diff --git a/src/arch/arm/armv7/mmu.c b/src/arch/arm/armv7/mmu.c
index b63497c..8ebb848 100644
--- a/src/arch/arm/armv7/mmu.c
+++ b/src/arch/arm/armv7/mmu.c
@@ -236,7 +236,7 @@
 	printk(BIOS_DEBUG, "Setting address range [%#.8x:%#.8x) as unmapped\n",
 	       start_kb * KiB, (start_kb + size_kb) * KiB);
 	mmu_fill_table(table, (start_kb & mask) / (PAGE_SIZE/KiB),
-		       div_round_up((start_kb + size_kb) & mask, PAGE_SIZE/KiB),
+		       DIV_ROUND_UP((start_kb + size_kb) & mask, PAGE_SIZE/KiB),
 		       (start_kb & ~mask) * KiB, PAGE_SHIFT, 0);
 }
 
@@ -246,7 +246,7 @@
 	       start_mb * MiB, (start_mb + size_mb) * MiB);
 	assert(start_mb + size_mb <= 4 * (GiB/MiB));
 	mmu_fill_table(ttb_buff, start_mb / (BLOCK_SIZE/MiB),
-		       div_round_up(start_mb + size_mb, BLOCK_SIZE/MiB),
+		       DIV_ROUND_UP(start_mb + size_mb, BLOCK_SIZE/MiB),
 		       0, BLOCK_SHIFT, 0);
 }
 
@@ -256,7 +256,7 @@
 	       start_mb * MiB, (start_mb + size_mb) * MiB, attrs[policy].name);
 	assert(start_mb + size_mb <= 4 * (GiB/MiB));
 	mmu_fill_table(ttb_buff, start_mb / (BLOCK_SIZE/MiB),
-		       div_round_up(start_mb + size_mb, BLOCK_SIZE/MiB),
+		       DIV_ROUND_UP(start_mb + size_mb, BLOCK_SIZE/MiB),
 		       0, BLOCK_SHIFT, ATTR_BLOCK | attrs[policy].value);
 }
 
diff --git a/src/commonlib/include/commonlib/helpers.h b/src/commonlib/include/commonlib/helpers.h
index f2acedc..03f4306 100644
--- a/src/commonlib/include/commonlib/helpers.h
+++ b/src/commonlib/include/commonlib/helpers.h
@@ -36,7 +36,6 @@
 #define MAX(a, b) ((a) > (b) ? (a) : (b))
 #endif
 #define ABS(a) (((a) < 0) ? (-(a)) : (a))
-#define CEIL_DIV(a, b)  (((a) + (b) - 1) / (b))
 #define IS_POWER_OF_2(x)  (((x) & ((x) - 1)) == 0)
 #define DIV_ROUND_UP(x, y)  (((x) + (y) - 1) / (y))
 /*
diff --git a/src/cpu/allwinner/a10/clock.c b/src/cpu/allwinner/a10/clock.c
index ae50e06..8f64f06 100644
--- a/src/cpu/allwinner/a10/clock.c
+++ b/src/cpu/allwinner/a10/clock.c
@@ -247,9 +247,9 @@
 	 * will always be in spec, as long as AHB is in spec, although the max
 	 * AHB0 clock we can get is 125 MHz
 	 */
-	axi = CEIL_DIV(actual_mhz, 450);	/* Max 450 MHz */
-	ahb = CEIL_DIV(actual_mhz/axi, 250);	/* Max 250 MHz */
-	apb0 = 2;				/* Max 150 MHz */
+	axi = DIV_ROUND_UP(actual_mhz, 450);		/* Max 450 MHz */
+	ahb = DIV_ROUND_UP(actual_mhz/axi, 250);	/* Max 250 MHz */
+	apb0 = 2;					/* Max 150 MHz */
 
 	ahb_exp = log2_ceil(ahb);
 	ahb = 1 << ahb_exp;
diff --git a/src/cpu/x86/tsc/delay_tsc.c b/src/cpu/x86/tsc/delay_tsc.c
index 04f709f..a589cdb 100644
--- a/src/cpu/x86/tsc/delay_tsc.c
+++ b/src/cpu/x86/tsc/delay_tsc.c
@@ -83,7 +83,7 @@
 		if (end.lo <= CALIBRATE_DIVISOR)
 			goto bad_ctc;
 
