Mainline Android logger project - Revision history

Frowand: /* [8] ttyprintk driver */

2012-03-26T19:38:47Z

‎[8] ttyprintk driver

Frowand: /* [8] ttyprintk driver */

2012-03-26T19:38:04Z

‎[8] ttyprintk driver

Tim Bird: /* [8] ttyprintk driver */ add sample code to cause console to go to ttyprintk

2012-03-26T18:55:03Z

‎[8] ttyprintk driver: add sample code to cause console to go to ttyprintk

Tim Bird: /* [8] printk tty driver */

2012-03-24T19:52:40Z

‎[8] printk tty driver

Tim Bird: /* [8] printk tty driver */

2012-03-24T19:10:08Z

‎[8] printk tty driver

Tim Bird: /* [8] printk tty driver */

2012-03-24T19:06:10Z

‎[8] printk tty driver

Tim Bird: /* [8] printk tty driver */

2012-03-24T18:24:18Z

‎[8] printk tty driver

Tim Bird: /* [8] printk tty driver */

2012-03-24T17:46:01Z

‎[8] printk tty driver

Tim Bird: /* [13] code cleanups for current driver */

2012-03-21T01:10:07Z

‎[13] code cleanups for current driver

Tim Bird: /* [10] Kay Siever's wishlist */

2012-02-07T19:15:58Z

‎[10] Kay Siever's wishlist

@@ Line 291: / Line 291: @@
 ** See http://www.kerneltrap.org/mailarchive/linux-kernel/2010/8/25/4611115 for Alan Cox's justification for this, compared to systemd on Fedora 14, or Kay Seiver's 'exec </dev/console >/dev/kmsg 2>&1'
-*** '''CAUTION: 'exec </dev/console >/dev/kmsg 2>&1' crashes the Xserver on my Fedora 14 system'''
+*** '''CAUTION: 'exec </dev/console >/dev/kmsg 2>&1' crashes the Xserver on my Fedora 14 system (frowand)'''
 *** answering the question - what's different about this from /dev/kmsg
 *** also: http://www.kerneltrap.org/mailarchive/linux-kernel/2010/8/25/4611254

@@ Line 290: / Line 290: @@
 *** I'm not sure how this is installed. (maybe with console=ttyprintk on the command line?)
-** See http://www.kerneltrap.org/mailarchive/linux-kernel/2010/8/25/4611115 for Alan Cox's justification for this, compared to systemd on Fedora 14, or Kay Seiver's 'exec <dev/console >/dev/kmsg 2>&1'
+** See http://www.kerneltrap.org/mailarchive/linux-kernel/2010/8/25/4611115 for Alan Cox's justification for this, compared to systemd on Fedora 14, or Kay Seiver's 'exec </dev/console >/dev/kmsg 2>&1'
 *** answering the question - what's different about this from /dev/kmsg
 *** also: http://www.kerneltrap.org/mailarchive/linux-kernel/2010/8/25/4611254

@@ Line 300: / Line 300: @@
   daemons etc, all of which work to produce that feature IFF you have pty
   support in your kernel, and for the current code also glibc.
 ==== [9] no need to switch to something new ====

@@ Line 275: / Line 275: @@
 ** I have no numbers, but suspect there is no problem
-==== [8] printk tty driver ====
+==== [8] ttyprintk driver ====
-* [8] userspace printk tty driver (Greg KH, Kay Sievers mentioned it)
+* [8] userspace printk tty (ttyprintk) driver (Greg KH, Kay Sievers mentioned it)
 ** I never heard of it - need to do some research I guess
 ** see linux/drivers/char/ttyprintk.c
@@ Line 293: / Line 293: @@
 *** answering the question - what's different about this from /dev/kmsg
 *** also: http://www.kerneltrap.org/mailarchive/linux-kernel/2010/8/25/4611254
 ==== [9] no need to switch to something new ====

@@ Line 292: / Line 292: @@
 ** See http://www.kerneltrap.org/mailarchive/linux-kernel/2010/8/25/4611115 for Alan Cox's justification for this, compared to systemd on Fedora 14, or Kay Seiver's 'exec <dev/console >/dev/kmsg 2>&1'
 *** answering the question - what's different about this from /dev/kmsg
 ==== [9] no need to switch to something new ====

@@ Line 279: / Line 279: @@
 ** I never heard of it - need to do some research I guess
 ** see linux/drivers/char/ttyprintk.c
 ** Results: ttyprintk allows user-space to submit messages to the kernel log buffer
 *** ttyprintk features:
@@ Line 285: / Line 286: @@
 *** splits long messages into fragments no longer than 508 bytes
 ** (How is this different from /dev/kmsg???)
 ==== [9] no need to switch to something new ====

@@ Line 316: / Line 316: @@
 ==== [13] code cleanups for current driver ====
 * [13] Specific code fixes in current driver (Andrew Morton, see https://lkml.org/lkml/2011/12/28/165)
-** [13.1] logger_offset macro depends on local 'log' variable
+** [13.1] ** MAINLINED in 3.4 - logger_offset macro depends on local 'log' variable
-** [13.2] should document endian-ness of numeric values in buffer
+** [13.2] ** MAINLINED in 3.4 - should document endian-ness of numeric values in buffer
-** [13.3] do get_entry_len more neatly (possibly using packed_struct.h)
+** [13.3] ** MAINLINED in 3.4 - do get_entry_len more neatly (possibly using packed_struct.h)
-** [13.4] use get_unaligned() instead of current get_entry_len() code??
+** [13.4] ** FIXED in 3.4 - use get_unaligned() instead of current get_entry_len() code??
-** [13.5] change order of prepare_to_wait and mutex_lock lines in logger_read()
+** [13.5] ** MAINLINED in 3.4 - change order of prepare_to_wait and mutex_lock lines in logger_read()
-** [13.6] make clock_interval function more obvious
+** [13.6] ** MAINLINED in 3.4 - make clock_interval function more obvious
 ** [13.7] don't make clock_interval explicitly inline (gcc should figure it out)
-** [13.8] in do_write_log_from_user, update log->w_off ??
+** [13.8] ** MAINLINED in 3.4 - in do_write_log_from_user, update log->w_off ??
 ** [13.9] resolve pid, tgid confusion
 ** [13.10] consider access control for logs, especially in container context
-** [13.11] does reader->list need to be locked, in logger_release?
+** [13.11] ** MAINLINED in 3.4 - does reader->list need to be locked, in logger_release?
 ** [13.12] get rid of ioctl?
 ** [13.13] get rid of hardcoded minor names in get_log_from_minor()

@@ Line 294: / Line 294: @@
 ==== [10] Kay Siever's wishlist ====
 * [10] Kay Siever's wishlist - kmsg with structured data (Kay Sievers - who else?)
-** supports structured data, with can use with current tools
+** supports structured data, that can be used with current tools
 ** each record has timestamp, log facility, log level, string (no pid)
 ** timestamp converted on output (not saved as text in log)