io_uring-6.5-2023-07-21

-----BEGIN PGP SIGNATURE-----
 
 iQJEBAABCAAuFiEEwPw5LcreJtl1+l5K99NY+ylx4KYFAmS62/kQHGF4Ym9lQGtl
 cm5lbC5kawAKCRD301j7KXHgpuYYD/9HMf8IG/2i/EmfgHa2O0yyaBprITcMAnCx
 3x5szTz2Q4V5Dwdib1xRNAiT5KF4uoio8JJJWeIjeHTucUnVA7hjwAma4qmmVNlM
 9dVelH5FL85oNEMVlIqpJ2KHWHH3FaQAyvQvOIXODr+DNAOn+NlgL18B6W5g43Xt
 e8/d+vmDflr3KfliNOfADdZebJzB6hR497btScr9bcmVNuxHnz8tPUCugdLz5LQh
 GEGW+iIUIS/Xmfrv2cleO5ANQ245gcCyBBGV/3klj8LndJpeKstFp3dl8mTA372/
 ZLXnckKQerErKEcuvB8XJ3JIjlQR6s4JeCBk8fBi04ibHdRQuSDQgNXoFIKuRI2o
 A4znCK7P/TP7yK1dn7PLFaPtsfUkoUOSuTeMPkh/f8iW/ckurPu7WTQTd8dcyXDj
 50U09Aa1V7TpnTCGkfQGTlX5ihlvRHQhGdqUOnR7NuS/KOaFqxpK3X1fuMwoea+s
 I+BC1hDnSm5uJTzcTY/G4OBXyujlslTbD2bq4Gmpb9JkZVkzhvIpjYHiEmZQJrC2
 EvyI0z8MLARjHia6snn5EOJqrX7EsV7GmYcoWadQh/h1zt5TPyEyY0S4u6HrwtwB
 Rlp/dH60VbrRzoOHVzOr53F505F3DJ/woZXKGIuqTSBNtYudJKkfUbWY4Ga23pPw
 +i1JhigYgg==
 =FL//
 -----END PGP SIGNATURE-----

Merge tag 'io_uring-6.5-2023-07-21' of git://git.kernel.dk/linux

Pull io_uring fixes from Jens Axboe:

 - Fix for io-wq not always honoring REQ_F_NOWAIT, if it was set and
   punted directly (eg via DRAIN) (me)

 - Capability check fix (Ondrej)

 - Regression fix for the mmap changes that went into 6.4, which
   apparently broke IA64 (Helge)

* tag 'io_uring-6.5-2023-07-21' of git://git.kernel.dk/linux:
  ia64: mmap: Consider pgoff when searching for free mapping
  io_uring: Fix io_uring mmap() by using architecture-provided get_unmapped_area()
  io_uring: treat -EAGAIN for REQ_F_NOWAIT as final for io-wq
  io_uring: don't audit the capability check in io_uring_create()

arch/ia64/kernel/sys_ia64.c

@@ -63,7 +63,7 @@ arch_get_unmapped_area (struct file *filp, unsigned long addr, unsigned long len
 	info.low_limit = addr;
 	info.high_limit = TASK_SIZE;
 	info.align_mask = align_mask;
-	info.align_offset = 0;
+	info.align_offset = pgoff << PAGE_SHIFT;
 	return vm_unmapped_area(&info);
 }
 

arch/parisc/kernel/sys_parisc.c

@@ -27,12 +27,17 @@
 #include <linux/elf-randomize.h>
 
 /*
- * Construct an artificial page offset for the mapping based on the physical
+ * Construct an artificial page offset for the mapping based on the virtual
  * address of the kernel file mapping variable.
+ * If filp is zero the calculated pgoff value aliases the memory of the given
+ * address. This is useful for io_uring where the mapping shall alias a kernel
+ * address and a userspace adress where both the kernel and the userspace
+ * access the same memory region.
  */
-#define GET_FILP_PGOFF(filp)		\
-	(filp ? (((unsigned long) filp->f_mapping) >> 8)	\
-		 & ((SHM_COLOUR-1) >> PAGE_SHIFT) : 0UL)
+#define GET_FILP_PGOFF(filp, addr)		\
+	((filp ? (((unsigned long) filp->f_mapping) >> 8)	\
+		 & ((SHM_COLOUR-1) >> PAGE_SHIFT) : 0UL)	\
+	  + (addr >> PAGE_SHIFT))
 
 static unsigned long shared_align_offset(unsigned long filp_pgoff,
 					 unsigned long pgoff)
@@ -112,7 +117,7 @@ static unsigned long arch_get_unmapped_area_common(struct file *filp,
 	do_color_align = 0;
 	if (filp || (flags & MAP_SHARED))
 		do_color_align = 1;
-	filp_pgoff = GET_FILP_PGOFF(filp);
+	filp_pgoff = GET_FILP_PGOFF(filp, addr);
 
 	if (flags & MAP_FIXED) {
 		/* Even MAP_FIXED mappings must reside within TASK_SIZE */

io_uring/io_uring.c

@@ -1948,6 +1948,14 @@ fail:
 		ret = io_issue_sqe(req, issue_flags);
 		if (ret != -EAGAIN)
 			break;
+
+		/*
+		 * If REQ_F_NOWAIT is set, then don't wait or retry with
+		 * poll. -EAGAIN is final for that case.
+		 */
+		if (req->flags & REQ_F_NOWAIT)
+			break;
+
 		/*
 		 * We can get EAGAIN for iopolled IO even though we're
 		 * forcing a sync submission from here, since we can't
@@ -3429,8 +3437,6 @@ static unsigned long io_uring_mmu_get_unmapped_area(struct file *filp,
 			unsigned long addr, unsigned long len,
 			unsigned long pgoff, unsigned long flags)
 {
-	const unsigned long mmap_end = arch_get_mmap_end(addr, len, flags);
-	struct vm_unmapped_area_info info;
 	void *ptr;
 
 	/*
@@ -3445,32 +3451,26 @@ static unsigned long io_uring_mmu_get_unmapped_area(struct file *filp,
 	if (IS_ERR(ptr))
 		return -ENOMEM;
 
-	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
-	info.length = len;
-	info.low_limit = max(PAGE_SIZE, mmap_min_addr);
-	info.high_limit = arch_get_mmap_base(addr, current->mm->mmap_base);
-#ifdef SHM_COLOUR
-	info.align_mask = PAGE_MASK & (SHM_COLOUR - 1UL);
-#else
-	info.align_mask = PAGE_MASK & (SHMLBA - 1UL);
-#endif
-	info.align_offset = (unsigned long) ptr;
-
 	/*
-	 * A failed mmap() very likely causes application failure,
-	 * so fall back to the bottom-up function here. This scenario
-	 * can happen with large stack limits and large mmap()
-	 * allocations.
+	 * Some architectures have strong cache aliasing requirements.
+	 * For such architectures we need a coherent mapping which aliases
+	 * kernel memory *and* userspace memory. To achieve that:
+	 * - use a NULL file pointer to reference physical memory, and
+	 * - use the kernel virtual address of the shared io_uring context
+	 *   (instead of the userspace-provided address, which has to be 0UL
+	 *   anyway).
+	 * For architectures without such aliasing requirements, the
+	 * architecture will return any suitable mapping because addr is 0.
 	 */
-	addr = vm_unmapped_area(&info);
-	if (offset_in_page(addr)) {
-		info.flags = 0;
-		info.low_limit = TASK_UNMAPPED_BASE;
-		info.high_limit = mmap_end;
-		addr = vm_unmapped_area(&info);
-	}
-
-	return addr;
+	filp = NULL;
+	flags |= MAP_SHARED;
+	pgoff = 0;	/* has been translated to ptr above */
+#ifdef SHM_COLOUR
+	addr = (uintptr_t) ptr;
+#else
+	addr = 0UL;
+#endif
+	return current->mm->get_unmapped_area(filp, addr, len, pgoff, flags);
 }
 
 #else /* !CONFIG_MMU */
@@ -3870,7 +3870,7 @@ static __cold int io_uring_create(unsigned entries, struct io_uring_params *p,
 		ctx->syscall_iopoll = 1;
 
 	ctx->compat = in_compat_syscall();
-	if (!capable(CAP_IPC_LOCK))
+	if (!ns_capable_noaudit(&init_user_ns, CAP_IPC_LOCK))
 		ctx->user = get_uid(current_user());
 
 	/*