linux-kernel - Re: [PATCH] PCI/PTM: Build debugfs code only if CONFIG_DEBUG

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <aoqcrof6edexuem4b2xtboiiwwihyolkfpwj7pchy32mgp2o4r@ib5rlna4z2v7>
Date: Mon, 9 Jun 2025 08:43:03 +0530
From: Manivannan Sadhasivam <mani@...nel.org>
To: Eric Biggers <ebiggers@...nel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>, 
	bhelgaas@...gle.com, linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] PCI/PTM: Build debugfs code only if CONFIG_DEBUG_FS is
 enabled

On Sun, Jun 08, 2025 at 05:08:09PM -0700, Eric Biggers wrote:
> On Sun, Jun 08, 2025 at 09:03:05AM +0530, Manivannan Sadhasivam wrote:
> > Otherwise, the following build error will happen for CONFIG_DEBUG_FS=n &&
> > CONFIG_PCIE_PTM=y.
> > 
> >     drivers/pci/pcie/ptm.c:498:25: error: redefinition of 'pcie_ptm_create_debugfs'
> >       498 | struct pci_ptm_debugfs *pcie_ptm_create_debugfs(struct device *dev, void *pdata,
> >           |                         ^
> >     ./include/linux/pci.h:1915:2: note: previous definition is here
> >      1915 | *pcie_ptm_create_debugfs(struct device *dev, void *pdata,
> >           |  ^
> >     drivers/pci/pcie/ptm.c:546:6: error: redefinition of 'pcie_ptm_destroy_debugfs'
> >       546 | void pcie_ptm_destroy_debugfs(struct pci_ptm_debugfs *ptm_debugfs)
> >           |      ^
> >     ./include/linux/pci.h:1918:1: note: previous definition is here
> >      1918 | pcie_ptm_destroy_debugfs(struct pci_ptm_debugfs *ptm_debugfs) { }
> >           |
> > 
> > Fixes: 132833405e61 ("PCI: Add debugfs support for exposing PTM context")
> > Reported-by: Eric Biggers <ebiggers@...nel.org>
> > Closes: https://lore.kernel.org/linux-pci/20250607025506.GA16607@sol
> > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
> > ---
> >  drivers/pci/pcie/ptm.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/drivers/pci/pcie/ptm.c b/drivers/pci/pcie/ptm.c
> > index ee5f615a9023..4bd73f038ffb 100644
> > --- a/drivers/pci/pcie/ptm.c
> > +++ b/drivers/pci/pcie/ptm.c
> > @@ -254,6 +254,7 @@ bool pcie_ptm_enabled(struct pci_dev *dev)
> >  }
> >  EXPORT_SYMBOL(pcie_ptm_enabled);
> >  
> > +#if IS_ENABLED(CONFIG_DEBUG_FS)
> >  static ssize_t context_update_write(struct file *file, const char __user *ubuf,
> >  			     size_t count, loff_t *ppos)
> >  {
> > @@ -552,3 +553,4 @@ void pcie_ptm_destroy_debugfs(struct pci_ptm_debugfs *ptm_debugfs)
> >  	debugfs_remove_recursive(ptm_debugfs->debugfs);
> >  }
> >  EXPORT_SYMBOL_GPL(pcie_ptm_destroy_debugfs);
> > +#endif
> 
> Tested-by: Eric Biggers <ebiggers@...nel.org>
> 
> DEBUG_FS is a bool, not a tristate.  It would be more conventional to write
> '#ifdef CONFIG_DEBUG_FS' instead of '#if IS_ENABLED(CONFIG_DEBUG_FS)'.
> 

I intentionally wanted to avoid using '#ifdef' since mixing both '#ifdef' and
'#if IS_ENABLED' always create issues when one mistakenly uses the former for
tristate symbols. So I always prefer using '#if IS_ENABLED'.

> Also, the #if guards 300 lines of code.  Perhaps they belong in a separate .c
> file?
> 

Perhaps, but not for this fix patch.

- Mani

-- 
மணிவண்ணன் சதாசிவம்