[RESEND PATCH] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[RESEND PATCH] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Lubomir Rintel]

Symbol name changed with the following commit:
62e32ac printk: rename struct log to struct printk_log

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Cc: Simon Horman <horms@verge.net.au>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
---

I believe this can now be applied, as the change went into 3.11 kernel release.
Thank you!

 vmcore-dmesg/vmcore-dmesg.c |   10 +++++++++-
 1 files changed, 9 insertions(+), 1 deletions(-)

diff --git a/vmcore-dmesg/vmcore-dmesg.c b/vmcore-dmesg/vmcore-dmesg.c
index 0f477c0..bfc19e9 100644
--- a/vmcore-dmesg/vmcore-dmesg.c
+++ b/vmcore-dmesg/vmcore-dmesg.c
@@ -324,19 +324,27 @@ static void scan_vmcoreinfo(char *start, size_t size)
 			*symbol[i].vaddr = vaddr;
 		}
 
-		/* Check for "SIZE(log)=" */
+		/* Check for "SIZE(printk_log)" or older "SIZE(log)=" */
 		if (memcmp("SIZE(log)=", pos, 10) == 0)
 			log_sz = strtoull(pos + 10, NULL, 10);
+		if (memcmp("SIZE(printk_log)=", pos, 17) == 0)
+			log_sz = strtoull(pos + 17, NULL, 10);
 
 		/* Check for struct log field offsets */
 		if (memcmp("OFFSET(log.ts_nsec)=", pos, 20) == 0)
 			log_offset_ts_nsec = strtoull(pos + 20, NULL, 10);
+		if (memcmp("OFFSET(printk_log.ts_nsec)=", pos, 27) == 0)
+			log_offset_ts_nsec = strtoull(pos + 27, NULL, 10);
 
 		if (memcmp("OFFSET(log.len)=", pos, 16) == 0)
 			log_offset_len = strtoul(pos + 16, NULL, 10);
+		if (memcmp("OFFSET(printk_log.len)=", pos, 23) == 0)
+			log_offset_len = strtoul(pos + 23, NULL, 10);
 
 		if (memcmp("OFFSET(log.text_len)=", pos, 21) == 0)
 			log_offset_text_len = strtoul(pos + 21, NULL, 10);
+		if (memcmp("OFFSET(printk_log.text_len)=", pos, 28) == 0)
+			log_offset_text_len = strtoul(pos + 28, NULL, 10);
 
 		if (last_line)
 			break;
-- 
1.7.1


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [RESEND PATCH] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Dave Young]

On 09/18/13 at 02:21pm, Lubomir Rintel wrote:
> Symbol name changed with the following commit:
> 62e32ac printk: rename struct log to struct printk_log
> 
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> Cc: Simon Horman <horms@verge.net.au>
> Cc: Vivek Goyal <vgoyal@redhat.com>
> Cc: Joe Perches <joe@perches.com>
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> ---
> 
> I believe this can now be applied, as the change went into 3.11 kernel release.
> Thank you!

It works for me, thanks.

But there's still a minor issue, can you also update other
comments besides of SIZEOF(log)? I think it's better to mention that
log has been renamed printk_log since 3.11 kernel.

