,
+		.get2		= dspregs_get,
 		.set		= dspregs_set,
 		.active		= dspregs_active,
 	},
-- 
2.11.0

[PATCH 26/41] arm: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/arm/kernel/ptrace.c | 52 +++++++++++-------------------------------------
 1 file changed, 12 insertions(+), 40 deletions(-)

diff --git a/arch/arm/kernel/ptrace.c b/arch/arm/kernel/ptrace.c
index d0f7c8896c96..895388288093 100644
--- a/arch/arm/kernel/ptrace.c
+++ b/arch/arm/kernel/ptrace.c
@@ -569,14 +569,9 @@ static int ptrace_sethbpregs(struct task_struct *tsk, long num,
 
 static int gpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	struct pt_regs *regs = task_pt_regs(target);
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   regs,
-				   0, sizeof(*regs));
+	return membuf_write(&to, task_pt_regs(target), sizeof(struct pt_regs));
 }
 
 static int gpr_set(struct task_struct *target,
@@ -602,12 +597,10 @@ static int gpr_set(struct task_struct *target,
 
 static int fpa_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &task_thread_info(target)->fpstate,
-				   0, sizeof(struct user_fp));
+	return membuf_write(&to, &task_thread_info(target)->fpstate,
+				 sizeof(struct user_fp));
 }
 
 static int fpa_set(struct task_struct *target,
@@ -642,41 +635,20 @@ static int fpa_set(struct task_struct *target,
  *	vfp_set() ignores this chunk
  *
  * 1 word for the FPSCR
- *
- * The bounds-checking logic built into user_regset_copyout and friends
- * means that we can make a simple sequence of calls to map the relevant data
- * to/from the specified slice of the user regset structure.
  */
 static int vfp_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	int ret;
 	struct thread_info *thread = task_thread_info(target);
 	struct vfp_hard_struct const *vfp = &thread->vfpstate.hard;
-	const size_t user_fpregs_offset = offsetof(struct user_vfp, fpregs);
 	const size_t user_fpscr_offset = offsetof(struct user_vfp, fpscr);
 
 	vfp_sync_hwstate(thread);
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &vfp->fpregs,
-				  user_fpregs_offset,
-				  user_fpregs_offset + sizeof(vfp->fpregs));
-	if (ret)
-		return ret;
-
-	ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-				       user_fpregs_offset + sizeof(vfp->fpregs),
-				       user_fpscr_offset);
-	if (ret)
-		return ret;
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &vfp->fpscr,
-				   user_fpscr_offset,
-				   user_fpscr_offset + sizeof(vfp->fpscr));
+	membuf_write(&to, vfp->fpregs, sizeof(vfp->fpregs));
+	membuf_zero(&to, user_fpscr_offset - sizeof(vfp->fpregs));
+	return membuf_store(&to, vfp->fpscr);
 }
 
 /*
@@ -739,7 +711,7 @@ static const struct user_regset arm_regsets[] = {
 		.n = ELF_NGREG,
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = gpr_get,
+		.get2 = gpr_get,
 		.set = gpr_set
 	},
 	[REGSET_FPR] = {
@@ -751,7 +723,7 @@ static const struct user_regset arm_regsets[] = {
 		.n = sizeof(struct user_fp) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = fpa_get,
+		.get2 = fpa_get,
 		.set = fpa_set
 	},
 #ifdef CONFIG_VFP
@@ -764,7 +736,7 @@ static const struct user_regset arm_regsets[] = {
 		.n = ARM_VFPREGS_SIZE / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = vfp_get,
+		.get2 = vfp_get,
 		.set = vfp_set
 	},
 #endif /* CONFIG_VFP */
-- 
2.11.0

[PATCH 27/41] arc: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

NB: it used to do short store; fix is needed earlier in the series.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/arc/kernel/ptrace.c | 148 +++++++++++++++++++----------------------------
 1 file changed, 58 insertions(+), 90 deletions(-)

diff --git a/arch/arc/kernel/ptrace.c b/arch/arc/kernel/ptrace.c
index f49a054a1016..350aefbd7fdb 100644
--- a/arch/arc/kernel/ptrace.c
+++ b/arch/arc/kernel/ptrace.c
@@ -18,88 +18,61 @@ static struct callee_regs *task_callee_regs(struct task_struct *tsk)
 
 static int genregs_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+		       struct membuf to)
 {
 	const struct pt_regs *ptregs = task_pt_regs(target);
 	const struct callee_regs *cregs = task_callee_regs(target);
-	int ret = 0;
 	unsigned int stop_pc_val;
 
-#define REG_O_CHUNK(START, END, PTR)	\
-	if (!ret)	\
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, PTR, \
-			offsetof(struct user_regs_struct, START), \
-			offsetof(struct user_regs_struct, END));
-
-#define REG_O_ONE(LOC, PTR)	\
-	if (!ret)		\
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, PTR, \
-			offsetof(struct user_regs_struct, LOC), \
-			offsetof(struct user_regs_struct, LOC) + 4);
-
-#define REG_O_ZERO(LOC)		\
-	if (!ret)		\
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf, \
-			offsetof(struct user_regs_struct, LOC), \
-			offsetof(struct user_regs_struct, LOC) + 4);
-
-	REG_O_ZERO(pad);
-	REG_O_ONE(scratch.bta, &ptregs->bta);
-	REG_O_ONE(scratch.lp_start, &ptregs->lp_start);
-	REG_O_ONE(scratch.lp_end, &ptregs->lp_end);
-	REG_O_ONE(scratch.lp_count, &ptregs->lp_count);
-	REG_O_ONE(scratch.status32, &ptregs->status32);
-	REG_O_ONE(scratch.ret, &ptregs->ret);
-	REG_O_ONE(scratch.blink, &ptregs->blink);
-	REG_O_ONE(scratch.fp, &ptregs->fp);
-	REG_O_ONE(scratch.gp, &ptregs->r26);
-	REG_O_ONE(scratch.r12, &ptregs->r12);
-	REG_O_ONE(scratch.r11, &ptregs->r11);
-	REG_O_ONE(scratch.r10, &ptregs->r10);
-	REG_O_ONE(scratch.r9, &ptregs->r9);
-	REG_O_ONE(scratch.r8, &ptregs->r8);
-	REG_O_ONE(scratch.r7, &ptregs->r7);
-	REG_O_ONE(scratch.r6, &ptregs->r6);
-	REG_O_ONE(scratch.r5, &ptregs->r5);
-	REG_O_ONE(scratch.r4, &ptregs->r4);
-	REG_O_ONE(scratch.r3, &ptregs->r3);
-	REG_O_ONE(scratch.r2, &ptregs->r2);
-	REG_O_ONE(scratch.r1, &ptregs->r1);
-	REG_O_ONE(scratch.r0, &ptregs->r0);
-	REG_O_ONE(scratch.sp, &ptregs->sp);
-
-	REG_O_ZERO(pad2);
-
-	REG_O_ONE(callee.r25, &cregs->r25);
-	REG_O_ONE(callee.r24, &cregs->r24);
-	REG_O_ONE(callee.r23, &cregs->r23);
-	REG_O_ONE(callee.r22, &cregs->r22);
-	REG_O_ONE(callee.r21, &cregs->r21);
-	REG_O_ONE(callee.r20, &cregs->r20);
-	REG_O_ONE(callee.r19, &cregs->r19);
-	REG_O_ONE(callee.r18, &cregs->r18);
-	REG_O_ONE(callee.r17, &cregs->r17);
-	REG_O_ONE(callee.r16, &cregs->r16);
-	REG_O_ONE(callee.r15, &cregs->r15);
-	REG_O_ONE(callee.r14, &cregs->r14);
-	REG_O_ONE(callee.r13, &cregs->r13);
-
-	REG_O_ONE(efa, &target->thread.fault_address);
-
-	if (!ret) {
-		if (in_brkpt_trap(ptregs)) {
-			stop_pc_val = target->thread.fault_address;
-			pr_debug("\t\tstop_pc (brk-pt)\n");
-		} else {
-			stop_pc_val = ptregs->ret;
-			pr_debug("\t\tstop_pc (others)\n");
-		}
-
-		REG_O_ONE(stop_pc, &stop_pc_val);
+	membuf_zero(&to, 4);	// pad
+	membuf_store(&to, ptregs->bta);
+	membuf_store(&to, ptregs->lp_start);
+	membuf_store(&to, ptregs->lp_end);
+	membuf_store(&to, ptregs->lp_count);
+	membuf_store(&to, ptregs->status32);
+	membuf_store(&to, ptregs->ret);
+	membuf_store(&to, ptregs->blink);
+	membuf_store(&to, ptregs->fp);
+	membuf_store(&to, ptregs->r26);	// gp
+	membuf_store(&to, ptregs->r12);
+	membuf_store(&to, ptregs->r11);
+	membuf_store(&to, ptregs->r10);
+	membuf_store(&to, ptregs->r9);
+	membuf_store(&to, ptregs->r8);
+	membuf_store(&to, ptregs->r7);
+	membuf_store(&to, ptregs->r6);
+	membuf_store(&to, ptregs->r5);
+	membuf_store(&to, ptregs->r4);
+	membuf_store(&to, ptregs->r3);
+	membuf_store(&to, ptregs->r2);
+	membuf_store(&to, ptregs->r1);
+	membuf_store(&to, ptregs->r0);
+	membuf_store(&to, ptregs->sp);
+	membuf_zero(&to, 4);	// pad2
+	membuf_store(&to, cregs->r25);
+	membuf_store(&to, cregs->r24);
+	membuf_store(&to, cregs->r23);
+	membuf_store(&to, cregs->r22);
+	membuf_store(&to, cregs->r21);
+	membuf_store(&to, cregs->r20);
+	membuf_store(&to, cregs->r19);
+	membuf_store(&to, cregs->r18);
+	membuf_store(&to, cregs->r17);
+	membuf_store(&to, cregs->r16);
+	membuf_store(&to, cregs->r15);
+	membuf_store(&to, cregs->r14);
+	membuf_store(&to, cregs->r13);
+	membuf_store(&to, target->thread.fault_address); // efa
+
+	if (in_brkpt_trap(ptregs)) {
+		stop_pc_val = target->thread.fault_address;
+		pr_debug("\t\tstop_pc (brk-pt)\n");
+	} else {
+		stop_pc_val = ptregs->ret;
+		pr_debug("\t\tstop_pc (others)\n");
 	}
 