-		return CEIL_DIV(end.lo, CALIBRATE_DIVISOR);
+		return DIV_ROUND_UP(end.lo, CALIBRATE_DIVISOR);
 	}
 
 	/*
diff --git a/src/include/stdlib.h b/src/include/stdlib.h
index 81b805f..34ef93a 100644
--- a/src/include/stdlib.h
+++ b/src/include/stdlib.h
@@ -11,12 +11,4 @@
 /* We never free memory */
 static inline void free(void *ptr) {}
 
-#ifndef __ROMCC__
-static inline unsigned long div_round_up(unsigned int n, unsigned int d)
-{
-	return (n + d - 1) / d;
-}
-#endif
-
-
 #endif /* STDLIB_H */
diff --git a/src/lib/edid.c b/src/lib/edid.c
index 957897e..8d18393 100644
--- a/src/lib/edid.c
+++ b/src/lib/edid.c
@@ -1689,7 +1689,7 @@
 
 	edid->framebuffer_bits_per_pixel = fb_bpp;
 	edid->bytes_per_line = ALIGN_UP(edid->mode.ha *
-		div_round_up(fb_bpp, 8), row_byte_alignment);
+		DIV_ROUND_UP(fb_bpp, 8), row_byte_alignment);
 	edid->x_resolution = edid->bytes_per_line / (fb_bpp / 8);
 	edid->y_resolution = edid->mode.va;
 }
diff --git a/src/mainboard/google/nyan/romstage.c b/src/mainboard/google/nyan/romstage.c
index f3304ae..b94d2dc 100644
--- a/src/mainboard/google/nyan/romstage.c
+++ b/src/mainboard/google/nyan/romstage.c
@@ -55,7 +55,7 @@
 	/* Device memory below DRAM is uncached. */
 	mmu_config_range(0, dram_start_mb, DCACHE_OFF);
 	/* SRAM is cached. MMU code will round size up to page size. */
-	mmu_config_range((uintptr_t)_sram/MiB, div_round_up(_sram_size, MiB),
+	mmu_config_range((uintptr_t)_sram/MiB, DIV_ROUND_UP(_sram_size, MiB),
 			 DCACHE_WRITEBACK);
 	/* DRAM is cached. */
 	mmu_config_range(dram_start_mb, dram_size_mb, DCACHE_WRITEBACK);
diff --git a/src/mainboard/google/nyan_big/romstage.c b/src/mainboard/google/nyan_big/romstage.c
index f3304ae..b94d2dc 100644
--- a/src/mainboard/google/nyan_big/romstage.c
+++ b/src/mainboard/google/nyan_big/romstage.c
@@ -55,7 +55,7 @@
 	/* Device memory below DRAM is uncached. */
 	mmu_config_range(0, dram_start_mb, DCACHE_OFF);
 	/* SRAM is cached. MMU code will round size up to page size. */
-	mmu_config_range((uintptr_t)_sram/MiB, div_round_up(_sram_size, MiB),
+	mmu_config_range((uintptr_t)_sram/MiB, DIV_ROUND_UP(_sram_size, MiB),
 			 DCACHE_WRITEBACK);
 	/* DRAM is cached. */
 	mmu_config_range(dram_start_mb, dram_size_mb, DCACHE_WRITEBACK);
diff --git a/src/mainboard/google/nyan_blaze/romstage.c b/src/mainboard/google/nyan_blaze/romstage.c
index 602989d..e44381f 100644
--- a/src/mainboard/google/nyan_blaze/romstage.c
+++ b/src/mainboard/google/nyan_blaze/romstage.c
@@ -57,7 +57,7 @@
 	/* Device memory below DRAM is uncached. */
 	mmu_config_range(0, dram_start_mb, DCACHE_OFF);
 	/* SRAM is cached. MMU code will round size up to page size. */
-	mmu_config_range((uintptr_t)_sram/MiB, div_round_up(_sram_size, MiB),
+	mmu_config_range((uintptr_t)_sram/MiB, DIV_ROUND_UP(_sram_size, MiB),
 			 DCACHE_WRITEBACK);
 	/* The space above DRAM is uncached. */
 	if (dram_end_mb < 4096)
diff --git a/src/northbridge/intel/gm45/raminit.c b/src/northbridge/intel/gm45/raminit.c
index d067dc2..2d7965f 100644
--- a/src/northbridge/intel/gm45/raminit.c
+++ b/src/northbridge/intel/gm45/raminit.c
@@ -361,8 +361,6 @@
 	}
 }
 