> 
>  vmcore-dmesg/vmcore-dmesg.c |   10 +++++++++-
>  1 files changed, 9 insertions(+), 1 deletions(-)
> 
> diff --git a/vmcore-dmesg/vmcore-dmesg.c b/vmcore-dmesg/vmcore-dmesg.c
> index 0f477c0..bfc19e9 100644
> --- a/vmcore-dmesg/vmcore-dmesg.c
> +++ b/vmcore-dmesg/vmcore-dmesg.c
> @@ -324,19 +324,27 @@ static void scan_vmcoreinfo(char *start, size_t size)
>  			*symbol[i].vaddr = vaddr;
>  		}
>  
> -		/* Check for "SIZE(log)=" */
> +		/* Check for "SIZE(printk_log)" or older "SIZE(log)=" */
>  		if (memcmp("SIZE(log)=", pos, 10) == 0)
>  			log_sz = strtoull(pos + 10, NULL, 10);
> +		if (memcmp("SIZE(printk_log)=", pos, 17) == 0)
> +			log_sz = strtoull(pos + 17, NULL, 10);
>  
>  		/* Check for struct log field offsets */
>  		if (memcmp("OFFSET(log.ts_nsec)=", pos, 20) == 0)
>  			log_offset_ts_nsec = strtoull(pos + 20, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.ts_nsec)=", pos, 27) == 0)
> +			log_offset_ts_nsec = strtoull(pos + 27, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.len)=", pos, 16) == 0)
>  			log_offset_len = strtoul(pos + 16, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.len)=", pos, 23) == 0)
> +			log_offset_len = strtoul(pos + 23, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.text_len)=", pos, 21) == 0)
>  			log_offset_text_len = strtoul(pos + 21, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.text_len)=", pos, 28) == 0)
> +			log_offset_text_len = strtoul(pos + 28, NULL, 10);
>  
>  		if (last_line)
>  			break;
> -- 
> 1.7.1
> 
> 
> _______________________________________________
> kexec mailing list
> kexec@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kexec

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

[PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Lubomir Rintel]

Symbol name changed with the following commit:
62e32ac printk: rename struct log to struct printk_log

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Cc: Simon Horman <horms@verge.net.au>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Joe Perches <joe@perches.com>
Cc: Dave Young <dyoung@redhat.com>
---

Changes for v2:
  * Clarified printk_log vs. log in some comments.

 vmcore-dmesg/vmcore-dmesg.c |   16 ++++++++++++----
 1 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/vmcore-dmesg/vmcore-dmesg.c b/vmcore-dmesg/vmcore-dmesg.c
index 0f477c0..dd222cd 100644
--- a/vmcore-dmesg/vmcore-dmesg.c
+++ b/vmcore-dmesg/vmcore-dmesg.c
@@ -34,10 +34,10 @@ static loff_t logged_chars_vaddr;
 static loff_t log_first_idx_vaddr;
 static loff_t log_next_idx_vaddr;
 
-/* struct log size */
+/* struct printk_log (or older log) size */
 static uint64_t log_sz;
 
-/* struct log field offsets */
+/* struct printk_log (or older log) field offsets */
 static uint64_t log_offset_ts_nsec = UINT64_MAX;
 static uint16_t log_offset_len = UINT16_MAX;
 static uint16_t log_offset_text_len = UINT16_MAX;
@@ -324,19 +324,27 @@ static void scan_vmcoreinfo(char *start, size_t size)
 			*symbol[i].vaddr = vaddr;
 		}
 
-		/* Check for "SIZE(log)=" */
+		/* Check for "SIZE(printk_log)" or older "SIZE(log)=" */
 		if (memcmp("SIZE(log)=", pos, 10) == 0)
 			log_sz = strtoull(pos + 10, NULL, 10);
+		if (memcmp("SIZE(printk_log)=", pos, 17) == 0)
+			log_sz = strtoull(pos + 17, NULL, 10);
 
-		/* Check for struct log field offsets */
+		/* Check for struct printk_log (or older log) field offsets */
 		if (memcmp("OFFSET(log.ts_nsec)=", pos, 20) == 0)
 			log_offset_ts_nsec = strtoull(pos + 20, NULL, 10);
+		if (memcmp("OFFSET(printk_log.ts_nsec)=", pos, 27) == 0)
+			log_offset_ts_nsec = strtoull(pos + 27, NULL, 10);
 
 		if (memcmp("OFFSET(log.len)=", pos, 16) == 0)
 			log_offset_len = strtoul(pos + 16, NULL, 10);
+		if (memcmp("OFFSET(printk_log.len)=", pos, 23) == 0)
+			log_offset_len = strtoul(pos + 23, NULL, 10);
 
 		if (memcmp("OFFSET(log.text_len)=", pos, 21) == 0)
 			log_offset_text_len = strtoul(pos + 21, NULL, 10);
+		if (memcmp("OFFSET(printk_log.text_len)=", pos, 28) == 0)
+			log_offset_text_len = strtoul(pos + 28, NULL, 10);
 
 		if (last_line)
 			break;
-- 
1.7.1


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Dave Young]

