[Qemu-devel] Re: lsi: Handle removal of selected devices

[Qemu-devel] Re: lsi: Handle removal of selected devices

[Aurelien Jarno]

This patch:

64d564094cac5f72eeaeb950c442b773a00d3586 is the first bad commit
commit 64d564094cac5f72eeaeb950c442b773a00d3586
Author: Jan Kiszka <jan.kiszka@siemens.com>
Date:   Tue May 4 14:21:03 2010 +0200

    lsi: Handle removal of selected devices

    We must not store references to selected devices as they may be
    hot-removed. Instead, look up the device based on its tag right before
    using it. If the device disappeared, throw an interrupt and disconnect.

    Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
    Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

breaks the versatile machine. qemu-system-arm segfaults during the boot.

-- 
Aurelien Jarno	                        GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

Re: [Qemu-devel] Re: lsi: Handle removal of selected devices

[Aurelien Jarno]

On Fri, May 21, 2010 at 05:49:26PM +0200, Aurelien Jarno wrote:
> This patch:
> 
> 64d564094cac5f72eeaeb950c442b773a00d3586 is the first bad commit
> commit 64d564094cac5f72eeaeb950c442b773a00d3586
> Author: Jan Kiszka <jan.kiszka@siemens.com>
> Date:   Tue May 4 14:21:03 2010 +0200
> 
>     lsi: Handle removal of selected devices
> 
>     We must not store references to selected devices as they may be
>     hot-removed. Instead, look up the device based on its tag right before
>     using it. If the device disappeared, throw an interrupt and disconnect.
> 
>     Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
> 
> breaks the versatile machine. qemu-system-arm segfaults during the boot.

Actually it only seems to be the case with old kernels. I have put the
material to reproduce the bug here:

http://temp.aurel32.net/qemu-versatile/

-- 
Aurelien Jarno	                        GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

Re: [Qemu-devel] Re: lsi: Handle removal of selected devices

[Jan Kiszka]

Aurelien Jarno wrote:
> On Fri, May 21, 2010 at 05:49:26PM +0200, Aurelien Jarno wrote:
>> This patch:
>>
>> 64d564094cac5f72eeaeb950c442b773a00d3586 is the first bad commit
>> commit 64d564094cac5f72eeaeb950c442b773a00d3586
>> Author: Jan Kiszka <jan.kiszka@siemens.com>
>> Date:   Tue May 4 14:21:03 2010 +0200
>>
>>     lsi: Handle removal of selected devices
>>
>>     We must not store references to selected devices as they may be
>>     hot-removed. Instead, look up the device based on its tag right before
>>     using it. If the device disappeared, throw an interrupt and disconnect.
>>
>>     Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>>     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
>>
>> breaks the versatile machine. qemu-system-arm segfaults during the boot.
> 
> Actually it only seems to be the case with old kernels. I have put the
> material to reproduce the bug here:
> 
> http://temp.aurel32.net/qemu-versatile/
> 

Ah, perfect. Will have a look.

Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux

Re: [Qemu-devel] Re: lsi: Handle removal of selected devices

[Aurelien Jarno]

Jan Kiszka a écrit :
> Aurelien Jarno wrote:
>> On Fri, May 21, 2010 at 05:49:26PM +0200, Aurelien Jarno wrote:
>>> This patch:
>>>
>>> 64d564094cac5f72eeaeb950c442b773a00d3586 is the first bad commit
>>> commit 64d564094cac5f72eeaeb950c442b773a00d3586
>>> Author: Jan Kiszka <jan.kiszka@siemens.com>
>>> Date:   Tue May 4 14:21:03 2010 +0200
>>>
>>>     lsi: Handle removal of selected devices
>>>
>>>     We must not store references to selected devices as they may be
>>>     hot-removed. Instead, look up the device based on its tag right before
>>>     using it. If the device disappeared, throw an interrupt and disconnect.
>>>
>>>     Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>>>     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
>>>
>>> breaks the versatile machine. qemu-system-arm segfaults during the boot.
>> Actually it only seems to be the case with old kernels. I have put the
>> material to reproduce the bug here:
>>
>> http://temp.aurel32.net/qemu-versatile/
>>
> 
> Ah, perfect. Will have a look.
> 

Thanks.


-- 
Aurelien Jarno                          GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

[Qemu-devel] [PATCH] lsi: Fix value overflow in request tag processing

[Jan Kiszka]

This fixes a mismerge of 64d564094cac5f72eeaeb950c442b773a00d3586 (wrong
patch version): We need to mask the tag value properly to obtain its
device ID.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
---
 hw/lsi53c895a.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/hw/lsi53c895a.c b/hw/lsi53c895a.c
index 9d3c44d..f5a91ba 100644
--- a/hw/lsi53c895a.c
+++ b/hw/lsi53c895a.c
@@ -543,7 +543,7 @@ static void lsi_do_dma(LSIState *s, int out)
         return;
     }
 
-    id = s->current->tag >> 8;
+    id = (s->current->tag >> 8) & 0xf;
     dev = s->bus.devs[id];
     if (!dev) {
         lsi_bad_selection(s, id);
@@ -745,7 +745,7 @@ static void lsi_do_command(LSIState *s)
     s->sfbr = buf[0];
     s->command_complete = 0;
 
-    id = s->select_tag >> 8;
+    id = (s->select_tag >> 8) & 0xf;
     dev = s->bus.devs[id];
     if (!dev) {
         lsi_bad_selection(s, id);
-- 
1.6.0.2

Re: [Qemu-devel] [PATCH] lsi: Fix value overflow in request tag processing

[Aurelien Jarno]

On Fri, May 21, 2010 at 06:44:59PM +0200, Jan Kiszka wrote:
> This fixes a mismerge of 64d564094cac5f72eeaeb950c442b773a00d3586 (wrong
> patch version): We need to mask the tag value properly to obtain its
> device ID.
> 
> Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>

Thanks for the quick patch, I have applied it.

> ---
>  hw/lsi53c895a.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/hw/lsi53c895a.c b/hw/lsi53c895a.c
> index 9d3c44d..f5a91ba 100644
> --- a/hw/lsi53c895a.c
> +++ b/hw/lsi53c895a.c
> @@ -543,7 +543,7 @@ static void lsi_do_dma(LSIState *s, int out)
>          return;
>      }
>  
> -    id = s->current->tag >> 8;
> +    id = (s->current->tag >> 8) & 0xf;
>      dev = s->bus.devs[id];
>      if (!dev) {
>          lsi_bad_selection(s, id);
> @@ -745,7 +745,7 @@ static void lsi_do_command(LSIState *s)
>      s->sfbr = buf[0];
>      s->command_complete = 0;
>  
> -    id = s->select_tag >> 8;
> +    id = (s->select_tag >> 8) & 0xf;
>      dev = s->bus.devs[id];
>      if (!dev) {
>          lsi_bad_selection(s, id);
> -- 
> 1.6.0.2
> 
> 

-- 
Aurelien Jarno                          GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

Re: [Qemu-devel] Re: lsi: Handle removal of selected devices

[Jan Kiszka]

Jan Kiszka wrote:
> Aurelien Jarno wrote:
>> On Fri, May 21, 2010 at 05:49:26PM +0200, Aurelien Jarno wrote:
>>> This patch:
>>>
>>> 64d564094cac5f72eeaeb950c442b773a00d3586 is the first bad commit
>>> commit 64d564094cac5f72eeaeb950c442b773a00d3586
>>> Author: Jan Kiszka <jan.kiszka@siemens.com>
>>> Date:   Tue May 4 14:21:03 2010 +0200
>>>
>>>     lsi: Handle removal of selected devices
>>>
>>>     We must not store references to selected devices as they may be
>>>     hot-removed. Instead, look up the device based on its tag right before
>>>     using it. If the device disappeared, throw an interrupt and disconnect.
>>>
>>>     Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>>>     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
>>>
>>> breaks the versatile machine. qemu-system-arm segfaults during the boot.
>> Actually it only seems to be the case with old kernels. I have put the
>> material to reproduce the bug here:
>>
>> http://temp.aurel32.net/qemu-versatile/
>>
> 
> Ah, perfect. Will have a look.
> 

Boots fine up to "Waiting for root file system". That's on a 64-bit
host. Are you on 32 bits?

Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux

Re: [Qemu-devel] Re: lsi: Handle removal of selected devices

[Jan Kiszka]

Jan Kiszka wrote:
> Jan Kiszka wrote:
>> Aurelien Jarno wrote:
>>> On Fri, May 21, 2010 at 05:49:26PM +0200, Aurelien Jarno wrote:
>>>> This patch:
>>>>
>>>> 64d564094cac5f72eeaeb950c442b773a00d3586 is the first bad commit
>>>> commit 64d564094cac5f72eeaeb950c442b773a00d3586
>>>> Author: Jan Kiszka <jan.kiszka@siemens.com>
>>>> Date:   Tue May 4 14:21:03 2010 +0200
>>>>
>>>>     lsi: Handle removal of selected devices
>>>>
>>>>     We must not store references to selected devices as they may be
>>>>     hot-removed. Instead, look up the device based on its tag right before
>>>>     using it. If the device disappeared, throw an interrupt and disconnect.
>>>>
>>>>     Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>>>>     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
>>>>
>>>> breaks the versatile machine. qemu-system-arm segfaults during the boot.
>>> Actually it only seems to be the case with old kernels. I have put the
>>> material to reproduce the bug here:
>>>
>>> http://temp.aurel32.net/qemu-versatile/
>>>
>> Ah, perfect. Will have a look.
>>
> 
> Boots fine up to "Waiting for root file system". That's on a 64-bit
> host. Are you on 32 bits?

Forget it - shouldn't blindly run the suggested command line. With the
right qemu is crashes properly.

Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux

[Qemu-devel] Re: lsi: Handle removal of selected devices

[Jan Kiszka]

Aurelien Jarno wrote:
> This patch:
> 
> 64d564094cac5f72eeaeb950c442b773a00d3586 is the first bad commit
> commit 64d564094cac5f72eeaeb950c442b773a00d3586
> Author: Jan Kiszka <jan.kiszka@siemens.com>
> Date:   Tue May 4 14:21:03 2010 +0200
> 
>     lsi: Handle removal of selected devices
> 
>     We must not store references to selected devices as they may be
>     hot-removed. Instead, look up the device based on its tag right before
>     using it. If the device disappeared, throw an interrupt and disconnect.
> 
>     Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
>     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
> 
> breaks the versatile machie. qemu-system-arm segfaults during the boot.

Do you have an image for me? Or some gdb backtrace?

Thanks,
Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux