From: Tom Hughes <tom@compton.nu>
Date: Mon, 14 Nov 2005 14:37:02 +0000 (+0000)
Subject: Document sign extension of si_code by the kernel.
X-Git-Tag: svn/VALGRIND_3_1_0~109
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=204dd5138c34c87a0d8bb4629d12dc65ece81458;p=thirdparty%2Fvalgrind.git

Document sign extension of si_code by the kernel.


git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5123
---

diff --git a/coregrind/m_signals.c b/coregrind/m_signals.c
index f11269bee4..438262dde7 100644
--- a/coregrind/m_signals.c
+++ b/coregrind/m_signals.c
@@ -1358,7 +1358,11 @@ void async_signalhandler ( Int sigNo, vki_siginfo_t *info, struct vki_ucontext *
    /* The linux kernel uses the top 16 bits of si_code for it's own
       use and only exports the bottom 16 bits to user space - at least
       that is the theory, but it turns out that there are some kernels
-      around that forget to mask out the top 16 bits so we do it here. */
+      around that forget to mask out the top 16 bits so we do it here.
+
+      The kernel treats the bottom 16 bits as signed and (when it does
+      mask them off) sign extends them when exporting to user space so
+      we do the same thing here. */
    info->si_code = (Short)info->si_code;
 #endif
 
@@ -1475,7 +1479,11 @@ void sync_signalhandler ( Int sigNo, vki_siginfo_t *info, struct vki_ucontext *u
    /* The linux kernel uses the top 16 bits of si_code for it's own
       use and only exports the bottom 16 bits to user space - at least
       that is the theory, but it turns out that there are some kernels
-      around that forget to mask out the top 16 bits so we do it here. */
+      around that forget to mask out the top 16 bits so we do it here.
+
+      The kernel treats the bottom 16 bits as signed and (when it does
+      mask them off) sign extends them when exporting to user space so
+      we do the same thing here. */
    info->si_code = (Short)info->si_code;
 #endif