On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> Symbol name changed with the following commit:
> 62e32ac printk: rename struct log to struct printk_log
> 
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> Cc: Simon Horman <horms@verge.net.au>
> Cc: Vivek Goyal <vgoyal@redhat.com>
> Cc: Joe Perches <joe@perches.com>
> Cc: Dave Young <dyoung@redhat.com>
> ---

Acked-by: Dave Young <dyoung@redhat.com>


> 
> Changes for v2:
>   * Clarified printk_log vs. log in some comments.
> 
>  vmcore-dmesg/vmcore-dmesg.c |   16 ++++++++++++----
>  1 files changed, 12 insertions(+), 4 deletions(-)
> 
> diff --git a/vmcore-dmesg/vmcore-dmesg.c b/vmcore-dmesg/vmcore-dmesg.c
> index 0f477c0..dd222cd 100644
> --- a/vmcore-dmesg/vmcore-dmesg.c
> +++ b/vmcore-dmesg/vmcore-dmesg.c
> @@ -34,10 +34,10 @@ static loff_t logged_chars_vaddr;
>  static loff_t log_first_idx_vaddr;
>  static loff_t log_next_idx_vaddr;
>  
> -/* struct log size */
> +/* struct printk_log (or older log) size */
>  static uint64_t log_sz;
>  
> -/* struct log field offsets */
> +/* struct printk_log (or older log) field offsets */
>  static uint64_t log_offset_ts_nsec = UINT64_MAX;
>  static uint16_t log_offset_len = UINT16_MAX;
>  static uint16_t log_offset_text_len = UINT16_MAX;
> @@ -324,19 +324,27 @@ static void scan_vmcoreinfo(char *start, size_t size)
>  			*symbol[i].vaddr = vaddr;
>  		}
>  
> -		/* Check for "SIZE(log)=" */
> +		/* Check for "SIZE(printk_log)" or older "SIZE(log)=" */
>  		if (memcmp("SIZE(log)=", pos, 10) == 0)
>  			log_sz = strtoull(pos + 10, NULL, 10);
> +		if (memcmp("SIZE(printk_log)=", pos, 17) == 0)
> +			log_sz = strtoull(pos + 17, NULL, 10);
>  
> -		/* Check for struct log field offsets */
> +		/* Check for struct printk_log (or older log) field offsets */
>  		if (memcmp("OFFSET(log.ts_nsec)=", pos, 20) == 0)
>  			log_offset_ts_nsec = strtoull(pos + 20, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.ts_nsec)=", pos, 27) == 0)
> +			log_offset_ts_nsec = strtoull(pos + 27, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.len)=", pos, 16) == 0)
>  			log_offset_len = strtoul(pos + 16, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.len)=", pos, 23) == 0)
> +			log_offset_len = strtoul(pos + 23, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.text_len)=", pos, 21) == 0)
>  			log_offset_text_len = strtoul(pos + 21, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.text_len)=", pos, 28) == 0)
> +			log_offset_text_len = strtoul(pos + 28, NULL, 10);
>  
>  		if (last_line)
>  			break;
> -- 
> 1.7.1
> 

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Joe Perches]

On Fri, 2013-09-20 at 09:50 +0800, Dave Young wrote:
> On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> > Symbol name changed with the following commit:
> > 62e32ac printk: rename struct log to struct printk_log
> > 
> > Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> > Cc: Simon Horman <horms@verge.net.au>
> > Cc: Vivek Goyal <vgoyal@redhat.com>
> > Cc: Joe Perches <joe@perches.com>
> > Cc: Dave Young <dyoung@redhat.com>
> > ---
> 
> Acked-by: Dave Young <dyoung@redhat.com>

It might be nice if there was a reference somewhere in
the linux kernel source tree to whatever it is that this
patch modifies.



