From: Valentine Sinitsyn <valesini@yandex-team.ru>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Tejun Heo <tj@kernel.org>, Bjorn Helgaas <bhelgaas@google.com>,
Dan Williams <dan.j.williams@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>, linux-kernel@vger.kernel.org
Subject: [PATCH v4 2/2] PCI: Implement custom llseek for sysfs resource entries
Date: Tue, 22 Aug 2023 16:54:55 +0500 [thread overview]
Message-ID: <20230822115455.310222-2-valesini@yandex-team.ru> (raw)
In-Reply-To: <20230822075107.224512-2-valesini@yandex-team.ru>
Since commit 636b21b50152 ("PCI: Revoke mappings like devmem"), mmappable
sysfs entries have started to receive their f_mapping from the iomem
pseudo filesystem, so that CONFIG_IO_STRICT_DEVMEM is honored in sysfs
(and procfs) as well as in /dev/[k]mem.
This resulted in a userspace-visible regression:
1. Open a sysfs PCI resource file (eg. /sys/bus/pci/devices/*/resource0)
2. Use lseek(fd, 0, SEEK_END) to determine its size
Expected result: a PCI region size is returned.
Actual result: 0 is returned.
The reason is that PCI resource files residing in sysfs use
generic_file_llseek(), which relies on f_mapping->host inode to get the
file size. As f_mapping is now redefined, f_mapping->host points to an
anonymous zero-sized iomem_inode which has nothing to do with sysfs file
in question.
Implement a custom llseek method for sysfs PCI resources, which is
almost the same as proc_bus_pci_lseek() used for procfs entries.
This makes sysfs and procfs entries consistent with regards to seeking,
but also introduces userspace-visible changes to seeking PCI resources
in sysfs:
- SEEK_DATA and SEEK_HOLE are no longer supported;
- Seeking past the end of the file is prohibited while previously
offsets up to MAX_NON_LFS were accepted (reading from these offsets
was always invalid).
Fixes: 636b21b50152 ("PCI: Revoke mappings like devmem")
Cc: stable@vger.kernel.org
Signed-off-by: Valentine Sinitsyn <valesini@yandex-team.ru>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
---
drivers/pci/pci-sysfs.c | 27 ++++++++++++++++++++++++++-
1 file changed, 26 insertions(+), 1 deletion(-)
diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c
index ab32a91f287b..2b9d93f11dd1 100644
--- a/drivers/pci/pci-sysfs.c
+++ b/drivers/pci/pci-sysfs.c
@@ -836,6 +836,12 @@ static const struct attribute_group pci_dev_config_attr_group = {
};
#ifdef HAVE_PCI_LEGACY
+
+static loff_t pci_llseek_resource(struct file *filep,
+ struct kobject *kobj,
+ struct bin_attribute *attr,
+ loff_t offset, int whence);
+
/**
* pci_read_legacy_io - read byte(s) from legacy I/O port space
* @filp: open sysfs file
@@ -967,6 +973,8 @@ void pci_create_legacy_files(struct pci_bus *b)
b->legacy_io->attr.mode = 0600;
b->legacy_io->read = pci_read_legacy_io;
b->legacy_io->write = pci_write_legacy_io;
+ /* See pci_create_attr() for motivation */
+ b->legacy_io->llseek = pci_llseek_resource;
b->legacy_io->mmap = pci_mmap_legacy_io;
b->legacy_io->f_mapping = iomem_get_mapping;
pci_adjust_legacy_attr(b, pci_mmap_io);
@@ -981,6 +989,8 @@ void pci_create_legacy_files(struct pci_bus *b)
b->legacy_mem->size = 1024*1024;
b->legacy_mem->attr.mode = 0600;
b->legacy_mem->mmap = pci_mmap_legacy_mem;
+ /* See pci_create_attr() for motivation */
+ b->legacy_io->llseek = pci_llseek_resource;
b->legacy_mem->f_mapping = iomem_get_mapping;
pci_adjust_legacy_attr(b, pci_mmap_mem);
error = device_create_bin_file(&b->dev, b->legacy_mem);
@@ -1138,6 +1148,14 @@ static ssize_t pci_write_resource_io(struct file *filp, struct kobject *kobj,
return pci_resource_io(filp, kobj, attr, buf, off, count, true);
}
+static loff_t pci_llseek_resource(struct file *filep,
+ struct kobject *kobj __always_unused,
+ struct bin_attribute *attr,
+ loff_t offset, int whence)
+{
+ return fixed_size_llseek(filep, offset, whence, attr->size);
+}
+
/**
* pci_remove_resource_files - cleanup resource files
* @pdev: dev to cleanup
@@ -1195,8 +1213,15 @@ static int pci_create_attr(struct pci_dev *pdev, int num, int write_combine)
res_attr->mmap = pci_mmap_resource_uc;
}
}
- if (res_attr->mmap)
+ if (res_attr->mmap) {
res_attr->f_mapping = iomem_get_mapping;
+ /*
+ * generic_file_llseek() consults f_mapping->host to determine
+ * the file size. As iomem_inode knows nothing about the
+ * attribute, it's not going to work, so override it as well.
