qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Richard Henderson <rth@twiddle.net>
To: qemu-devel@nongnu.org
Cc: Blue Swirl <blauwirbel@gmail.com>, Riku Voipio <riku.voipio@iki.fi>
Subject: [Qemu-devel] [PATCH 6/8] linux-user: Rewrite __get_user/__put_user with __builtin_choose_expr
Date: Fri,  4 Jan 2013 16:39:31 -0800	[thread overview]
Message-ID: <1357346373-13898-7-git-send-email-rth@twiddle.net> (raw)
In-Reply-To: <1357346373-13898-1-git-send-email-rth@twiddle.net>

The previous formuation with multiple assignments to __typeof(*hptr) falls
down when hptr is qualified const.  E.g. with const struct S *p, p->f is
also qualified const.

With this formulation, there's no assignment to any local variable.

Signed-off-by: Richard Henderson <rth@twiddle.net>
---
 linux-user/qemu.h | 63 +++++++++++++++++++++++++++++--------------------------
 1 file changed, 33 insertions(+), 30 deletions(-)

diff --git a/linux-user/qemu.h b/linux-user/qemu.h
index 8a3538c..31a220a 100644
--- a/linux-user/qemu.h
+++ b/linux-user/qemu.h
@@ -287,36 +287,39 @@ static inline int access_ok(int type, abi_ulong addr, abi_ulong size)
                             (type == VERIFY_READ) ? PAGE_READ : (PAGE_READ | PAGE_WRITE)) == 0;
 }
 
-/* NOTE __get_user and __put_user use host pointers and don't check access. */
-/* These are usually used to access struct data members once the
- * struct has been locked - usually with lock_user_struct().
- */
-#define __put_user(x, hptr)\
-({ __typeof(*hptr) pu_ = (x);\
-    switch(sizeof(*hptr)) {\
-    case 1: break;\
-    case 2: pu_ = tswap16(pu_); break; \
-    case 4: pu_ = tswap32(pu_); break; \
-    case 8: pu_ = tswap64(pu_); break; \
-    default: abort();\
-    }\
-    memcpy(hptr, &pu_, sizeof(pu_)); \
-    0;\
-})
-
-#define __get_user(x, hptr) \
-({ __typeof(*hptr) gu_; \
-    memcpy(&gu_, hptr, sizeof(gu_)); \
-    switch(sizeof(*hptr)) {\
-    case 1: break; \
-    case 2: gu_ = tswap16(gu_); break; \
-    case 4: gu_ = tswap32(gu_); break; \
-    case 8: gu_ = tswap64(gu_); break; \
-    default: abort();\
-    }\
-    (x) = gu_; \
-    0;\
-})
+/* NOTE __get_user and __put_user use host pointers and don't check access.
+   These are usually used to access struct data members once the struct has
+   been locked - usually with lock_user_struct.  */
+
+/* Tricky points:
+   - Use __builtin_choose_expr to avoid type promotion from ?:,
+   - Invalid sizes result in a compile time error stemming from
+     the fact that abort has no parameters.
+   - It's easier to use the endian-specific unaligned load/store
+     functions than host-endian unaligned load/store plus tswapN.  */
+
+#define __put_user_e(x, hptr, e)                                        \
+  (__builtin_choose_expr(sizeof(*(hptr)) == 1, stb_p,                   \
+   __builtin_choose_expr(sizeof(*(hptr)) == 2, stw_##e##_p,             \
+   __builtin_choose_expr(sizeof(*(hptr)) == 4, stl_##e##_p,             \
+   __builtin_choose_expr(sizeof(*(hptr)) == 8, stq_##e##_p, abort))))   \
+     ((hptr), (x)), 0)
+
+#define __get_user_e(x, hptr, e)                                        \
+  ((x) =                                                                \
+   __builtin_choose_expr(sizeof(*(hptr)) == 1, ldub_p,                  \
+   __builtin_choose_expr(sizeof(*(hptr)) == 2, lduw_##e##_p,            \
+   __builtin_choose_expr(sizeof(*(hptr)) == 4, ldl_##e##_p,             \
+   __builtin_choose_expr(sizeof(*(hptr)) == 8, ldq_##e##_p, abort))))   \
+     (hptr), 0)
+
+#ifdef TARGET_WORDS_BIGENDIAN
+# define __put_user(x, hptr)  __put_user_e(x, hptr, be)
+# define __get_user(x, hptr)  __get_user_e(x, hptr, be)
+#else
+# define __put_user(x, hptr)  __put_user_e(x, hptr, le)
+# define __get_user(x, hptr)  __get_user_e(x, hptr, le)
+#endif
 
 /* put_user()/get_user() take a guest address and check access */
 /* These are usually used to access an atomic data type, such as an int,
-- 
1.7.11.7

  parent reply	other threads:[~2013-01-05  0:39 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-01-05  0:39 [Qemu-devel] [PATCH v2 0/8] linux-user fixes Richard Henderson
2013-01-05  0:39 ` [Qemu-devel] [PATCH 1/8] fdt: Use bswapN instead of bswap_N Richard Henderson
2013-01-05  0:39 ` [Qemu-devel] [PATCH 2/8] bswap: Tidy base definitions of bswapN Richard Henderson
2013-01-05  0:39 ` [Qemu-devel] [PATCH 3/8] bswap: Add host endian unaligned access functions Richard Henderson
2013-01-05  0:39 ` [Qemu-devel] [PATCH 4/8] bswap: Rewrite all ld<type>_<endian>_p functions Richard Henderson
2013-01-05  0:39 ` [Qemu-devel] [PATCH 5/8] bswap: Rewrite cpu_to_<endian><type>u with {ld, st}<type>_<endian>_p Richard Henderson
2013-01-05  0:39 ` Richard Henderson [this message]
2013-01-05  0:39 ` [Qemu-devel] [PATCH 7/8] alpha-linux-user: Fix sigaction Richard Henderson
2013-01-05  0:39 ` [Qemu-devel] [PATCH 8/8] user: Consider symbolic links as possible directories Richard Henderson
2013-01-12 16:07 ` [Qemu-devel] [PATCH v2 0/8] linux-user fixes Blue Swirl

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1357346373-13898-7-git-send-email-rth@twiddle.net \
    --to=rth@twiddle.net \
    --cc=blauwirbel@gmail.com \
    --cc=qemu-devel@nongnu.org \
    --cc=riku.voipio@iki.fi \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).