_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Dave Young]

On 09/19/13 at 06:55pm, Joe Perches wrote:
> On Fri, 2013-09-20 at 09:50 +0800, Dave Young wrote:
> > On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> > > Symbol name changed with the following commit:
> > > 62e32ac printk: rename struct log to struct printk_log
> > > 
> > > Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> > > Cc: Simon Horman <horms@verge.net.au>
> > > Cc: Vivek Goyal <vgoyal@redhat.com>
> > > Cc: Joe Perches <joe@perches.com>
> > > Cc: Dave Young <dyoung@redhat.com>
> > > ---
> > 
> > Acked-by: Dave Young <dyoung@redhat.com>
> 
> It might be nice if there was a reference somewhere in
> the linux kernel source tree to whatever it is that this
> patch modifies.
> 

Agree with you, it will be better to add the info in comment. 

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Vivek Goyal]

On Thu, Sep 19, 2013 at 04:49:46PM +0200, Lubomir Rintel wrote:
> Symbol name changed with the following commit:
> 62e32ac printk: rename struct log to struct printk_log
> 
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> Cc: Simon Horman <horms@verge.net.au>
> Cc: Vivek Goyal <vgoyal@redhat.com>
> Cc: Joe Perches <joe@perches.com>
> Cc: Dave Young <dyoung@redhat.com>
> ---

Looks good to me. Agreed that let us put kernel verion number in
changelog since this change became effective.

Acked-by: Vivek Goyal <vgoyal@redhat.com>

Vivek

> 
> Changes for v2:
>   * Clarified printk_log vs. log in some comments.
> 
>  vmcore-dmesg/vmcore-dmesg.c |   16 ++++++++++++----
>  1 files changed, 12 insertions(+), 4 deletions(-)
> 
> diff --git a/vmcore-dmesg/vmcore-dmesg.c b/vmcore-dmesg/vmcore-dmesg.c
> index 0f477c0..dd222cd 100644
> --- a/vmcore-dmesg/vmcore-dmesg.c
> +++ b/vmcore-dmesg/vmcore-dmesg.c
> @@ -34,10 +34,10 @@ static loff_t logged_chars_vaddr;
>  static loff_t log_first_idx_vaddr;
>  static loff_t log_next_idx_vaddr;
>  
> -/* struct log size */
> +/* struct printk_log (or older log) size */
>  static uint64_t log_sz;
>  
> -/* struct log field offsets */
> +/* struct printk_log (or older log) field offsets */
>  static uint64_t log_offset_ts_nsec = UINT64_MAX;
>  static uint16_t log_offset_len = UINT16_MAX;
>  static uint16_t log_offset_text_len = UINT16_MAX;
> @@ -324,19 +324,27 @@ static void scan_vmcoreinfo(char *start, size_t size)
>  			*symbol[i].vaddr = vaddr;
>  		}
>  
> -		/* Check for "SIZE(log)=" */
> +		/* Check for "SIZE(printk_log)" or older "SIZE(log)=" */
>  		if (memcmp("SIZE(log)=", pos, 10) == 0)
>  			log_sz = strtoull(pos + 10, NULL, 10);
> +		if (memcmp("SIZE(printk_log)=", pos, 17) == 0)
> +			log_sz = strtoull(pos + 17, NULL, 10);
>  
> -		/* Check for struct log field offsets */
> +		/* Check for struct printk_log (or older log) field offsets */
>  		if (memcmp("OFFSET(log.ts_nsec)=", pos, 20) == 0)
>  			log_offset_ts_nsec = strtoull(pos + 20, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.ts_nsec)=", pos, 27) == 0)
> +			log_offset_ts_nsec = strtoull(pos + 27, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.len)=", pos, 16) == 0)
>  			log_offset_len = strtoul(pos + 16, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.len)=", pos, 23) == 0)
> +			log_offset_len = strtoul(pos + 23, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.text_len)=", pos, 21) == 0)
>  			log_offset_text_len = strtoul(pos + 21, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.text_len)=", pos, 28) == 0)
> +			log_offset_text_len = strtoul(pos + 28, NULL, 10);
>  
>  		if (last_line)
>  			break;
> -- 
> 1.7.1

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[WANG Chao]

On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> Symbol name changed with the following commit:
> 62e32ac printk: rename struct log to struct printk_log
> 
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> Cc: Simon Horman <horms@verge.net.au>
> Cc: Vivek Goyal <vgoyal@redhat.com>
> Cc: Joe Perches <joe@perches.com>
> Cc: Dave Young <dyoung@redhat.com>
> ---

Hi,

vmcore-dmesg is really broken here since 3.11 kernel is released. Are
you going to send an update v3 to address the comments to v2 from Vivek,
Joe and Dave

Thanks
WANG Chao

> 
> Changes for v2:
>   * Clarified printk_log vs. log in some comments.
> 
>  vmcore-dmesg/vmcore-dmesg.c |   16 ++++++++++++----
>  1 files changed, 12 insertions(+), 4 deletions(-)
> 
> diff --git a/vmcore-dmesg/vmcore-dmesg.c b/vmcore-dmesg/vmcore-dmesg.c
> index 0f477c0..dd222cd 100644
> --- a/vmcore-dmesg/vmcore-dmesg.c
> +++ b/vmcore-dmesg/vmcore-dmesg.c
> @@ -34,10 +34,10 @@ static loff_t logged_chars_vaddr;
>  static loff_t log_first_idx_vaddr;
>  static loff_t log_next_idx_vaddr;
>  
> -/* struct log size */
> +/* struct printk_log (or older log) size */
>  static uint64_t log_sz;
>  
> -/* struct log field offsets */
> +/* struct printk_log (or older log) field offsets */
>  static uint64_t log_offset_ts_nsec = UINT64_MAX;
>  static uint16_t log_offset_len = UINT16_MAX;
>  static uint16_t log_offset_text_len = UINT16_MAX;
> @@ -324,19 +324,27 @@ static void scan_vmcoreinfo(char *start, size_t size)
>  			*symbol[i].vaddr = vaddr;
>  		}
>  
> -		/* Check for "SIZE(log)=" */
> +		/* Check for "SIZE(printk_log)" or older "SIZE(log)=" */
>  		if (memcmp("SIZE(log)=", pos, 10) == 0)
>  			log_sz = strtoull(pos + 10, NULL, 10);
> +		if (memcmp("SIZE(printk_log)=", pos, 17) == 0)
> +			log_sz = strtoull(pos + 17, NULL, 10);
>  
> -		/* Check for struct log field offsets */
> +		/* Check for struct printk_log (or older log) field offsets */
>  		if (memcmp("OFFSET(log.ts_nsec)=", pos, 20) == 0)
>  			log_offset_ts_nsec = strtoull(pos + 20, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.ts_nsec)=", pos, 27) == 0)
> +			log_offset_ts_nsec = strtoull(pos + 27, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.len)=", pos, 16) == 0)
>  			log_offset_len = strtoul(pos + 16, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.len)=", pos, 23) == 0)
> +			log_offset_len = strtoul(pos + 23, NULL, 10);
>  
>  		if (memcmp("OFFSET(log.text_len)=", pos, 21) == 0)
>  			log_offset_text_len = strtoul(pos + 21, NULL, 10);
> +		if (memcmp("OFFSET(printk_log.text_len)=", pos, 28) == 0)
> +			log_offset_text_len = strtoul(pos + 28, NULL, 10);
>  
>  		if (last_line)
>  			break;
> -- 
> 1.7.1
> 
> 
> _______________________________________________
> kexec mailing list
> kexec@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kexec

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Lubomir Rintel]

Hi,

On Wed, 2013-10-30 at 11:12 +0800, WANG Chao wrote:
> On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> > Symbol name changed with the following commit:
> > 62e32ac printk: rename struct log to struct printk_log
...
> vmcore-dmesg is really broken here since 3.11 kernel is released. Are
> you going to send an update v3 to address the comments to v2 from Vivek,
> Joe and Dave

I'm not really sure. I thought the suggestions were concerning the Linux
kernel tree, not kexec-tools tree. I might be wrong though, but I'd need
the suggestions to be a bit more specific for me to understand.