+ */
+ res_attr->llseek = pci_llseek_resource;
+ }
res_attr->attr.name = res_attr_name;
res_attr->attr.mode = 0600;
res_attr->size = pci_resource_len(pdev, num);
--
2.34.1
next prev parent reply other threads:[~2023-08-22 11:55 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-14 19:08 [PATCH] kernfs: implement custom llseek method to fix userspace regression Valentine Sinitsyn
2023-08-14 20:01 ` Dan Williams
2023-08-15 8:25 ` Valentin Sinitsyn
2023-08-15 15:48 ` Dan Williams
2023-08-17 18:53 ` Valentin Sinitsyn
2023-08-21 7:29 ` [RESEND PATCH v2 1/2] kernfs: sysfs: support custom llseek method for sysfs entries Valentine Sinitsyn
2023-08-21 7:29 ` [RESEND PATCH v2 2/2] PCI: implement custom llseek method for PCI resource entries in sysfs Valentine Sinitsyn
2023-08-21 19:55 ` Bjorn Helgaas
2023-08-22 7:51 ` [PATCH v3 1/2] kernfs: sysfs: support custom llseek method for sysfs entries Valentine Sinitsyn
2023-08-22 7:51 ` [PATCH v3 2/2] PCI: Implement custom llseek for sysfs resource entries Valentine Sinitsyn
2023-08-22 11:54 ` [PATCH v4 1/2] kernfs: sysfs: support custom llseek method for sysfs entries Valentine Sinitsyn
2023-08-22 11:54 ` Valentine Sinitsyn [this message]
2023-08-22 17:43 ` [PATCH v4 2/2] PCI: Implement custom llseek for sysfs resource entries kernel test robot
2023-08-23 12:58 ` [PATCH v5 1/2] kernfs: sysfs: support custom llseek method for sysfs entries Valentine Sinitsyn
2023-08-23 14:37 ` Greg Kroah-Hartman
2023-08-23 14:39 ` Greg Kroah-Hartman
2023-08-23 20:11 ` Valentin Sinitsyn
2023-08-23 12:58 ` [PATCH v5 2/2] PCI: Implement custom llseek for sysfs resource entries Valentine Sinitsyn
2023-08-25 14:10 ` kernel test robot
2023-08-28 7:22 ` Valentin Sinitsyn
2023-08-29 20:02 ` Bjorn Helgaas
2023-08-29 20:26 ` Bjorn Helgaas
2023-09-02 15:50 ` [PATCH v6 1/2] kernfs: sysfs: support custom llseek method for sysfs entries Valentine Sinitsyn
2023-09-05 7:02 ` kernel test robot
2023-09-08 13:49 ` [PATCH v7 " Valentine Sinitsyn
2023-09-08 13:49 ` [PATCH v7 2/2] PCI: Implement custom llseek for sysfs resource entries Valentine Sinitsyn
2023-09-08 13:56 ` [PATCH v8 1/2] kernfs: sysfs: support custom llseek method for sysfs entries Valentine Sinitsyn
2023-09-08 13:56 ` [PATCH v8 2/2] PCI: Implement custom llseek for sysfs resource entries Valentine Sinitsyn
2023-09-02 15:50 ` [PATCH v6 " Valentine Sinitsyn
2023-08-21 23:15 ` [RESEND PATCH v2 2/2] PCI: implement custom llseek method for PCI resource entries in sysfs kernel test robot
2023-08-22 8:09 ` Valentin Sinitsyn
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=20230822115455.310222-2-valesini@yandex-team.ru \
--to=valesini@yandex-team.ru \
--cc=bhelgaas@google.com \
--cc=dan.j.williams@intel.com \
--cc=daniel.vetter@ffwll.ch \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=tj@kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).