mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Hans Zhang <18255117159@163.com>
To: Gerd Bayer <gbayer@linux.ibm.com>,
	Manivannan Sadhasivam <mani@kernel.org>,
	Hans Zhang <hans.zhang@cixtech.com>
Cc: "Arnd Bergmann" <arnd@kernel.org>,
	"Bjorn Helgaas" <helgaas@kernel.org>,
	bhelgaas@google.com, "Alexander Gordeev" <agordeev@linux.ibm.com>,
	"Christian Borntraeger" <borntraeger@linux.ibm.com>,
	"Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>,
	jingoohan1@gmail.com,
	"Krzysztof Wilczyński" <kwilczynski@kernel.org>,
	linux-kernel@vger.kernel.org, linux-s390@vger.kernel.org,
	linux-next <linux-next@vger.kernel.org>,
	linux-pci@vger.kernel.org,
	"Lorenzo Pieralisi" <lpieralisi@kernel.org>,
	"Rob Herring" <robh@kernel.org>,
	"Niklas Schnelle" <schnelle@linux.ibm.com>,
	geert@linux-m68k.org
Subject: Re: [PATCH] PCI: Fix endianness issues in pci_bus_read_config()
Date: Sat, 2 Aug 2025 00:54:27 +0800	[thread overview]
Message-ID: <659b8389-16a7-423b-a231-5489c7cc0da9@163.com> (raw)
In-Reply-To: <06f16b1a55eede3dc3e0bf31ff14eca89ab6f009.camel@linux.ibm.com>



On 2025/8/1 19:30, Gerd Bayer wrote:
> On Fri, 2025-08-01 at 16:24 +0530, Manivannan Sadhasivam wrote:
> 
> <--- snip --->
> 
>>>>>>> The pci_bus_read_config() interface itself may have been a
>>>>>>> mistake, can't the callers just use the underlying helpers
>>>>>>> directly?
>>>>>>>
>>>>>>
>>>>>> They can! Since the callers of this API is mostly the macros, we can easily
>>>>>> implement the logic to call relevant accessors based on the requested size.
>>>>>>
>>>>>> Hans, could you please respin the series based the feedback since the series is
>>>>>> dropped for 6.17.
>>>>>>
>>>>>
>>>>> Dear all,
>>>>>
>>>>> I am once again deeply sorry for the problems that occurred in this series.
>>>>> I only test pulling the ARM platform.
>>>>>
>>>>> Thank you very much, Gerd, for reporting the problem.
> 
> no worries!
> 
>>>>> Thank you all for your discussions and suggestions for revision.
>>>>>
>>>>> Hi Mani,
>>>>>
>>>>> Geert provided a solution. My patch based on this is as follows. Please
>>>>> check if there are any problems.
>>>>> https://lore.kernel.org/linux-pci/CAMuHMdVwFeV46oCid_sMHjXfP+yyGTpBfs9t3uaa=wRxNcSOAQ@mail.gmail.com/
>>>>>
>>>>> Also, please ask Gerd to help test whether it works properly. Thank you very
>>>>> much.
>>>>>
> 
> I found Geert's proposal intriguing for a quick resolution of the
> issue. Yet, I have not tried that proposal, though.
> 

Hi Gerd,

As I mentioned in my reply to Mani's email, the data ultimately read 
here is also a forced type conversion.

#define PCI_OP_READ(size, type, len) \
int noinline pci_bus_read_config_##size \
	(struct pci_bus *bus, unsigned int devfn, int pos, type *value)	\
{									\
	unsigned long flags;						\
	u32 data = 0;							\
	int res;							\
									\
	if (PCI_##size##_BAD)						\
		return PCIBIOS_BAD_REGISTER_NUMBER;			\
									\
	pci_lock_config(flags);						\
	res = bus->ops->read(bus, devfn, pos, len, &data);		\
	if (res)							\
		PCI_SET_ERROR_RESPONSE(value);				\
	else								\
		*value = (type)data;					\
	pci_unlock_config(flags);					\
									\
	return res;							\
}

And this function. Could it be that I misunderstood something?

int pci_generic_config_read(struct pci_bus *bus, unsigned int devfn,
			    int where, int size, u32 *val)
{
	void __iomem *addr;

	addr = bus->ops->map_bus(bus, devfn, where);
	if (!addr)
		return PCIBIOS_DEVICE_NOT_FOUND;

	if (size == 1)
		*val = readb(addr);
	else if (size == 2)
		*val = readw(addr);
	else
		*val = readl(addr);

	return PCIBIOS_SUCCESSFUL;
}
EXPORT_SYMBOL_GPL(pci_generic_config_read);

> Instead I spent some more cycles on Lukas' and Mani's question about
> the value of the pci_bus_read_config() helper. So I changed
> PCI_FIND_NEXT_CAP and PCI_FIND_NEXT_EXT_CAP to use size-aware versions
> of read_cfg accessor functions like this:
> 
> diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h
> index ac954584d991..9e2f75ede95f 100644
> --- a/drivers/pci/pci.h
> +++ b/drivers/pci/pci.h
> @@ -109,17 +109,17 @@ int pci_bus_read_config(void *priv, unsigned int
> devfn, int where, u32 size,
>   ({
> \
>          int __ttl = PCI_FIND_CAP_TTL;
> \
>          u8 __id, __found_pos = 0;
> \
> -       u32 __pos = (start);
> \
> -       u32 __ent;
> \
> +       u8 __pos = (start);
> \
> +       u16 __ent;
> \
>                                                                        
> \
> -       read_cfg(args, __pos, 1, &__pos);
> \
> +       read_cfg##_byte(args, __pos, &__pos);
> \
>                                                                        
> \
>          while (__ttl--) {
> \
>                  if (__pos < PCI_STD_HEADER_SIZEOF)
> \
>                          break;
> \
>                                                                        
> \
>                  __pos = ALIGN_DOWN(__pos, 4);
> \
> -               read_cfg(args, __pos, 2, &__ent);
> \
> +               read_cfg##_word(args, __pos, &__ent);
> \
>                                                                        
> \
>                  __id = FIELD_GET(PCI_CAP_ID_MASK, __ent);
> \
>                  if (__id == 0xff)
> \
> @@ -158,7 +158,7 @@ int pci_bus_read_config(void *priv, unsigned int
> devfn, int where, u32 size,
>                                                                        
> \
>          __ttl = (PCI_CFG_SPACE_EXP_SIZE - PCI_CFG_SPACE_SIZE) / 8;
> \
>          while (__ttl-- > 0 && __pos >= PCI_CFG_SPACE_SIZE) {
> \
> -               __ret = read_cfg(args, __pos, 4, &__header);
> \
> +               __ret = read_cfg##_dword(args, __pos, &__header);
> \
>                  if (__ret != PCIBIOS_SUCCESSFUL)
> \
>                          break;
> \
>                                                                        
> \
> 
> 
> This fixes the issue for s390's use-cases. With that
> pci_bus_read_config() becomes unused - and could be removed in further
> refinements.
>                                                                        
> However, this probably breaks your dwc and cdns use-cases. I think,
> with the accessor functions for dwc and cadence changed to follow the
> {_byte|_word|_dword} naming pattern they could be used straight out of
> PCI_FIND_NEXT_{EXT_}CAP, too. Then, dw_pcie_read_cfg() and
> cdns_pcie_read_cfg become obsolete as well.
> 
> Thoughts?

In my opinion, it's no problem. I can provide the corresponding function 
of {_byte / _word / _dword}. But it is necessary to know Bjorn, Mani, 
Arnd, Lukas... Their viewpoints or suggestions.

Best regards,
Hans


  reply	other threads:[~2025-08-01 16:55 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-07-16 16:11 [PATCH v14 0/7] Refactor capability search into common macros Hans Zhang
2025-07-16 16:11 ` [PATCH v14 1/7] PCI: Introduce generic bus config read helper function Hans Zhang
2025-07-16 16:11 ` [PATCH v14 2/7] PCI: Clean up __pci_find_next_cap_ttl() readability Hans Zhang
2025-07-16 16:11 ` [PATCH v14 3/7] PCI: Refactor standard capability search into common macro Hans Zhang
2025-07-16 16:12 ` [PATCH v14 4/7] PCI: Refactor extended " Hans Zhang
2025-07-16 16:12 ` [PATCH v14 5/7] PCI: dwc: Use common PCI host bridge APIs for finding the capabilities Hans Zhang
2025-07-16 16:12 ` [PATCH v14 6/7] PCI: cadence: " Hans Zhang
2025-07-16 16:12 ` [PATCH v14 7/7] PCI: cadence: Use cdns_pcie_find_*capability to avoid hardcode Hans Zhang
2025-07-16 23:11 ` [PATCH v14 0/7] Refactor capability search into common macros Bjorn Helgaas
2025-07-31  7:32   ` [REGRESSION] next/master: suspect endianness issue in common PCI capability search macro Gerd Bayer
2025-07-31 17:38     ` [PATCH] PCI: Fix endianness issues in pci_bus_read_config() Gerd Bayer
2025-07-31 18:39       ` Bjorn Helgaas
2025-07-31 19:01         ` Arnd Bergmann
2025-08-01  8:18           ` Manivannan Sadhasivam
2025-08-01  9:25             ` Hans Zhang
2025-08-01  9:47               ` Manivannan Sadhasivam
2025-08-01 10:06                 ` Hans Zhang
2025-08-01 10:54                   ` Manivannan Sadhasivam
2025-08-01 11:30                     ` Gerd Bayer
2025-08-01 16:54                       ` Hans Zhang [this message]
2025-08-01 18:08                         ` Keith Busch
2025-08-02 15:23                           ` Hans Zhang
2025-08-02 15:40                             ` Arnd Bergmann
2025-08-04  3:06                       ` Hans Zhang
2025-08-04  8:03                         ` Arnd Bergmann
2025-08-04  8:25                           ` Hans Zhang
2025-08-04 10:09                         ` Gerd Bayer
2025-08-12 14:44                           ` Hans Zhang
2025-08-13  7:47                             ` Niklas Schnelle
2025-08-13  7:50                               ` Hans Zhang
2025-08-04 14:33                         ` Bjorn Helgaas
2025-08-04 15:04                           ` Hans Zhang
2025-08-01 16:47                     ` Hans Zhang
2025-07-31 18:53       ` Lukas Wunner
2025-08-01  7:52       ` Geert Uytterhoeven
2025-08-01 13:00   ` [PATCH v14 0/7] Refactor capability search into common macros Bjorn Helgaas

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=659b8389-16a7-423b-a231-5489c7cc0da9@163.com \
    --to=18255117159@163.com \
    --cc=agordeev@linux.ibm.com \
    --cc=arnd@kernel.org \
    --cc=bhelgaas@google.com \
    --cc=borntraeger@linux.ibm.com \
    --cc=gbayer@linux.ibm.com \
    --cc=geert@linux-m68k.org \
    --cc=hans.zhang@cixtech.com \
    --cc=helgaas@kernel.org \
    --cc=ilpo.jarvinen@linux.intel.com \
    --cc=jingoohan1@gmail.com \
    --cc=kwilczynski@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-next@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=lpieralisi@kernel.org \
    --cc=mani@kernel.org \
    --cc=robh@kernel.org \
    --cc=schnelle@linux.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®