Lubo


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Taras Kondratiuk]

On 10/30/2013 10:06 AM, Lubomir Rintel wrote:
> Hi,
> 
> On Wed, 2013-10-30 at 11:12 +0800, WANG Chao wrote:
>> On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
>>> Symbol name changed with the following commit:
>>> 62e32ac printk: rename struct log to struct printk_log
> ...
>> vmcore-dmesg is really broken here since 3.11 kernel is released. Are
>> you going to send an update v3 to address the comments to v2 from Vivek,
>> Joe and Dave
> 
> I'm not really sure. I thought the suggestions were concerning the Linux
> kernel tree, not kexec-tools tree. I might be wrong though, but I'd need
> the suggestions to be a bit more specific for me to understand.
> 

Hi

I've posted a bit different version of fix [1].
Could you please check out that approach?

[1] http://permalink.gmane.org/gmane.linux.kernel.kexec/9522

-- 
Taras Kondratiuk

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[WANG Chao]

On 10/30/13 at 06:06pm, Lubomir Rintel wrote:
> Hi,
> 
> On Wed, 2013-10-30 at 11:12 +0800, WANG Chao wrote:
> > On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> > > Symbol name changed with the following commit:
> > > 62e32ac printk: rename struct log to struct printk_log
> ...
> > vmcore-dmesg is really broken here since 3.11 kernel is released. Are
> > you going to send an update v3 to address the comments to v2 from Vivek,
> > Joe and Dave
> 
> I'm not really sure. I thought the suggestions were concerning the Linux
> kernel tree, not kexec-tools tree. I might be wrong though, but I'd need
> the suggestions to be a bit more specific for me to understand.

That wouldn't be your concern any more. I think your patch is fine.
Thank you for fix.

Hi Simon,

Will you pick up this patch? Vivek and Dave has acked this patch.
And also this patch works for me.

Thanks
WANG Chao

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[WANG Chao]

On 10/31/13 at 01:38pm, WANG Chao wrote:
> On 10/30/13 at 06:06pm, Lubomir Rintel wrote:
> > Hi,
> > 
> > On Wed, 2013-10-30 at 11:12 +0800, WANG Chao wrote:
> > > On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> > > > Symbol name changed with the following commit:
> > > > 62e32ac printk: rename struct log to struct printk_log
> > ...
> > > vmcore-dmesg is really broken here since 3.11 kernel is released. Are
> > > you going to send an update v3 to address the comments to v2 from Vivek,
> > > Joe and Dave
> > 
> > I'm not really sure. I thought the suggestions were concerning the Linux
> > kernel tree, not kexec-tools tree. I might be wrong though, but I'd need
> > the suggestions to be a bit more specific for me to understand.
> 
> That wouldn't be your concern any more.

Hi, Lubomir,

I'm sorry. I was told that this phrase is not friendly and nearly rude.
By that I realy mean I think your patch has no problem and you don't
need to concern what I brought up in my first email.

I'm not a native speaker and that was not my intention. I should have
choosed my words more carefully. Sorry again.

Thanks
WANG Chao

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Simon Horman]

On Thu, Oct 31, 2013 at 01:38:08PM +0800, WANG Chao wrote:
> On 10/30/13 at 06:06pm, Lubomir Rintel wrote:
> > Hi,
> > 
> > On Wed, 2013-10-30 at 11:12 +0800, WANG Chao wrote:
> > > On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> > > > Symbol name changed with the following commit:
> > > > 62e32ac printk: rename struct log to struct printk_log
> > ...
> > > vmcore-dmesg is really broken here since 3.11 kernel is released. Are
> > > you going to send an update v3 to address the comments to v2 from Vivek,
> > > Joe and Dave
> > 
> > I'm not really sure. I thought the suggestions were concerning the Linux
> > kernel tree, not kexec-tools tree. I might be wrong though, but I'd need
> > the suggestions to be a bit more specific for me to understand.
> 
> That wouldn't be your concern any more. I think your patch is fine.
> Thank you for fix.
> 
> Hi Simon,
> 
> Will you pick up this patch? Vivek and Dave has acked this patch.
> And also this patch works for me.

Sure. Please repost it with the Acks.

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Taras Kondratiuk]

On 10/30/2013 11:20 PM, Simon Horman wrote:
> On Thu, Oct 31, 2013 at 01:38:08PM +0800, WANG Chao wrote:
>> On 10/30/13 at 06:06pm, Lubomir Rintel wrote:
>>> Hi,
>>>
>>> On Wed, 2013-10-30 at 11:12 +0800, WANG Chao wrote:
>>>> On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
>>>>> Symbol name changed with the following commit:
>>>>> 62e32ac printk: rename struct log to struct printk_log
>>> ...
>>>> vmcore-dmesg is really broken here since 3.11 kernel is released. Are
>>>> you going to send an update v3 to address the comments to v2 from Vivek,
>>>> Joe and Dave
>>>
>>> I'm not really sure. I thought the suggestions were concerning the Linux
>>> kernel tree, not kexec-tools tree. I might be wrong though, but I'd need
>>> the suggestions to be a bit more specific for me to understand.
>>
>> That wouldn't be your concern any more. I think your patch is fine.
>> Thank you for fix.
>>
>> Hi Simon,
>>
>> Will you pick up this patch? Vivek and Dave has acked this patch.
>> And also this patch works for me.
> 
> Sure. Please repost it with the Acks.

I have concerns about this patch.
There were a lot of hardcoded lenght values in the original code.
Now there are twice as much and we need to keep adding hardcoded
values if any additional parameter should be read of renamed.

Please eveluate this patch [1]. It should remove both:
hardcoded values and duplicated reading for two buffer names.

[1] http://permalink.gmane.org/gmane.linux.kernel.kexec/9522

-- 
Taras Kondratiuk

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Simon Horman]

On Thu, Oct 31, 2013 at 09:42:32AM -0700, Taras Kondratiuk wrote:
> On 10/30/2013 11:20 PM, Simon Horman wrote:
> > On Thu, Oct 31, 2013 at 01:38:08PM +0800, WANG Chao wrote:
> >> On 10/30/13 at 06:06pm, Lubomir Rintel wrote:
> >>> Hi,
> >>>
> >>> On Wed, 2013-10-30 at 11:12 +0800, WANG Chao wrote:
> >>>> On 09/19/13 at 04:49pm, Lubomir Rintel wrote:
> >>>>> Symbol name changed with the following commit:
> >>>>> 62e32ac printk: rename struct log to struct printk_log
> >>> ...
> >>>> vmcore-dmesg is really broken here since 3.11 kernel is released. Are
> >>>> you going to send an update v3 to address the comments to v2 from Vivek,
> >>>> Joe and Dave
> >>>
> >>> I'm not really sure. I thought the suggestions were concerning the Linux
> >>> kernel tree, not kexec-tools tree. I might be wrong though, but I'd need
> >>> the suggestions to be a bit more specific for me to understand.
> >>
> >> That wouldn't be your concern any more. I think your patch is fine.
> >> Thank you for fix.
> >>
> >> Hi Simon,
> >>
> >> Will you pick up this patch? Vivek and Dave has acked this patch.
> >> And also this patch works for me.
> > 
> > Sure. Please repost it with the Acks.
> 
> I have concerns about this patch.
> There were a lot of hardcoded lenght values in the original code.
> Now there are twice as much and we need to keep adding hardcoded
> values if any additional parameter should be read of renamed.
> 
> Please eveluate this patch [1]. It should remove both:
> hardcoded values and duplicated reading for two buffer names.
> 
> [1] http://permalink.gmane.org/gmane.linux.kernel.kexec/9522

Please repost the patch.

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH v2] vmcore-dmesg: Understand >= v3.11-rc4 dmesg

[Vivek Goyal]

On Thu, Oct 31, 2013 at 09:42:32AM -0700, Taras Kondratiuk wrote:

[..]

This patch has not gone it for quite some time and vmcore-dmesg continues
to fail for me with upstream kernels. So it is time to do something
about it.

> I have concerns about this patch.
> There were a lot of hardcoded lenght values in the original code.

One can easily fix that by using strlen().

> Now there are twice as much and we need to keep adding hardcoded
> values if any additional parameter should be read of renamed.
> 
> Please eveluate this patch [1]. It should remove both:
> hardcoded values and duplicated reading for two buffer names.

I don't like excessive usage of macros. And in your patch you are
using nested marcros.

While it might make look code smaller, readability of code becomes
a problem. 

So I personally like simpler approach. And please use strlen() for
fixing the hardcoding of length issue.

Thanks
Vivek

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec