Function displaying console buffer always assumes its content lenght
equal to console buffer size. This is not true and sometimes it sends
garbage to the screen. This patch fixes this issue.

Additionally, remove references to unused conringc variable.

Signed-off-by: Daniel Kiper <[email protected]>

diff -Npru crash-6.1.4.orig/xen_hyper_command.c crash-6.1.4/xen_hyper_command.c
--- crash-6.1.4.orig/xen_hyper_command.c        2013-02-14 21:38:54.000000000 
+0100
+++ crash-6.1.4/xen_hyper_command.c     2013-03-14 15:03:58.045075917 +0100
@@ -585,11 +585,10 @@ xen_hyper_cmd_log(void)
 void
 xen_hyper_dump_log(void)
 {
-       uint conringc, conringp;
-       uint warp, start, len, idx, i;
+       uint conringp, warp, len, idx, i;
        ulong conring;
        char *buf;
-       char last;
+       char last = 0;
        uint32_t conring_size;
 
        if (get_symbol_type("conring", NULL, NULL) == TYPE_CODE_ARRAY)
@@ -597,7 +596,6 @@ xen_hyper_dump_log(void)
        else
                get_symbol_data("conring", sizeof(ulong), &conring);
 
-       get_symbol_data("conringc", sizeof(uint), &conringc);
        get_symbol_data("conringp", sizeof(uint), &conringp);
 
        if (symbol_exists("conring_size"))
@@ -605,21 +603,19 @@ xen_hyper_dump_log(void)
        else
                conring_size = XEN_HYPER_CONRING_SIZE;
 
-       warp = FALSE;
        if (conringp >= conring_size) {
-               if ((start = conringp & (conring_size - 1))) {
-                       warp = TRUE;
-               }
+               idx = conringp & (conring_size - 1);
+               len = conring_size;
+               warp = TRUE;
        } else {
-               start = 0;
+               idx = 0;
+               len = conringp;
+               warp = FALSE;
        }
 
        buf = GETBUF(conring_size);
        readmem(conring, KVADDR, buf, conring_size,
                "conring contents", FAULT_ON_ERROR);
-       idx = start;
-       len = conring_size;
-       last = 0;
 
 wrap_around:
        for (i = idx; i < len; i++) {

--
Crash-utility mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/crash-utility

Reply via email to