| 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
| ||
|
Message-Id: <1734678.Uaq7DTajoE@hactar>
Date: Tue, 16 Aug 2016 14:00:31 -0300
From: Thiago Jung Bauermann <bauerman@...ux.vnet.ibm.com>
To: Andrew Morton <akpm@...ux-foundation.org>
Cc: kexec@...ts.infradead.org, linux-security-module@...r.kernel.org,
linux-ima-devel@...ts.sourceforge.net,
linuxppc-dev@...ts.ozlabs.org, linux-kernel@...r.kernel.org,
x86@...nel.org, Eric Biederman <ebiederm@...ssion.com>,
Dave Young <dyoung@...hat.com>,
Vivek Goyal <vgoyal@...hat.com>, Baoquan He <bhe@...hat.com>,
Michael Ellerman <mpe@...erman.id.au>,
Benjamin Herrenschmidt <benh@...nel.crashing.org>,
Paul Mackerras <paulus@...ba.org>,
Stewart Smith <stewart@...ux.vnet.ibm.com>,
Samuel Mendoza-Jonas <sam@...dozajonas.com>,
Mimi Zohar <zohar@...ux.vnet.ibm.com>,
Eric Richter <erichte@...ux.vnet.ibm.com>,
Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>,
Petko Manolov <petkan@...-labs.com>,
David Laight <David.Laight@...lab.com>,
Balbir Singh <bsingharora@...il.com>
Subject: Re: [PATCH v2 4/6] kexec_file: Add mechanism to update kexec segments.
Hello Andrew,
Thank you for your review!
Am Montag, 15 August 2016, 15:27:56 schrieb Andrew Morton:
> On Sat, 13 Aug 2016 00:18:23 -0300 Thiago Jung Bauermann <bauerman@...ux.vnet.ibm.com> wrote:
> > +/**
> > + * kexec_update_segment - update the contents of a kimage segment
> > + * @buffer: New contents of the segment.
> > + * @bufsz: @buffer size.
> > + * @load_addr: Segment's physical address in the next kernel.
> > + * @memsz: Segment size.
> > + *
> > + * This function assumes kexec_mutex is held.
> > + *
> > + * Return: 0 on success, negative errno on error.
> > + */
> > +int kexec_update_segment(const char *buffer, unsigned long bufsz,
> > + unsigned long load_addr, unsigned long memsz)
> > +{
> > + int i;
> > + unsigned long entry;
> > + unsigned long *ptr = NULL;
> > + void *dest = NULL;
> > +
> > + if (kexec_image == NULL) {
> > + pr_err("Can't update segment: no kexec image loaded.\n");
> > + return -EINVAL;
> > + }
> > +
> > + /*
> > + * kexec_add_buffer rounds up segment sizes to PAGE_SIZE, so
> > + * we have to do it here as well.
> > + */
> > + memsz = ALIGN(memsz, PAGE_SIZE);
> > +
> > + for (i = 0; i < kexec_image->nr_segments; i++)
> > + /* We only support updating whole segments. */
> > + if (load_addr == kexec_image->segment[i].mem &&
> > + memsz == kexec_image->segment[i].memsz) {
> > + if (kexec_image->segment[i].do_checksum) {
> > + pr_err("Trying to update non-modifiable segment.\n");
> > + return -EINVAL;
> > + }
> > +
> > + break;
> > + }
> > + if (i == kexec_image->nr_segments) {
> > + pr_err("Couldn't find segment to update: 0x%lx, size 0x%lx\n",
> > + load_addr, memsz);
> > + return -EINVAL;
> > + }
> > +
> > + for (entry = kexec_image->head; !(entry & IND_DONE) && memsz;
> > + entry = *ptr++) {
> > + void *addr = (void *) (entry & PAGE_MASK);
> > +
> > + switch (entry & IND_FLAGS) {
> > + case IND_DESTINATION:
> > + dest = addr;
> > + break;
> > + case IND_INDIRECTION:
> > + ptr = __va(addr);
> > + break;
> > + case IND_SOURCE:
> > + /* Shouldn't happen, but verify just to be safe. */
> > + if (dest == NULL) {
> > + pr_err("Invalid kexec entries list.");
> > + return -EINVAL;
> > + }
> > +
> > + if (dest == (void *) load_addr) {
> > + struct page *page;
> > + char *ptr;
> > + size_t uchunk, mchunk;
> > +
> > + page = kmap_to_page(addr);
> > +
> > + ptr = kmap(page);
>
> kmap_atomic() could be used here, and it is appreciably faster.
Good idea. The patch below implements your suggestion.
This has a consequence for patch 5/6 in this series, because it makes
this code be used in the path of the kexec_file_load and
kexec_load syscalls.
In the latter case, there's a call to copy_from_user and thus kmap_atomic
can't be used. I can change the patch to use kmap_atomic if
state->from_kernel is true and kmap otherwise, but perhaps this is one more
hint that patch 5/6 is not a very good idea after all.
--
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center
Subject: [PATCH v2 4/6] kexec_file: Add mechanism to update kexec segments.
kexec_update_segment allows a given segment in kexec_image to have
its contents updated. This is useful if the current kernel wants to
send information to the next kernel that is up-to-date at the time of
reboot.
Signed-off-by: Thiago Jung Bauermann <bauerman@...ux.vnet.ibm.com>
---
include/linux/kexec.h | 2 ++
kernel/kexec_core.c | 99 +++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 101 insertions(+)
diff --git a/include/linux/kexec.h b/include/linux/kexec.h
index 37eea32fdff1..14dda81e3e01 100644
--- a/include/linux/kexec.h
+++ b/include/linux/kexec.h
@@ -259,6 +259,8 @@ extern int kexec_purgatory_get_set_symbol(struct kimage *image,
unsigned int size, bool get_value);
extern void *kexec_purgatory_get_symbol_addr(struct kimage *image,
const char *name);
+int kexec_update_segment(const char *buffer, unsigned long bufsz,
+ unsigned long load_addr, unsigned long memsz);
extern void __crash_kexec(struct pt_regs *);
extern void crash_kexec(struct pt_regs *);
int kexec_should_crash(struct task_struct *);
diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
index 561675589511..9782b292714e 100644
--- a/kernel/kexec_core.c
+++ b/kernel/kexec_core.c
@@ -721,6 +721,105 @@ static struct page *kimage_alloc_page(struct kimage *image,
return page;
}
+/**
+ * kexec_update_segment - update the contents of a kimage segment
+ * @buffer: New contents of the segment.
+ * @bufsz: @buffer size.
+ * @load_addr: Segment's physical address in the next kernel.
+ * @memsz: Segment size.
+ *
+ * This function assumes kexec_mutex is held.
+ *
+ * Return: 0 on success, negative errno on error.
+ */
+int kexec_update_segment(const char *buffer, unsigned long bufsz,
+ unsigned long load_addr, unsigned long memsz)
+{
+ int i;
+ unsigned long entry;
+ unsigned long *ptr = NULL;
+ void *dest = NULL;
+
+ if (kexec_image == NULL) {
+ pr_err("Can't update segment: no kexec image loaded.\n");
+ return -EINVAL;
+ }
+
+ /*
+ * kexec_add_buffer rounds up segment sizes to PAGE_SIZE, so
+ * we have to do it here as well.
+ */
+ memsz = ALIGN(memsz, PAGE_SIZE);
+
+ for (i = 0; i < kexec_image->nr_segments; i++)
+ /* We only support updating whole segments. */
+ if (load_addr == kexec_image->segment[i].mem &&
+ memsz == kexec_image->segment[i].memsz) {
+ if (kexec_image->segment[i].do_checksum) {
+ pr_err("Trying to update non-modifiable segment.\n");
+ return -EINVAL;
+ }
+
+ break;
+ }
+ if (i == kexec_image->nr_segments) {
+ pr_err("Couldn't find segment to update: 0x%lx, size 0x%lx\n",
+ load_addr, memsz);
+ return -EINVAL;
+ }
+
+ for (entry = kexec_image->head; !(entry & IND_DONE) && memsz;
+ entry = *ptr++) {
+ void *addr = (void *) (entry & PAGE_MASK);
+
+ switch (entry & IND_FLAGS) {
+ case IND_DESTINATION:
+ dest = addr;
+ break;
+ case IND_INDIRECTION:
+ ptr = __va(addr);
+ break;
+ case IND_SOURCE:
+ /* Shouldn't happen, but verify just to be safe. */
+ if (dest == NULL) {
+ pr_err("Invalid kexec entries list.");
+ return -EINVAL;
+ }
+
+ if (dest == (void *) load_addr) {
+ struct page *page;
+ char *ptr;
+ size_t uchunk, mchunk;
+
+ page = kmap_to_page(addr);
+
+ ptr = kmap_atomic(page);
+ ptr += load_addr & ~PAGE_MASK;
+ mchunk = min_t(size_t, memsz,
+ PAGE_SIZE - (load_addr & ~PAGE_MASK));
+ uchunk = min(bufsz, mchunk);
+ memcpy(ptr, buffer, uchunk);
+
+ kunmap_atomic(ptr);
+
+ bufsz -= uchunk;
+ load_addr += mchunk;
+ buffer += mchunk;
+ memsz -= mchunk;
+ }
+ dest += PAGE_SIZE;
+ }
+
+ /* Shouldn't happen, but verify just to be safe. */
+ if (ptr == NULL) {
+ pr_err("Invalid kexec entries list.");
+ return -EINVAL;
+ }
+ }
+
+ return 0;
+}
+
static int kimage_load_normal_segment(struct kimage *image,
struct kexec_segment *segment)
{
--
1.9.1
Powered by blists - more mailing lists