[PATCH] Revert "printk: export symbols for debug modules"

[PATCH] Revert "printk: export symbols for debug modules"

[Christoph Hellwig]

This reverts commit 3e00123a13d824d63072b1824c9da59cd78356d9.

No, we never export random symbols for out of tree modules.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 kernel/printk/printk.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index 96fc38cb2e8415..7e0b4dd02398db 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -538,14 +538,12 @@ char *log_buf_addr_get(void)
 {
 	return log_buf;
 }
-EXPORT_SYMBOL_GPL(log_buf_addr_get);
 
 /* Return log buffer size */
 u32 log_buf_len_get(void)
 {
 	return log_buf_len;
 }
-EXPORT_SYMBOL_GPL(log_buf_len_get);
 
 /*
  * Define how much of the log buffer we could take at maximum. The value
-- 
2.39.2

Re: [PATCH] Revert "printk: export symbols for debug modules"

[Greg KH]

On Tue, Sep 05, 2023 at 10:19:02AM +0200, Christoph Hellwig wrote:
> This reverts commit 3e00123a13d824d63072b1824c9da59cd78356d9.
> 
> No, we never export random symbols for out of tree modules.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  kernel/printk/printk.c | 2 --
>  1 file changed, 2 deletions(-)

Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Re: [PATCH] Revert "printk: export symbols for debug modules"

[Petr Mladek]

Added people from the original thread into Cc.

On Tue 2023-09-05 10:19:02, Christoph Hellwig wrote:
> This reverts commit 3e00123a13d824d63072b1824c9da59cd78356d9.
> 
> No, we never export random symbols for out of tree modules.

Yeah, I did not have a good feeling about this patch.

I accepted it because there was an intention to upstream
the module, see
https://lore.kernel.org/all/CAAfh-jM0B2Cn579B0CkCrW44pJGGvjs112K+oMuViib+jDKafg@mail.gmail.com/

I think that it was bad decision. We should not export
symbols when there is no in-tree user.

> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Petr Mladek <pmladek@suse.com>

Best Regards,
Petr

Re: [PATCH] Revert "printk: export symbols for debug modules"

[Christoph Hellwig]

On Tue, Sep 05, 2023 at 11:54:37AM +0200, Petr Mladek wrote:
> Yeah, I did not have a good feeling about this patch.
> 
> I accepted it because there was an intention to upstream
> the module, see
> https://lore.kernel.org/all/CAAfh-jM0B2Cn579B0CkCrW44pJGGvjs112K+oMuViib+jDKafg@mail.gmail.com/
> 
> I think that it was bad decision. We should not export
> symbols when there is no in-tree user.

I think raw address to the printk buffer is a bad idea even for an
in-tree module.

Re: [PATCH] Revert "printk: export symbols for debug modules"

[Petr Mladek]

On Tue 2023-09-05 15:20:22, Christoph Hellwig wrote:
> On Tue, Sep 05, 2023 at 11:54:37AM +0200, Petr Mladek wrote:
> > Yeah, I did not have a good feeling about this patch.
> > 
> > I accepted it because there was an intention to upstream
> > the module, see
> > https://lore.kernel.org/all/CAAfh-jM0B2Cn579B0CkCrW44pJGGvjs112K+oMuViib+jDKafg@mail.gmail.com/
> > 
> > I think that it was bad decision. We should not export
> > symbols when there is no in-tree user.
> 
> I think raw address to the printk buffer is a bad idea even for an
> in-tree module.

Yeah, I suggested to use the kmsg_dump() API in the original thread.

Sigh, I made the bad decision in a hurry before vacation. I know
that it is a bad excuse but... Thanks for catching it.

Best Regards,
Petr

PS: I have just sent a pull request with the revert.