-	return ret;
+	return membuf_store(&to, stop_pc_val); // stop_pc
 }
 
 static int genregs_set(struct task_struct *target,
@@ -184,25 +157,20 @@ static int genregs_set(struct task_struct *target,
 #ifdef CONFIG_ISA_ARCV2
 static int arcv2regs_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+		       struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	int ret, copy_sz;
 
 	if (IS_ENABLED(CONFIG_ARC_HAS_ACCL_REGS))
-		copy_sz = sizeof(struct user_regs_arcv2);
-	else
-		copy_sz = 4;	/* r30 only */
+		/*
+		 * itemized copy not needed like above as layout of regs (r30,r58,r59)
+		 * is exactly same in kernel (pt_regs) and userspace (user_regs_arcv2)
+		 */
+		return membuf_write(&to, &regs->r30, sizeof(struct user_regs_arcv2));
 
-	/*
-	 * itemized copy not needed like above as layout of regs (r30,r58,r59)
-	 * is exactly same in kernel (pt_regs) and userspace (user_regs_arcv2)
-	 */
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, &regs->r30,
-				  0, copy_sz);
 
-	return ret;
+	membuf_write(&to, &regs->r30, 4); /* r30 only */
+	return membuf_zero(&to, sizeof(struct user_regs_arcv2) - 4);
 }
 
 static int arcv2regs_set(struct task_struct *target,
@@ -237,7 +205,7 @@ static const struct user_regset arc_regsets[] = {
 	       .n = ELF_NGREG,
 	       .size = sizeof(unsigned long),
 	       .align = sizeof(unsigned long),
-	       .get = genregs_get,
+	       .get2 = genregs_get,
 	       .set = genregs_set,
 	},
 #ifdef CONFIG_ISA_ARCV2
@@ -246,7 +214,7 @@ static const struct user_regset arc_regsets[] = {
 	       .n = ELF_ARCV2REG,
 	       .size = sizeof(unsigned long),
 	       .align = sizeof(unsigned long),
-	       .get = arcv2regs_get,
+	       .get2 = arcv2regs_get,
 	       .set = arcv2regs_set,
 	},
 #endif
-- 
2.11.0

[PATCH 28/41] ia64: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/ia64/kernel/ptrace.c | 149 ++++++++++++++++++----------------------------
 1 file changed, 58 insertions(+), 91 deletions(-)

diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c
index eae7b094f608..990ac8caae8c 100644
--- a/arch/ia64/kernel/ptrace.c
+++ b/arch/ia64/kernel/ptrace.c
@@ -1489,9 +1489,17 @@ access_elf_reg(struct task_struct *target, struct unw_frame_info *info,
 		return access_elf_areg(target, info, addr, data, write_access);
 }
 
+struct regset_membuf {
+	struct membuf to;
+	int ret;
+};
+
 void do_gpregs_get(struct unw_frame_info *info, void *arg)
 {
-	struct regset_getset *dst = arg;
+	struct regset_membuf *dst = arg;
+	struct membuf to = dst->to;
+	unsigned int n;
+	elf_greg_t reg;
 
 	if (unw_unwind_to_user(info) < 0)
 		return;
@@ -1509,35 +1517,13 @@ void do_gpregs_get(struct unw_frame_info *info, void *arg)
 
 
 	/* Skip r0 */
-	if (dst->count > 0 && dst->pos < ELF_GR_OFFSET(1)) {
-		dst->ret = user_regset_copyout_zero(&dst->pos, &dst->count,
-						      &dst->u.get.kbuf,
-						      &dst->u.get.ubuf,
-						      0, ELF_GR_OFFSET(1));
-		if (dst->ret)
+	membuf_zero(&to, 8);
+	for (n = 8; to.left && n < ELF_AR_END_OFFSET; n += 8) {
+		if (access_elf_reg(info->task, info, n, &reg, 0) < 0) {
+			dst->ret = -EIO;
 			return;
-	}
-
-	while (dst->count && dst->pos < ELF_AR_END_OFFSET) {
-		unsigned int n, from, to;
-		elf_greg_t tmp[16];
-
-		from = dst->pos;
-		to = from + min(dst->count, (unsigned)sizeof(tmp));
-		if (to > ELF_AR_END_OFFSET)
-			to = ELF_AR_END_OFFSET;
-		for (n = 0; from < to; from += sizeof(elf_greg_t), n++) {
-			if (access_elf_reg(dst->target, info, from,
-						&tmp[n], 0) < 0) {
-				dst->ret = -EIO;
-				return;
-			}
 		}
-		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
-				&dst->u.get.kbuf, &dst->u.get.ubuf, tmp,
-				dst->pos, to);
-		if (dst->ret)
-			return;
+		membuf_store(&to, reg);
 	}
 }
 
@@ -1588,60 +1574,36 @@ void do_gpregs_set(struct unw_frame_info *info, void *arg)
 
 void do_fpregs_get(struct unw_frame_info *info, void *arg)
 {
-	struct regset_getset *dst = arg;
-	struct task_struct *task = dst->target;
-	elf_fpreg_t tmp[30];
-	int index, min_copy, i;
+	struct task_struct *task = info->task;
+	struct regset_membuf *dst = arg;
+	struct membuf to = dst->to;
+	elf_fpreg_t reg;
+	unsigned int n;
 
 	if (unw_unwind_to_user(info) < 0)
 		return;
 
 	/* Skip pos 0 and 1 */
-	if (dst->count > 0 && dst->pos < ELF_FP_OFFSET(2)) {
-		dst->ret = user_regset_copyout_zero(&dst->pos, &dst->count,
-						      &dst->u.get.kbuf,
-						      &dst->u.get.ubuf,
-						      0, ELF_FP_OFFSET(2));
-		if (dst->count == 0 || dst->ret)
-			return;
-	}
+	membuf_zero(&to, 2 * sizeof(elf_fpreg_t));
 
 	/* fr2-fr31 */
-	if (dst->count > 0 && dst->pos < ELF_FP_OFFSET(32)) {
-		index = (dst->pos - ELF_FP_OFFSET(2)) / sizeof(elf_fpreg_t);
-
-		min_copy = min(((unsigned int)ELF_FP_OFFSET(32)),
-				dst->pos + dst->count);
-		for (i = dst->pos; i < min_copy; i += sizeof(elf_fpreg_t),
-				index++)
-			if (unw_get_fr(info, i / sizeof(elf_fpreg_t),
-					 &tmp[index])) {
-				dst->ret = -EIO;
-				return;
-			}
-		dst->ret = user_regset_copyout(&dst->pos, &dst->count,
-				&dst->u.get.kbuf, &dst->u.get.ubuf, tmp,
-				ELF_FP_OFFSET(2), ELF_FP_OFFSET(32));
-		if (dst->count == 0 || dst->ret)
+	for (n = 2; to.left && n < 32; n++) {
+		if (unw_get_fr(info, n, &reg)) {
+			dst->ret = -EIO;
 			return;
+		}
+		membuf_write(&to, &reg, sizeof(reg));
 	}
 
 	/* fph */
-	if (dst->count > 0) {
-		ia64_flush_fph(dst->target);
-		if (task->thread.flags & IA64_THREAD_FPH_VALID)
-			dst->ret = user_regset_copyout(
-				&dst->pos, &dst->count,
-				&dst->u.get.kbuf, &dst->u.get.ubuf,
-				&dst->target->thread.fph,
-				ELF_FP_OFFSET(32), -1);
-		else
-			/* Zero fill instead.  */
-			dst->ret = user_regset_copyout_zero(
-				&dst->pos, &dst->count,
-				&dst->u.get.kbuf, &dst->u.get.ubuf,
-				ELF_FP_OFFSET(32), -1);
-	}
+	if (!to.left)
+		return;
+
+	ia64_flush_fph(task);
+	if (task->thread.flags & IA64_THREAD_FPH_VALID)
+		membuf_write(&to, &task->thread.fph, 96 * sizeof(reg));
+	else
+		membuf_zero(&to, 96 * sizeof(reg));
 }
 
 void do_fpregs_set(struct unw_frame_info *info, void *arg)
@@ -1717,6 +1679,20 @@ void do_fpregs_set(struct unw_frame_info *info, void *arg)
 	}
 }
 
+static void
+unwind_and_call(void (*call)(struct unw_frame_info *, void *),
+	       struct task_struct *target, void *data)
+{
+	if (target == current)
+		unw_init_running(call, data);
+	else {
+		struct unw_frame_info info;
+		memset(&info, 0, sizeof(info));
+		unw_init_from_blocked_task(&info, target);
+		(*call)(&info, data);
+	}
+}
+
 static int
 do_regset_call(void (*call)(struct unw_frame_info *, void *),
 	       struct task_struct *target,
@@ -1728,27 +1704,18 @@ do_regset_call(void (*call)(struct unw_frame_info *, void *),
 				 .pos = pos, .count = count,
 				 .u.set = { .kbuf = kbuf, .ubuf = ubuf },
 				 .ret = 0 };
-
-	if (target == current)
-		unw_init_running(call, &info);
-	else {
-		struct unw_frame_info ufi;
-		memset(&ufi, 0, sizeof(ufi));
-		unw_init_from_blocked_task(&ufi, target);
-		(*call)(&ufi, &info);
-	}
-
+	unwind_and_call(call, target, &info);
 	return info.ret;
 }
 
 static int
 gpregs_get(struct task_struct *target,
 	   const struct user_regset *regset,
-	   unsigned int pos, unsigned int count,
-	   void *kbuf, void __user *ubuf)
+	   struct membuf to)
 {
-	return do_regset_call(do_gpregs_get, target, regset, pos, count,
-		kbuf, ubuf);
+	struct regset_membuf info = {.to = to};
+	unwind_and_call(do_gpregs_get, target, &info);
+	return info.ret;
 }
 
 static int gpregs_set(struct task_struct *target,
@@ -1790,11 +1757,11 @@ fpregs_active(struct task_struct *target, const struct user_regset *regset)
 
 static int fpregs_get(struct task_struct *target,
 		const struct user_regset *regset,
-		unsigned int pos, unsigned int count,
-		void *kbuf, void __user *ubuf)
+		struct membuf to)
 {
-	return do_regset_call(do_fpregs_get, target, regset, pos, count,
-		kbuf, ubuf);
+	struct regset_membuf info = {.to = to};
+	unwind_and_call(do_fpregs_get, target, &info);
+	return info.ret;
 }
 
 static int fpregs_set(struct task_struct *target,
@@ -2033,14 +2000,14 @@ static const struct user_regset native_regsets[] = {
 		.core_note_type = NT_PRSTATUS,
 		.n = ELF_NGREG,
 		.size = sizeof(elf_greg_t), .align = sizeof(elf_greg_t),
-		.get = gpregs_get, .set = gpregs_set,
+		.get2 = gpregs_get, .set = gpregs_set,
 		.writeback = gpregs_writeback
 	},
 	{
 		.core_note_type = NT_PRFPREG,
 		.n = ELF_NFPREG,
 		.size = sizeof(elf_fpreg_t), .align = sizeof(elf_fpreg_t),
-		.get = fpregs_get, .set = fpregs_set, .active = fpregs_active
+		.get2 = fpregs_get, .set = fpregs_set, .active = fpregs_active
 	},
 };
 
-- 
2.11.0

[PATCH 29/41] c6x: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/c6x/kernel/ptrace.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/arch/c6x/kernel/ptrace.c b/arch/c6x/kernel/ptrace.c
index 67af1562da86..d2402de2bc14 100644
--- a/arch/c6x/kernel/ptrace.c
+++ b/arch/c6x/kernel/ptrace.c
@@ -57,14 +57,9 @@ static inline int put_reg(struct task_struct *task,
 
 static int gpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	struct pt_regs *regs = task_pt_regs(target);
-
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   regs,
-				   0, sizeof(*regs));
+	return membuf_write(&to, task_pt_regs(target), sizeof(struct pt_regs));
 }
 
 enum c6x_regset {
@@ -77,7 +72,7 @@ static const struct user_regset c6x_regsets[] = {
 		.n = ELF_NGREG,
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = gpr_get,
+		.get2 = gpr_get,
 	},
 };
 
-- 
2.11.0

[PATCH 30/41] riscv: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Note: riscv_fpr_get() used to forget to zero-pad at the end.
Not worth -stable...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/riscv/kernel/ptrace.c | 33 +++++++++++----------------------
 1 file changed, 11 insertions(+), 22 deletions(-)

diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c
index 444dc7b0fd78..a6896449eff8 100644
--- a/arch/riscv/kernel/ptrace.c
+++ b/arch/riscv/kernel/ptrace.c
@@ -30,13 +30,10 @@ enum riscv_regset {
 
 static int riscv_gpr_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
-	struct pt_regs *regs;
-
-	regs = task_pt_regs(target);
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, regs, 0, -1);
+	return membuf_write(&to, task_pt_regs(target),
+			    sizeof(struct user_regs_struct));
 }
 
 static int riscv_gpr_set(struct task_struct *target,
@@ -55,21 +52,13 @@ static int riscv_gpr_set(struct task_struct *target,
 #ifdef CONFIG_FPU
 static int riscv_fpr_get(struct task_struct *target,
 			 const struct user_regset *regset,
-			 unsigned int pos, unsigned int count,
-			 void *kbuf, void __user *ubuf)
+			 struct membuf to)
 {
-	int ret;
 	struct __riscv_d_ext_state *fstate = &target->thread.fstate;
 
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, fstate, 0,
-				  offsetof(struct __riscv_d_ext_state, fcsr));
-	if (!ret) {
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, fstate, 0,
-					  offsetof(struct __riscv_d_ext_state, fcsr) +
-					  sizeof(fstate->fcsr));
-	}
-
-	return ret;
+	membuf_write(&to, fstate, offsetof(struct __riscv_d_ext_state, fcsr));
+	membuf_store(&to, fstate->fcsr);
+	return membuf_zero(&to, 4);	// explicitly pad
 }
 
 static int riscv_fpr_set(struct task_struct *target,
@@ -98,8 +87,8 @@ static const struct user_regset riscv_user_regset[] = {
 		.n = ELF_NGREG,
 		.size = sizeof(elf_greg_t),
 		.align = sizeof(elf_greg_t),
-		.get = &riscv_gpr_get,
-		.set = &riscv_gpr_set,
+		.get2 = riscv_gpr_get,
+		.set = riscv_gpr_set,
 	},
 #ifdef CONFIG_FPU
 	[REGSET_F] = {
@@ -107,8 +96,8 @@ static const struct user_regset riscv_user_regset[] = {
 		.n = ELF_NFPREG,
 		.size = sizeof(elf_fpreg_t),
 		.align = sizeof(elf_fpreg_t),
-		.get = &riscv_fpr_get,
-		.set = &riscv_fpr_set,
+		.get2 = riscv_fpr_get,
+		.set = riscv_fpr_set,
 	},
 #endif
 };
-- 
2.11.0

[PATCH 31/41] openrisc: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/openrisc/kernel/ptrace.c | 26 ++++++--------------------
 1 file changed, 6 insertions(+), 20 deletions(-)

diff --git a/arch/openrisc/kernel/ptrace.c b/arch/openrisc/kernel/ptrace.c
index c8f47a623754..0b762fc00999 100644
--- a/arch/openrisc/kernel/ptrace.c
+++ b/arch/openrisc/kernel/ptrace.c
@@ -44,29 +44,15 @@
  */
 static int genregs_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user * ubuf)
+		       struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
-	int ret;
 
 	/* r0 */
-	ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf, 0, 4);
-
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  regs->gpr+1, 4, 4*32);
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &regs->pc, 4*32, 4*33);
-	if (!ret)
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-					  &regs->sr, 4*33, 4*34);
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					       4*34, -1);
-
-	return ret;
+	membuf_zero(&to, 4);
+	membuf_write(&to, regs->gpr + 1, 31 * 4);
+	membuf_store(&to, regs->pc);
+	return membuf_store(&to, regs->sr);
 }
 
 /*
@@ -114,7 +100,7 @@ static const struct user_regset or1k_regsets[] = {
 			    .n = ELF_NGREG,
 			    .size = sizeof(long),
 			    .align = sizeof(long),
-			    .get = genregs_get,
+			    .get2 = genregs_get,
 			    .set = genregs_set,
 			    },
 };
-- 
2.11.0

[PATCH 32/41] h8300: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/h8300/kernel/ptrace.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)

diff --git a/arch/h8300/kernel/ptrace.c b/arch/h8300/kernel/ptrace.c
index 0dc1c8f622bc..a99cbe91a6d3 100644
--- a/arch/h8300/kernel/ptrace.c
+++ b/arch/h8300/kernel/ptrace.c
@@ -87,20 +87,15 @@ int h8300_put_reg(struct task_struct *task, int regno, unsigned long data)
 
 static int regs_get(struct task_struct *target,
 		    const struct user_regset *regset,
-		    unsigned int pos, unsigned int count,
-		    void *kbuf, void __user *ubuf)
+		    struct membuf to)
 {
 	int r;
-	struct user_regs_struct regs;
-	long *reg = (long *)&regs;
 
-	/* build user regs in buffer */
-	BUILD_BUG_ON(sizeof(regs) % sizeof(long) != 0);
-	for (r = 0; r < sizeof(regs) / sizeof(long); r++)
-		*reg++ = h8300_get_reg(target, r);
+	BUILD_BUG_ON(sizeof(struct user_regs_struct) % sizeof(long) != 0);
+	for (r = 0; r < ELF_NGREG; r++)
+		membuf_store(&to, h8300_get_reg(target, r));
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &regs, 0, sizeof(regs));
+	return 0;
 }
 
 static int regs_set(struct task_struct *target,
@@ -139,7 +134,7 @@ static const struct user_regset h8300_regsets[] = {
 		.n		= ELF_NGREG,
 		.size		= sizeof(long),
 		.align		= sizeof(long),
-		.get		= regs_get,
+		.get2		= regs_get,
 		.set		= regs_set,
 	},
 };
-- 
2.11.0

[PATCH 33/41] hexagon: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/hexagon/kernel/ptrace.c | 62 +++++++++++++++-----------------------------
 1 file changed, 21 insertions(+), 41 deletions(-)

diff --git a/arch/hexagon/kernel/ptrace.c b/arch/hexagon/kernel/ptrace.c
index dcbf7ea960cc..fa6287d1a061 100644
--- a/arch/hexagon/kernel/ptrace.c
+++ b/arch/hexagon/kernel/ptrace.c
@@ -35,58 +35,38 @@ void user_disable_single_step(struct task_struct *child)
 
 static int genregs_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   srtuct membuf to)
 {
-	int ret;
-	unsigned int dummy;
 	struct pt_regs *regs = task_pt_regs(target);
 
-
-	if (!regs)
-		return -EIO;
-
 	/* The general idea here is that the copyout must happen in
 	 * exactly the same order in which the userspace expects these
 	 * regs. Now, the sequence in userspace does not match the
 	 * sequence in the kernel, so everything past the 32 gprs
 	 * happens one at a time.
 	 */
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  &regs->r00, 0, 32*sizeof(unsigned long));
-
-#define ONEXT(KPT_REG, USR_REG) \
-	if (!ret) \
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, \
-			KPT_REG, offsetof(struct user_regs_struct, USR_REG), \
-			offsetof(struct user_regs_struct, USR_REG) + \
-				 sizeof(unsigned long));
-
+	membuf_write(&to, &regs->r00, 32*sizeof(unsigned long));
 	/* Must be exactly same sequence as struct user_regs_struct */
-	ONEXT(&regs->sa0, sa0);
-	ONEXT(&regs->lc0, lc0);
-	ONEXT(&regs->sa1, sa1);
-	ONEXT(&regs->lc1, lc1);
-	ONEXT(&regs->m0, m0);
-	ONEXT(&regs->m1, m1);
-	ONEXT(&regs->usr, usr);
-	ONEXT(&regs->preds, p3_0);
-	ONEXT(&regs->gp, gp);
-	ONEXT(&regs->ugp, ugp);
-	ONEXT(&pt_elr(regs), pc);
-	dummy = pt_cause(regs);
-	ONEXT(&dummy, cause);
-	ONEXT(&pt_badva(regs), badva);
+	membuf_store(&to, regs->sa0);
+	membuf_store(&to, regs->lc0);
+	membuf_store(&to, regs->sa1);
+	membuf_store(&to, regs->lc1);
+	membuf_store(&to, regs->m0);
+	membuf_store(&to, regs->m1);
+	membuf_store(&to, regs->usr);
+	membuf_store(&to, regs->p3_0);
+	membuf_store(&to, regs->gp);
+	membuf_store(&to, regs->ugp);
+	membuf_store(&to, pt_elr(regs)); // pc
+	membuf_store(&to, (unsigned long)pt_cause(regs)); // cause
+	membuf_store(&to, pt_badva(regs)); // badva
 #if CONFIG_HEXAGON_ARCH_VERSION >=4
-	ONEXT(&regs->cs0, cs0);
-	ONEXT(&regs->cs1, cs1);
+	membuf_store(&to, regs->cs0);
+	membuf_store(&to, regs->cs1);
+	return membuf_zero(&to, sizeof(unsigned long));
+#else
+	return membuf_zero(&to, 3 * sizeof(unsigned long));
 #endif
-
-	/* Pad the rest with zeros, if needed */
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					offsetof(struct user_regs_struct, pad1), -1);
-	return ret;
 }
 
 static int genregs_set(struct task_struct *target,
@@ -159,7 +139,7 @@ static const struct user_regset hexagon_regsets[] = {
 		.n = ELF_NGREG,
 		.size = sizeof(unsigned long),
 		.align = sizeof(unsigned long),
-		.get = genregs_get,
+		.get2 = genregs_get,
 		.set = genregs_set,
 	},
 };
-- 
2.11.0

[PATCH 34/41] nios2: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/nios2/kernel/ptrace.c | 51 ++++++++++++++--------------------------------
 1 file changed, 15 insertions(+), 36 deletions(-)

diff --git a/arch/nios2/kernel/ptrace.c b/arch/nios2/kernel/ptrace.c
index de97bcb7dd44..2214f95847cd 100644
--- a/arch/nios2/kernel/ptrace.c
+++ b/arch/nios2/kernel/ptrace.c
@@ -21,45 +21,24 @@
 
 static int genregs_get(struct task_struct *target,
 		       const struct user_regset *regset,
-		       unsigned int pos, unsigned int count,
-		       void *kbuf, void __user *ubuf)
+		       struct membuf to)
 {
 	const struct pt_regs *regs = task_pt_regs(target);
 	const struct switch_stack *sw = (struct switch_stack *)regs - 1;
-	int ret = 0;
-
-#define REG_O_ZERO_RANGE(START, END)		\
-	if (!ret)					\
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf, \
-			START * 4, (END * 4) + 4);
-
-#define REG_O_ONE(PTR, LOC)	\
-	if (!ret)			\
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, PTR, \
-			LOC * 4, (LOC * 4) + 4);
 
-#define REG_O_RANGE(PTR, START, END)	\
-	if (!ret)				\
-		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, PTR, \
-			START * 4, (END * 4) + 4);
-
-	REG_O_ZERO_RANGE(PTR_R0, PTR_R0);
-	REG_O_RANGE(&regs->r1, PTR_R1, PTR_R7);
-	REG_O_RANGE(&regs->r8, PTR_R8, PTR_R15);
-	REG_O_RANGE(sw, PTR_R16, PTR_R23);
-	REG_O_ZERO_RANGE(PTR_R24, PTR_R25); /* et and bt */
-	REG_O_ONE(&regs->gp, PTR_GP);
-	REG_O_ONE(&regs->sp, PTR_SP);
-	REG_O_ONE(&regs->fp, PTR_FP);
-	REG_O_ONE(&regs->ea, PTR_EA);
-	REG_O_ZERO_RANGE(PTR_BA, PTR_BA);
-	REG_O_ONE(&regs->ra, PTR_RA);
-	REG_O_ONE(&regs->ea, PTR_PC); /* use ea for PC */
-	if (!ret)
-		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					 PTR_STATUS * 4, -1);
-
-	return ret;
+	membuf_zero(&to, 4); // R0
+	membuf_write(&to, &regs->r1, 7 * 4); // R1..R7
+	membuf_write(&to, &regs->r8, 8 * 4); // R8..R15
+	membuf_write(&to, sw, 8 * 4); // R16..R23
+	membuf_zero(&to, 2 * 4); /* et and bt */
+	membuf_store(&to, regs->gp);
+	membuf_store(&to, regs->sp);
+	membuf_store(&to, regs->fp);
+	membuf_store(&to, regs->ea);
+	membuf_zero(&to, 4); // PTR_BA
+	membuf_store(&to, regs->ra);
+	membuf_store(&to, regs->ea); /* use ea for PC */
+	return membuf_zero(&to, (NUM_PTRACE_REG - PTR_PC) * 4);
 }
 
 /*
@@ -121,7 +100,7 @@ static const struct user_regset nios2_regsets[] = {
 		.n = NUM_PTRACE_REG,
 		.size = sizeof(unsigned long),
 		.align = sizeof(unsigned long),
-		.get = genregs_get,
+		.get2 = genregs_get,
 		.set = genregs_set,
 	}
 };
-- 
2.11.0

[PATCH 35/41] nds32: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/nds32/kernel/ptrace.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/arch/nds32/kernel/ptrace.c b/arch/nds32/kernel/ptrace.c
index eaaf7a999b20..feaf9b33f517 100644
--- a/arch/nds32/kernel/ptrace.c
+++ b/arch/nds32/kernel/ptrace.c
@@ -13,11 +13,10 @@ enum nds32_regset {
 
 static int gpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user * ubuf)
+		   struct membuf to)
 {
-	struct user_pt_regs *uregs = &task_pt_regs(target)->user_regs;
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, uregs, 0, -1);
+	return membuf_write(&to, &task_pt_regs(target)->user_regs,
+				sizeof(struct user_pt_regs));
 }
 
 static int gpr_set(struct task_struct *target, const struct user_regset *regset,
@@ -41,7 +40,7 @@ static const struct user_regset nds32_regsets[] = {
 			.n = sizeof(struct user_pt_regs) / sizeof(u32),
 			.size = sizeof(elf_greg_t),
 			.align = sizeof(elf_greg_t),
-			.get = gpr_get,
+			.get2 = gpr_get,
 			.set = gpr_set}
 };
 
-- 
2.11.0

[PATCH 36/41] parisc: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/parisc/kernel/ptrace.c | 84 +++++++++------------------------------------
 1 file changed, 16 insertions(+), 68 deletions(-)

diff --git a/arch/parisc/kernel/ptrace.c b/arch/parisc/kernel/ptrace.c
index b51418ad8655..6dd7a4debb9c 100644
--- a/arch/parisc/kernel/ptrace.c
+++ b/arch/parisc/kernel/ptrace.c
@@ -391,31 +391,11 @@ void do_syscall_trace_exit(struct pt_regs *regs)
 
 static int fpr_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
 	struct pt_regs *regs = task_regs(target);
-	__u64 *k = kbuf;
-	__u64 __user *u = ubuf;
-	__u64 reg;
-
-	pos /= sizeof(reg);
-	count /= sizeof(reg);
-
-	if (kbuf)
-		for (; count > 0 && pos < ELF_NFPREG; --count)
-			*k++ = regs->fr[pos++];
-	else
-		for (; count > 0 && pos < ELF_NFPREG; --count)
-			if (__put_user(regs->fr[pos++], u++))
-				return -EFAULT;
 
-	kbuf = k;
-	ubuf = u;
-	pos *= sizeof(reg);
-	count *= sizeof(reg);
-	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					ELF_NFPREG * sizeof(reg), -1);
+	return membuf_write(&to, regs->fr, ELF_NFPREG * sizeof(__u64));
 }
 
 static int fpr_set(struct task_struct *target,
@@ -527,30 +507,14 @@ static void set_reg(struct pt_regs *regs, int num, unsigned long val)
 
 static int gpr_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
 	struct pt_regs *regs = task_regs(target);
-	unsigned long *k = kbuf;
-	unsigned long __user *u = ubuf;
-	unsigned long reg;
+	unsigned int pos;
 
-	pos /= sizeof(reg);
-	count /= sizeof(reg);
-
-	if (kbuf)
-		for (; count > 0 && pos < ELF_NGREG; --count)
-			*k++ = get_reg(regs, pos++);
-	else
-		for (; count > 0 && pos < ELF_NGREG; --count)
-			if (__put_user(get_reg(regs, pos++), u++))
-				return -EFAULT;
-	kbuf = k;
-	ubuf = u;
-	pos *= sizeof(reg);
-	count *= sizeof(reg);
-	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					ELF_NGREG * sizeof(reg), -1);
+	for (pos = 0; pos < ELF_NGREG; pos++)
+		membuf_store(&to, get_reg(regs, pos));
+	return 0;
 }
 
 static int gpr_set(struct task_struct *target,
@@ -588,12 +552,12 @@ static const struct user_regset native_regsets[] = {
 	[REGSET_GENERAL] = {
 		.core_note_type = NT_PRSTATUS, .n = ELF_NGREG,
 		.size = sizeof(long), .align = sizeof(long),
-		.get = gpr_get, .set = gpr_set
+		.get2 = gpr_get, .set = gpr_set
 	},
 	[REGSET_FP] = {
 		.core_note_type = NT_PRFPREG, .n = ELF_NFPREG,
 		.size = sizeof(__u64), .align = sizeof(__u64),
-		.get = fpr_get, .set = fpr_set
+		.get2 = fpr_get, .set = fpr_set
 	}
 };
 
@@ -607,31 +571,15 @@ static const struct user_regset_view user_parisc_native_view = {
 
 static int gpr32_get(struct task_struct *target,
 		     const struct user_regset *regset,
-		     unsigned int pos, unsigned int count,
-		     void *kbuf, void __user *ubuf)
+		     struct membuf to)
 {
 	struct pt_regs *regs = task_regs(target);
-	compat_ulong_t *k = kbuf;
-	compat_ulong_t __user *u = ubuf;
-	compat_ulong_t reg;
+	unsigned int pos;
 
-	pos /= sizeof(reg);
-	count /= sizeof(reg);
+	for (pos = 0; pos < ELF_NGREG; pos++)
+		membuf_store(&to, (compat_ulong_t)get_reg(regs, pos));
 
-	if (kbuf)
-		for (; count > 0 && pos < ELF_NGREG; --count)
-			*k++ = get_reg(regs, pos++);
-	else
-		for (; count > 0 && pos < ELF_NGREG; --count)
-			if (__put_user((compat_ulong_t) get_reg(regs, pos++), u++))
-				return -EFAULT;
-
-	kbuf = k;
-	ubuf = u;
-	pos *= sizeof(reg);
-	count *= sizeof(reg);
-	return user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
-					ELF_NGREG * sizeof(reg), -1);
+	return 0;
 }
 
 static int gpr32_set(struct task_struct *target,
@@ -672,12 +620,12 @@ static const struct user_regset compat_regsets[] = {
 	[REGSET_GENERAL] = {
 		.core_note_type = NT_PRSTATUS, .n = ELF_NGREG,
 		.size = sizeof(compat_long_t), .align = sizeof(compat_long_t),
-		.get = gpr32_get, .set = gpr32_set
+		.get2 = gpr32_get, .set = gpr32_set
 	},
 	[REGSET_FP] = {
 		.core_note_type = NT_PRFPREG, .n = ELF_NFPREG,
 		.size = sizeof(__u64), .align = sizeof(__u64),
-		.get = fpr_get, .set = fpr_set
+		.get2 = fpr_get, .set = fpr_set
 	}
 };
 
-- 
2.11.0

[PATCH 37/41] xtensa: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/xtensa/kernel/ptrace.c | 16 ++++++----------
 1 file changed, 6 insertions(+), 10 deletions(-)

diff --git a/arch/xtensa/kernel/ptrace.c b/arch/xtensa/kernel/ptrace.c
index b4c07bd890fe..d46180b69600 100644
--- a/arch/xtensa/kernel/ptrace.c
+++ b/arch/xtensa/kernel/ptrace.c
@@ -37,8 +37,7 @@
 
 static int gpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	struct pt_regs *regs = task_pt_regs(target);
 	struct user_pt_regs newregs = {
@@ -60,8 +59,7 @@ static int gpr_get(struct task_struct *target,
 	       regs->areg,
 	       (WSBITS - regs->windowbase) * 16);
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				   &newregs, 0, -1);
+	return membuf_write(&to, &newregs, sizeof(newregs));
 }
 
 static int gpr_set(struct task_struct *target,
@@ -115,8 +113,7 @@ static int gpr_set(struct task_struct *target,
 
 static int tie_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	int ret;
 	struct pt_regs *regs = task_pt_regs(target);
@@ -141,8 +138,7 @@ static int tie_get(struct task_struct *target,
 	newregs->cp6 = ti->xtregs_cp.cp6;
 	newregs->cp7 = ti->xtregs_cp.cp7;
 #endif
-	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
-				  newregs, 0, -1);
+	ret = membuf_write(&to, newregs, sizeof(*newregs));
 	kfree(newregs);
 	return ret;
 }
@@ -197,7 +193,7 @@ static const struct user_regset xtensa_regsets[] = {
 		.n = sizeof(struct user_pt_regs) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = gpr_get,
+		.get2 = gpr_get,
 		.set = gpr_set,
 	},
 	[REGSET_TIE] = {
@@ -205,7 +201,7 @@ static const struct user_regset xtensa_regsets[] = {
 		.n = sizeof(elf_xtregs_t) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = tie_get,
+		.get2 = tie_get,
 		.set = tie_set,
 	},
 };
-- 
2.11.0

[PATCH 38/41] csky: switch to ->get2()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

NB: WTF is fpregs_get() playing at???

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/csky/kernel/ptrace.c | 24 ++++++++++--------------
 1 file changed, 10 insertions(+), 14 deletions(-)

diff --git a/arch/csky/kernel/ptrace.c b/arch/csky/kernel/ptrace.c
index 944ca2fdcdd9..acd36aeb5222 100644
--- a/arch/csky/kernel/ptrace.c
+++ b/arch/csky/kernel/ptrace.c
@@ -76,17 +76,14 @@ enum csky_regset {
 
 static int gpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
-	struct pt_regs *regs;
-
-	regs = task_pt_regs(target);
+	struct pt_regs *regs = task_pt_regs(target);
 
 	/* Abiv1 regs->tls is fake and we need sync here. */
 	regs->tls = task_thread_info(target)->tp_value;
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, regs, 0, -1);
+	return membuf_write(&to, regs, sizeof(regs));
 }
 
 static int gpr_set(struct task_struct *target,
@@ -114,8 +111,7 @@ static int gpr_set(struct task_struct *target,
 
 static int fpr_get(struct task_struct *target,
 		   const struct user_regset *regset,
-		   unsigned int pos, unsigned int count,
-		   void *kbuf, void __user *ubuf)
+		   struct membuf to)
 {
 	struct user_fp *regs = (struct user_fp *)&target->thread.user_fp;
 
@@ -131,9 +127,9 @@ static int fpr_get(struct task_struct *target,
 	for (i = 0; i < 32; i++)
 		tmp.vr[64 + i] = regs->vr[32 + i];
 
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &tmp, 0, -1);
+	return membuf_write(&to, &tmp, sizeof(tmp));
 #else
-	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, regs, 0, -1);
+	return membuf_write(&to, regs, sizeof(*regs));
 #endif
 }
 
@@ -173,16 +169,16 @@ static const struct user_regset csky_regsets[] = {
 		.n = sizeof(struct pt_regs) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = &gpr_get,
-		.set = &gpr_set,
+		.get2 = gpr_get,
+		.set = gpr_set,
 	},
 	[REGSET_FPR] = {
 		.core_note_type = NT_PRFPREG,
 		.n = sizeof(struct user_fp) / sizeof(u32),
 		.size = sizeof(u32),
 		.align = sizeof(u32),
-		.get = &fpr_get,
-		.set = &fpr_set,
+		.get2 = fpr_get,
+		.set = fpr_set,
 	},
 };
 
-- 
2.11.0

[PATCH 39/41] regset: kill ->get()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

no instances left

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 include/linux/regset.h | 22 ----------------------
 kernel/regset.c        | 24 +++++-------------------
 2 files changed, 5 insertions(+), 41 deletions(-)

diff --git a/include/linux/regset.h b/include/linux/regset.h
index 567a4b63f469..eeabf5cbbbd7 100644
--- a/include/linux/regset.h
+++ b/include/linux/regset.h
@@ -82,27 +82,6 @@ static inline int membuf_write(struct membuf *s, const void *v, size_t size)
 typedef int user_regset_active_fn(struct task_struct *target,
 				  const struct user_regset *regset);
 
-/**
- * user_regset_get_fn - type of @get function in &struct user_regset
- * @target:	thread being examined
- * @regset:	regset being examined
- * @pos:	offset into the regset data to access, in bytes
- * @count:	amount of data to copy, in bytes
- * @kbuf:	if not %NULL, a kernel-space pointer to copy into
- * @ubuf:	if @kbuf is %NULL, a user-space pointer to copy into
- *
- * Fetch register values.  Return %0 on success; -%EIO or -%ENODEV
- * are usual failure returns.  The @pos and @count values are in
- * bytes, but must be properly aligned.  If @kbuf is non-null, that
- * buffer is used and @ubuf is ignored.  If @kbuf is %NULL, then
- * ubuf gives a userland pointer to access directly, and an -%EFAULT
- * return value is possible.
- */
-typedef int user_regset_get_fn(struct task_struct *target,
-			       const struct user_regset *regset,
-			       unsigned int pos, unsigned int count,
-			       void *kbuf, void __user *ubuf);
-
 typedef int user_regset_get2_fn(struct task_struct *target,
 			       const struct user_regset *regset,
 			       struct membuf to);
@@ -235,7 +214,6 @@ typedef unsigned int user_regset_get_size_fn(struct task_struct *target,
  * omitted when there is an @active function and it returns zero.
  */
 struct user_regset {
-	user_regset_get_fn		*get;
 	user_regset_get2_fn		*get2;
 	user_regset_set_fn		*set;
 	user_regset_active_fn		*active;
diff --git a/kernel/regset.c b/kernel/regset.c
index c548fb9f0943..592d90fc13f9 100644
--- a/kernel/regset.c
+++ b/kernel/regset.c
@@ -11,7 +11,7 @@ static int __regset_get(struct task_struct *target,
 	void *p = *data, *to_free = NULL;
 	int res;
 
-	if (!regset->get && !regset->get2)
+	if (!regset->get2)
 		return -EOPNOTSUPP;
 	if (size > regset->n * regset->size)
 		size = regset->n * regset->size;
@@ -20,28 +20,14 @@ static int __regset_get(struct task_struct *target,
 		if (!p)
 			return -ENOMEM;
 	}
-	if (regset->get2) {
-		res = regset->get2(target, regset,
-				   (struct membuf){.p = p, .left = size});
-		if (res < 0) {
-			kfree(to_free);
-			return res;
-		}
-		*data = p;
-		return size - res;
-	}
-	res = regset->get(target, regset, 0, size, p, NULL);
-	if (unlikely(res < 0)) {
+	res = regset->get2(target, regset,
+			   (struct membuf){.p = p, .left = size});
+	if (res < 0) {
 		kfree(to_free);
 		return res;
 	}
 	*data = p;
-	if (regset->get_size) { // arm64-only kludge, will go away
-		unsigned max_size = regset->get_size(target, regset);
-		if (size > max_size)
-			size = max_size;
-	}
-	return size;
+	return size - res;
 }
 
 int regset_get(struct task_struct *target,
-- 
2.11.0

[PATCH 40/41] regset(): kill ->get_size()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

not used anymore

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 arch/arm64/kernel/ptrace.c | 13 -------------
 include/linux/regset.h     | 48 +---------------------------------------------
 2 files changed, 1 insertion(+), 60 deletions(-)

diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index 8745aecffcae..5fb2623b7f13 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -740,18 +740,6 @@ static unsigned int sve_size_from_header(struct user_sve_header const *header)
 	return ALIGN(header->size, SVE_VQ_BYTES);
 }
 
-static unsigned int sve_get_size(struct task_struct *target,
-				 const struct user_regset *regset)
-{
-	struct user_sve_header header;
-
-	if (!system_supports_sve())
-		return 0;
-
-	sve_init_header_from_task(&header, target);
-	return sve_size_from_header(&header);
-}
-
 static int sve_get(struct task_struct *target,
 		   const struct user_regset *regset,
 		   struct membuf to)
@@ -1130,7 +1118,6 @@ static const struct user_regset aarch64_regsets[] = {
 		.align = SVE_VQ_BYTES,
 		.get2 = sve_get,
 		.set = sve_set,
-		.get_size = sve_get_size,
 	},
 #endif
 #ifdef CONFIG_ARM64_PTR_AUTH
diff --git a/include/linux/regset.h b/include/linux/regset.h
index eeabf5cbbbd7..d82bb32e434c 100644
--- a/include/linux/regset.h
+++ b/include/linux/regset.h
@@ -133,28 +133,6 @@ typedef int user_regset_writeback_fn(struct task_struct *target,
 				     int immediate);
 
 /**
- * user_regset_get_size_fn - type of @get_size function in &struct user_regset
- * @target:	thread being examined
- * @regset:	regset being examined
- *
- * This call is optional; usually the pointer is %NULL.
- *
- * When provided, this function must return the current size of regset
- * data, as observed by the @get function in &struct user_regset.  The
- * value returned must be a multiple of @size.  The returned size is
- * required to be valid only until the next time (if any) @regset is
- * modified for @target.
- *
- * This function is intended for dynamically sized regsets.  A regset
- * that is statically sized does not need to implement it.
- *
- * This function should not be called directly: instead, callers should
- * call regset_size() to determine the current size of a regset.
- */
-typedef unsigned int user_regset_get_size_fn(struct task_struct *target,
-					     const struct user_regset *regset);
-
-/**
  * struct user_regset - accessible thread CPU state
  * @n:			Number of slots (registers).
  * @size:		Size in bytes of a slot (register).
@@ -165,7 +143,6 @@ typedef unsigned int user_regset_get_size_fn(struct task_struct *target,
  * @set:		Function to store values.
  * @active:		Function to report if regset is active, or %NULL.
  * @writeback:		Function to write data back to user memory, or %NULL.
- * @get_size:		Function to return the regset's size, or %NULL.
  *
  * This data structure describes a machine resource we call a register set.
  * This is part of the state of an individual thread, not necessarily
@@ -173,12 +150,7 @@ typedef unsigned int user_regset_get_size_fn(struct task_struct *target,
  * similar slots, given by @n.  Each slot is @size bytes, and aligned to
  * @align bytes (which is at least @size).  For dynamically-sized
  * regsets, @n must contain the maximum possible number of slots for the
- * regset, and @get_size must point to a function that returns the
- * current regset size.
- *
- * Callers that need to know only the current size of the regset and do
- * not care about its internal structure should call regset_size()
- * instead of inspecting @n or calling @get_size.
+ * regset.
  *
  * For backward compatibility, the @get and @set methods must pad to, or
  * accept, @n * @size bytes, even if the current regset size is smaller.
@@ -218,7 +190,6 @@ struct user_regset {
 	user_regset_set_fn		*set;
 	user_regset_active_fn		*active;
 	user_regset_writeback_fn	*writeback;
-	user_regset_get_size_fn		*get_size;
 	unsigned int			n;
 	unsigned int 			size;
 	unsigned int 			align;
@@ -422,21 +393,4 @@ static inline int copy_regset_from_user(struct task_struct *target,
 	return regset->set(target, regset, offset, size, NULL, data);
 }
 
-/**
- * regset_size - determine the current size of a regset
- * @target:	thread to be examined
- * @regset:	regset to be examined
- *
- * Note that the returned size is valid only until the next time
- * (if any) @regset is modified for @target.
- */
-static inline unsigned int regset_size(struct task_struct *target,
-				       const struct user_regset *regset)
-{
-	if (!regset->get_size)
-		return regset->n * regset->size;
-	else
-		return regset->get_size(target, regset);
-}
-
 #endif	/* <linux/regset.h> */
-- 
2.11.0

[PATCH 41/41] regset: kill user_regset_copyout{,_zero}()

[Al Viro]

From: Al Viro <viro@zeniv.linux.org.uk>

no callers left

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 include/linux/regset.h | 67 --------------------------------------------------
 1 file changed, 67 deletions(-)

diff --git a/include/linux/regset.h b/include/linux/regset.h
index d82bb32e434c..facfaf300b44 100644
--- a/include/linux/regset.h
+++ b/include/linux/regset.h
@@ -238,44 +238,6 @@ struct user_regset_view {
  */
 const struct user_regset_view *task_user_regset_view(struct task_struct *tsk);
 
-
-/*
- * These are helpers for writing regset get/set functions in arch code.
- * Because @start_pos and @end_pos are always compile-time constants,
- * these are inlined into very little code though they look large.
- *
- * Use one or more calls sequentially for each chunk of regset data stored
- * contiguously in memory.  Call with constants for @start_pos and @end_pos,
- * giving the range of byte positions in the regset that data corresponds
- * to; @end_pos can be -1 if this chunk is at the end of the regset layout.
- * Each call updates the arguments to point past its chunk.
- */
-
-static inline int user_regset_copyout(unsigned int *pos, unsigned int *count,
-				      void **kbuf,
-				      void __user **ubuf, const void *data,
-				      const int start_pos, const int end_pos)
-{
-	if (*count == 0)
-		return 0;
-	BUG_ON(*pos < start_pos);
-	if (end_pos < 0 || *pos < end_pos) {
-		unsigned int copy = (end_pos < 0 ? *count
-				     : min(*count, end_pos - *pos));
-		data += *pos - start_pos;
-		if (*kbuf) {
-			memcpy(*kbuf, data, copy);
-			*kbuf += copy;
-		} else if (__copy_to_user(*ubuf, data, copy))
-			return -EFAULT;
-		else
-			*ubuf += copy;
-		*pos += copy;
-		*count -= copy;
-	}
-	return 0;
-}
-
 static inline int user_regset_copyin(unsigned int *pos, unsigned int *count,
 				     const void **kbuf,
 				     const void __user **ubuf, void *data,
@@ -301,35 +263,6 @@ static inline int user_regset_copyin(unsigned int *pos, unsigned int *count,
 	return 0;
 }
 
-/*
- * These two parallel the two above, but for portions of a regset layout
- * that always read as all-zero or for which writes are ignored.
- */
-static inline int user_regset_copyout_zero(unsigned int *pos,
-					   unsigned int *count,
-					   void **kbuf, void __user **ubuf,
-					   const int start_pos,
-					   const int end_pos)
-{
-	if (*count == 0)
-		return 0;
-	BUG_ON(*pos < start_pos);
-	if (end_pos < 0 || *pos < end_pos) {
-		unsigned int copy = (end_pos < 0 ? *count
-				     : min(*count, end_pos - *pos));
-		if (*kbuf) {
-			memset(*kbuf, 0, copy);
-			*kbuf += copy;
-		} else if (clear_user(*ubuf, copy))
-			return -EFAULT;
-		else
-			*ubuf += copy;
-		*pos += copy;
-		*count -= copy;
-	}
-	return 0;
-}
-
 static inline int user_regset_copyin_ignore(unsigned int *pos,
 					    unsigned int *count,
 					    const void **kbuf,
-- 
2.11.0

Re: [PATCH 18/41] regset: new method and helpers for it

[Linus Torvalds]

On Mon, Jun 29, 2020 at 11:28 AM Al Viro <viro@zeniv.linux.org.uk> wrote:
>
> ->get2() takes task+regset+buffer, returns the amount of free space
> left in the buffer on success and -E... on error.

Can we please give it a better name than "get2"?

That's not a great name to begin with, and it's a completely
nonsensical name by the end of this series when you've removed the
original "get" function.

So either:

 (a) add one final patch to rename "get2" all back to "get" after you
got rid of the old "get"

 (b) or just call it something better to begin with. Maybe just
"get_regset" instead?

I'd prefer (b) just because I think it will be a lot clearer if we
ever end up having old patches forward-ported (or, more likely,
newpatches back-ported), so having a "get" function that changed
semantics but got back the old name sounds bad to me.

Other than that, I can't really argue with the numbers:

 41 files changed, 1368 insertions(+), 2347 deletions(-)

looks good to me, and the code seems more understandable.

But I only scanned the individual patches, maybe I missed some other horror.

               Linus

Re: [PATCH 18/41] regset: new method and helpers for it

[Al Viro]

On Mon, Jun 29, 2020 at 12:23:34PM -0700, Linus Torvalds wrote:
> On Mon, Jun 29, 2020 at 11:28 AM Al Viro <viro@zeniv.linux.org.uk> wrote:
> >
> > ->get2() takes task+regset+buffer, returns the amount of free space
> > left in the buffer on success and -E... on error.
> 
> Can we please give it a better name than "get2"?
> 
> That's not a great name to begin with, and it's a completely
> nonsensical name by the end of this series when you've removed the
> original "get" function.
> 
> So either:
> 
>  (a) add one final patch to rename "get2" all back to "get" after you
> got rid of the old "get"

Bad idea, IMO...

>  (b) or just call it something better to begin with. Maybe just
> "get_regset" instead?

<nod>

Frankly, other field names there are also nasty - 'set' is not
fun to grep for, but there are 'n' and 'size' as well.  There's
also 'bias'  and 'align' (both completely unused)...

> I'd prefer (b) just because I think it will be a lot clearer if we
> ever end up having old patches forward-ported (or, more likely,
> newpatches back-ported), so having a "get" function that changed
> semantics but got back the old name sounds bad to me.
> 
> Other than that, I can't really argue with the numbers:
> 
>  41 files changed, 1368 insertions(+), 2347 deletions(-)
> 
> looks good to me, and the code seems more understandable.

FWIW, there's also ->set() side of that thing.  Tons of boilerplate
in those; I hadn't seriously looked into that part, but I suspect
that "do copyin in the caller, pass the kernel buffer to the
method, always start at offset 0" would also trim a lot of crap.
I'd rather leave that one for later - this series had been painful
enough.

Other things in the same area: conversion of fdpic coredumps to
regset (fairly easy, I've a patch series doing that in the local
tree), conversion of the few remaining non-regset architectures
to regset-based coredump (kill a large chunk of rotting code in
fs/binfmt_elf.c, along with quite a few pieces of old per-arch
coredump helpers), unification of compat ELF (done locally,
mipsn32/mipso32 gone, all compat done via compat_binfmt_elf.c,
x32 horrors sanitized - IMO
#define PRSTATUS_SIZE \
       (test_thread_flag(TIF_X32) \
               ? sizeof(struct compat_elf_prstatus) \
               : sizeof(struct i386_elf_prstatus))
#define SET_PR_FPVALID(S) \
       (*(test_thread_flag(TIF_X32) \
               ? &(S)->pr_fpvalid      \
               : &((struct i386_elf_prstatus *)(S))->pr_fpvalid) = 1)
is much better than
#define PRSTATUS_SIZE(S, R) (R != sizeof(S.pr_reg) ? 144 : 296)
#define SET_PR_FPVALID(S, V, R) \
  do { *(int *) (((void *) &((S)->pr_reg)) + R) = (V); } \
  while (0)
) and a bunch of assorted cleanups that got trimmed from the
regset series...

I'll probably post fdpic and compat series tonight and get
back to uaccess and VFS stuff.

Re: [PATCH 22/41] sparc: switch to ->get2()

[Al Viro]

On Mon, Jun 29, 2020 at 07:26:09PM +0100, Al Viro wrote:
  
>  static int getregs64_get(struct task_struct *target,
>  			 const struct user_regset *regset,
> -			 unsigned int pos, unsigned int count,
> -			 void *kbuf, void __user *ubuf)
> +			 struct membuf to)
>  {
>  	const struct pt_regs *regs = task_pt_regs(target);
> -	int ret;
>  
>  	if (target == current)
>  		flushw_user();
>  
> -	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> -				  regs->u_regs + 1,
> -				  0, 15 * sizeof(u64));
> -	if (!ret)
> -		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
> -				  15 * sizeof(u64), 16 * sizeof(u64));
> -	if (!ret) {
> -		/* TSTATE, TPC, TNPC */
> -		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> -					  &regs->tstate,
> -					  16 * sizeof(u64),
> -					  19 * sizeof(u64));
> -	}
> -	if (!ret) {
> -		unsigned long y = regs->y;
> -
> -		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> -					  &y,
> -					  19 * sizeof(u64),
> -					  20 * sizeof(u64));
> -	}
> -	return ret;
> +	membuf_write(&to, regs->u_regs + 1, 15 * sizeof(u64));
> +	return membuf_store(&to, (u64)0);
	^^^^^^
Er...  That should be simply

+	membuf_store(&to, (u64)0);

> +	membuf_write(&to, &regs->tstate, 3 * sizeof(u64));
> +	return membuf_store(&to, (u64)regs->y);
>  }

Re: [PATCH 18/41] regset: new method and helpers for it

[Al Viro]

On Mon, Jun 29, 2020 at 09:30:28PM +0100, Al Viro wrote:

> >  (b) or just call it something better to begin with. Maybe just
> > "get_regset" instead?
> 
> <nod>
> 
> Frankly, other field names there are also nasty - 'set' is not
> fun to grep for, but there are 'n' and 'size' as well.  There's
> also 'bias'  and 'align' (both completely unused)...

How about ->regset_get()?  I've force-pushed a variant with that change +
removal of leftover unused variables (caught by Intel build bot) +
braino fix in sparc conversion (spotted when looking through the
patchset again); same branch names.  Not sure if it's worth a repost
of the entire patchbomb...

Re: [PATCH 18/41] regset: new method and helpers for it

[Linus Torvalds]

On Tue, Jun 30, 2020 at 6:25 AM Al Viro <viro@zeniv.linux.org.uk> wrote:
>
> How about ->regset_get()?

Sounds good to me. And if you ever do something similar for 'set', you
have a natural name to pick.

No need to repost as far as I'm concerned if it's just the name change
and some minor nots found by the bots.

                     Linus

Re: [PATCH 18/41] regset: new method and helpers for it

[Al Viro]

On Tue, Jun 30, 2020 at 09:53:12AM -0700, Linus Torvalds wrote:
> On Tue, Jun 30, 2020 at 6:25 AM Al Viro <viro@zeniv.linux.org.uk> wrote:
> >
> > How about ->regset_get()?
> 
> Sounds good to me. And if you ever do something similar for 'set', you
> have a natural name to pick.

Umm...  Something similar for 'set' would, AFAICS, work reasonably well
with the following primitives.
membuf_read(&from, data, size) -- copy min(size, amount left)
membuf_fetch(&from, data) -- copy min(sizeof(data), amount left) into data
			(obviously a macro)
membuf_skip(&from, size) -- obvious
membuf_pick(&from, size), along the lines of
	if (unlikely(size > p->left)
		return ERR_PTR(-EINVAL);
	res = p->p;
	p->p += size;
	p->left -= size;
	return res;

So it would be something like

int regset_tls_set(struct task_struct *target, const struct user_regset *regset,
                   struct membuf from)
{
        const struct user_desc *info;
	int n = from.left / sizeof(*info);
        int i;

	info = membuf_pick(&from, n * sizeof(*info)); // can't fail here
        for (i = 0; i < n; i++)
                if (!tls_desc_okay(info + i))
                        return -EINVAL;

        set_tls_desc(target, GDT_ENTRY_TLS_MIN, info, n);
        return 0;
}

and

static int genregs32_set(struct task_struct *target,
                         const struct user_regset *regset,
                         struct membuf from)
{
	int err;
	unsigned pos, v;

	for (pos = 0; from.left; pos += sizeof(u32)) {
		membuf_fetch(&from, v);
		err = putreg32(target, pos, v);
		if (err)
			return err;
	}
	return 0;
}

or

/*
 * Copy the supplied 64-bit NT_MIPS_DSP buffer to the DSP context.
 */
static int dsp64_set(struct task_struct *target,
                     const struct user_regset *regset,
                     struct membuf from)
{
        if (!cpu_has_dsp)
                return -EIO;

	membuf_read(&from, target->thread.dsp.dspr, NUM_DSP_REGS * sizeof(u64));
	return membuf_read(&from, target->thread.dsp.dspcontrol, sizeof(u64));
}

etc.

Re: [RFC][PATCHSET] regset ->get() rework

[Christoph Hellwig]

Hi Al,

what's the state of the series?  I noticed it hasn't shown up in
linux-next so far.

Re: [RFC][PATCHSET] regset ->get() rework

[Al Viro]

On Thu, Jul 23, 2020 at 09:00:29AM +0100, Christoph Hellwig wrote:
> Hi Al,
> 
> what's the state of the series?  I noticed it hasn't shown up in
> linux-next so far.

It (as well as fdpic coredump stuff) will go into -next as soon as I post
the second variant of csum_and_copy stuff (later today).

RE: [PATCH 33/41] hexagon: switch to ->get2()

[Brian Cain]

> -----Original Message-----
> From: linux-arch-owner@vger.kernel.org <linux-arch-owner@vger.kernel.org>
> On Behalf Of Al Viro

Acked-by: Brian Cain <bcain@codeaurora.org>

> From: Al Viro <viro@zeniv.linux.org.uk>
> 
> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
> ---
>  arch/hexagon/kernel/ptrace.c | 62
+++++++++++++++-----------------------------
>  1 file changed, 21 insertions(+), 41 deletions(-)
> 
> diff --git a/arch/hexagon/kernel/ptrace.c b/arch/hexagon/kernel/ptrace.c
index
> dcbf7ea960cc..fa6287d1a061 100644
> --- a/arch/hexagon/kernel/ptrace.c
> +++ b/arch/hexagon/kernel/ptrace.c
> @@ -35,58 +35,38 @@ void user_disable_single_step(struct task_struct
*child)
> 
>  static int genregs_get(struct task_struct *target,
>  		   const struct user_regset *regset,
> -		   unsigned int pos, unsigned int count,
> -		   void *kbuf, void __user *ubuf)
> +		   srtuct membuf to)
>  {
> -	int ret;
> -	unsigned int dummy;
>  	struct pt_regs *regs = task_pt_regs(target);
> 
> -
> -	if (!regs)
> -		return -EIO;
> -
>  	/* The general idea here is that the copyout must happen in
>  	 * exactly the same order in which the userspace expects these
>  	 * regs. Now, the sequence in userspace does not match the
>  	 * sequence in the kernel, so everything past the 32 gprs
>  	 * happens one at a time.
>  	 */
> -	ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf,
> -				  &regs->r00, 0, 32*sizeof(unsigned long));
> -
> -#define ONEXT(KPT_REG, USR_REG) \
> -	if (!ret) \
> -		ret = user_regset_copyout(&pos, &count, &kbuf, &ubuf, \
> -			KPT_REG, offsetof(struct user_regs_struct, USR_REG),
> \
> -			offsetof(struct user_regs_struct, USR_REG) + \
> -				 sizeof(unsigned long));
> -
> +	membuf_write(&to, &regs->r00, 32*sizeof(unsigned long));
>  	/* Must be exactly same sequence as struct user_regs_struct */
> -	ONEXT(&regs->sa0, sa0);
> -	ONEXT(&regs->lc0, lc0);
> -	ONEXT(&regs->sa1, sa1);
> -	ONEXT(&regs->lc1, lc1);
> -	ONEXT(&regs->m0, m0);
> -	ONEXT(&regs->m1, m1);
> -	ONEXT(&regs->usr, usr);
> -	ONEXT(&regs->preds, p3_0);
> -	ONEXT(&regs->gp, gp);
> -	ONEXT(&regs->ugp, ugp);
> -	ONEXT(&pt_elr(regs), pc);
> -	dummy = pt_cause(regs);
> -	ONEXT(&dummy, cause);
> -	ONEXT(&pt_badva(regs), badva);
> +	membuf_store(&to, regs->sa0);
> +	membuf_store(&to, regs->lc0);
> +	membuf_store(&to, regs->sa1);
> +	membuf_store(&to, regs->lc1);
> +	membuf_store(&to, regs->m0);
> +	membuf_store(&to, regs->m1);
> +	membuf_store(&to, regs->usr);
> +	membuf_store(&to, regs->p3_0);
> +	membuf_store(&to, regs->gp);
> +	membuf_store(&to, regs->ugp);
> +	membuf_store(&to, pt_elr(regs)); // pc
> +	membuf_store(&to, (unsigned long)pt_cause(regs)); // cause
> +	membuf_store(&to, pt_badva(regs)); // badva
>  #if CONFIG_HEXAGON_ARCH_VERSION >=4
> -	ONEXT(&regs->cs0, cs0);
> -	ONEXT(&regs->cs1, cs1);
> +	membuf_store(&to, regs->cs0);
> +	membuf_store(&to, regs->cs1);
> +	return membuf_zero(&to, sizeof(unsigned long)); #else
> +	return membuf_zero(&to, 3 * sizeof(unsigned long));
>  #endif
> -
> -	/* Pad the rest with zeros, if needed */
> -	if (!ret)
> -		ret = user_regset_copyout_zero(&pos, &count, &kbuf, &ubuf,
> -					offsetof(struct user_regs_struct,
> pad1), -1);
> -	return ret;
>  }
> 
>  static int genregs_set(struct task_struct *target, @@ -159,7 +139,7 @@
static
> const struct user_regset hexagon_regsets[] = {
>  		.n = ELF_NGREG,
>  		.size = sizeof(unsigned long),
>  		.align = sizeof(unsigned long),
> -		.get = genregs_get,
> +		.get2 = genregs_get,
>  		.set = genregs_set,
>  	},
>  };
> --
> 2.11.0