-#define ROUNDUP_DIV(val, by) CEIL_DIV(val, by)
-#define ROUNDUP_DIV_THIS(val, by) val = ROUNDUP_DIV(val, by)
 static fsb_clock_t read_fsb_clock(void)
 {
 	switch (MCHBAR32(CLKCFG_MCHBAR) & CLKCFG_FSBCLK_MASK) {
@@ -452,7 +450,7 @@
 		if (!clock)
 			die("Couldn't find compatible clock / CAS settings.\n");
 		tCKproposed = 8000 / clock;
-		CAS = ROUNDUP_DIV(tAAmin, tCKproposed);
+		CAS = DIV_ROUND_UP(tAAmin, tCKproposed);
 		printk(BIOS_SPEW, "Trying CAS %u, tCK %u.\n", CAS, tCKproposed);
 		for (; CAS <= DDR3_MAX_CAS; ++CAS)
 			if (cas_latencies & (1 << CAS))
@@ -488,10 +486,10 @@
 		if (spdinfo->channel[i].tWR > tWRmin)
 			tWRmin = spdinfo->channel[i].tWR;
 	}
-	ROUNDUP_DIV_THIS(tRASmin, tCLK);
-	ROUNDUP_DIV_THIS(tRPmin, tCLK);
-	ROUNDUP_DIV_THIS(tRCDmin, tCLK);
-	ROUNDUP_DIV_THIS(tWRmin, tCLK);
+	tRASmin = DIV_ROUND_UP(tRASmin, tCLK);
+	tRPmin = DIV_ROUND_UP(tRPmin, tCLK);
+	tRCDmin = DIV_ROUND_UP(tRCDmin, tCLK);
+	tWRmin = DIV_ROUND_UP(tWRmin, tCLK);
 
 	/* Lookup tRFC and calculate common tRFCmin. */
 	const unsigned int tRFC_from_clock_and_cap[][4] = {
diff --git a/src/northbridge/intel/nehalem/raminit.c b/src/northbridge/intel/nehalem/raminit.c
index d0dfb75..54ef278 100644
--- a/src/northbridge/intel/nehalem/raminit.c
+++ b/src/northbridge/intel/nehalem/raminit.c
@@ -602,7 +602,7 @@
 			break;
 		}
 	}
-	min_cas_latency = CEIL_DIV(cas_latency_time, cycletime);
+	min_cas_latency = DIV_ROUND_UP(cas_latency_time, cycletime);
 	cas_latency = 0;
 	while (supported_cas_latencies) {
 		cas_latency = find_highest_bit_set(supported_cas_latencies) + 3;
@@ -3231,7 +3231,7 @@
 
 static inline int div_roundup(int a, int b)
 {
-	return CEIL_DIV(a, b);
+	return DIV_ROUND_UP(a, b);
 }
 
 static unsigned lcm(unsigned a, unsigned b)
diff --git a/src/northbridge/via/vx900/raminit_ddr3.c b/src/northbridge/via/vx900/raminit_ddr3.c
index 1d05fa7..7acab31 100644
--- a/src/northbridge/via/vx900/raminit_ddr3.c
+++ b/src/northbridge/via/vx900/raminit_ddr3.c
@@ -572,7 +572,7 @@
 	printram("Selected DRAM frequency: %u MHz\n", val32);
 
 	/* Find CAS and CWL latencies */
-	val = CEIL_DIV(ctrl->tAA, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tAA, ctrl->tCK);
 	printram("Minimum  CAS latency   : %uT\n", val);
 	/* Find lowest supported CAS latency that satisfies the minimum value */
 	while (!((ctrl->cas_supported >> (val - 4)) & 1)
@@ -591,30 +591,30 @@
 	pci_write_config8(MCU, 0xc0, reg8);
 
 	/* Find tRCD */
-	val = CEIL_DIV(ctrl->tRCD, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tRCD, ctrl->tCK);
 	printram("Selected tRCD          : %uT\n", val);
 	reg8 = ((val - 4) & 0x7) << 4;
 	/* Find tRP */
-	val = CEIL_DIV(ctrl->tRP, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tRP, ctrl->tCK);
 	printram("Selected tRP           : %uT\n", val);
 	reg8 |= ((val - 4) & 0x7);
 	pci_write_config8(MCU, 0xc1, reg8);
 
 	/* Find tRAS */
-	val = CEIL_DIV(ctrl->tRAS, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tRAS, ctrl->tCK);
 	printram("Selected tRAS          : %uT\n", val);
 	reg8 = ((val - 15) & 0x7) << 4;
 	/* Find tWR */
-	ctrl->WR = CEIL_DIV(ctrl->tWR, ctrl->tCK);
+	ctrl->WR = DIV_ROUND_UP(ctrl->tWR, ctrl->tCK);
 	printram("Selected tWR           : %uT\n", ctrl->WR);
 	reg8 |= ((ctrl->WR - 4) & 0x7);
 	pci_write_config8(MCU, 0xc2, reg8);
 
 	/* Find tFAW */
-	tFAW = CEIL_DIV(ctrl->tFAW, ctrl->tCK);
+	tFAW = DIV_ROUND_UP(ctrl->tFAW, ctrl->tCK);
 	printram("Selected tFAW          : %uT\n", tFAW);
 	/* Find tRRD */
-	tRRD = CEIL_DIV(ctrl->tRRD, ctrl->tCK);
+	tRRD = DIV_ROUND_UP(ctrl->tRRD, ctrl->tCK);
 	printram("Selected tRRD          : %uT\n", tRRD);
 	val = tFAW - 4 * tRRD;	/* number of cycles above 4*tRRD */
 	reg8 = ((val - 0) & 0x7) << 4;
@@ -622,11 +622,11 @@
 	pci_write_config8(MCU, 0xc3, reg8);
 
 	/* Find tRTP */
-	val = CEIL_DIV(ctrl->tRTP, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tRTP, ctrl->tCK);
 	printram("Selected tRTP          : %uT\n", val);
 	reg8 = ((val & 0x3) << 4);
 	/* Find tWTR */
-	val = CEIL_DIV(ctrl->tWTR, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tWTR, ctrl->tCK);
 	printram("Selected tWTR          : %uT\n", val);
 	reg8 |= ((val - 2) & 0x7);
 	pci_mod_config8(MCU, 0xc4, 0x3f, reg8);
@@ -639,7 +639,7 @@
 	 *     Since we previously set RxC4[7]
 	 */
 	reg8 = pci_read_config8(MCU, 0xc5);
-	val = CEIL_DIV(ctrl->tRFC, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tRFC, ctrl->tCK);
 	printram("Minimum  tRFC          : %uT\n", val);
 	if (val < 30) {
 		val = 0;
@@ -652,7 +652,7 @@
 	pci_write_config8(MCU, 0xc5, reg8);
 
 	/* Where does this go??? */
-	val = CEIL_DIV(ctrl->tRC, ctrl->tCK);
+	val = DIV_ROUND_UP(ctrl->tRC, ctrl->tCK);
 	printram("Required tRC           : %uT\n", val);
 }
 
diff --git a/src/soc/mediatek/common/spi.c b/src/soc/mediatek/common/spi.c
index 5d2dab4..4e669f9 100644
--- a/src/soc/mediatek/common/spi.c
+++ b/src/soc/mediatek/common/spi.c
@@ -62,11 +62,11 @@
 	struct mtk_spi_regs *regs = slave->regs;
 
 	if (speed_hz < SPI_HZ / 2)
-		div = div_round_up(SPI_HZ, speed_hz);
+		div = DIV_ROUND_UP(SPI_HZ, speed_hz);
 	else
 		div = 1;
 
-	sck_ticks = div_round_up(div, 2);
+	sck_ticks = DIV_ROUND_UP(div, 2);
 	cs_ticks = sck_ticks * 2;
 
 	printk(BIOS_DEBUG, "SPI%u(PAD%u) initialized at %u Hz\n",
@@ -153,7 +153,7 @@
 		 * therefore we need arbitrary data on MOSI which the slave
 		 * must ignore.
 		 */
-		uint32_t word_count = div_round_up(size, sizeof(u32));
+		uint32_t word_count = DIV_ROUND_UP(size, sizeof(u32));
 		for (i = 0; i < word_count; i++)
 			write32(&regs->spi_tx_data_reg, MTK_ARBITRARY_VALUE);
 	}
diff --git a/src/soc/mediatek/mt8173/dsi.c b/src/soc/mediatek/mt8173/dsi.c
index b445d6c..c68cbbb 100644
--- a/src/soc/mediatek/mt8173/dsi.c
+++ b/src/soc/mediatek/mt8173/dsi.c
@@ -216,11 +216,11 @@
 	timcon0 = (8 << 24) | (0xa << 16) | (0x6 << 8) | lpx;
 	timcon1 = (7 << 24) | (5 * lpx << 16) | ((3 * lpx) / 2) << 8 |
 		  (4 * lpx);
-	timcon2 = ((div_round_up(0x64, cycle_time) + 0xa) << 24) |
-		  (div_round_up(0x150, cycle_time) << 16);
+	timcon2 = ((DIV_ROUND_UP(0x64, cycle_time) + 0xa) << 24) |
+		  (DIV_ROUND_UP(0x150, cycle_time) << 16);
 	timcon3 = (2 * lpx) << 16 |
-		  div_round_up(80 + 52 * ui, cycle_time) << 8 |
-		  div_round_up(0x40, cycle_time);
+		  DIV_ROUND_UP(80 + 52 * ui, cycle_time) << 8 |
+		  DIV_ROUND_UP(0x40, cycle_time);
 
 	dsi_write32(&dsi0->dsi_phy_timecon0, timcon0);
 	dsi_write32(&dsi0->dsi_phy_timecon1, timcon1);
diff --git a/src/soc/mediatek/mt8173/i2c.c b/src/soc/mediatek/mt8173/i2c.c
index 9ca6803..b4c3aa9 100644
--- a/src/soc/mediatek/mt8173/i2c.c
+++ b/src/soc/mediatek/mt8173/i2c.c
@@ -98,7 +98,7 @@
 
 	/* Calculate i2c frequency */
 	sample_div = 1;
-	step_div = div_round_up(I2C_CLK_HZ, (400 * KHz * sample_div * 2));
+	step_div = DIV_ROUND_UP(I2C_CLK_HZ, (400 * KHz * sample_div * 2));
 	i2c_freq = I2C_CLK_HZ / (step_div * sample_div * 2);
 	assert(sample_div < 8 && step_div < 64 && i2c_freq < 400 * KHz &&
 	       i2c_freq >= 380 * KHz);
diff --git a/src/soc/nvidia/tegra124/clock.c b/src/soc/nvidia/tegra124/clock.c
index 9173e62..b9a4cd1 100644
--- a/src/soc/nvidia/tegra124/clock.c
+++ b/src/soc/nvidia/tegra124/clock.c
@@ -247,17 +247,17 @@
 		 1 <<  8);			/* (rst) phy_divm */
 
 	write32(&clk_rst->utmip_pll_cfg1,
-		CEIL_DIV(khz, 8000) << 27 |	/* pllu_enbl_cnt / 8 (1us) */
+		DIV_ROUND_UP(khz, 8000) << 27 |	/* pllu_enbl_cnt / 8 (1us) */
 				  0 << 16 |	/* PLLU pwrdn */
 				  0 << 14 |	/* pll_enable pwrdn */
 				  0 << 12 |	/* pll_active pwrdn */
-		 CEIL_DIV(khz, 102) << 0);	/* phy_stbl_cnt / 256 (2.5ms) */
+		 DIV_ROUND_UP(khz, 102) << 0);	/* phy_stbl_cnt / 256 (2.5ms) */
 
 	/* TODO: TRM can't decide if actv is 5us or 10us, keep an eye on it */
 	write32(&clk_rst->utmip_pll_cfg2,
 				  0 << 24 |	/* SAMP_D/XDEV pwrdn */
-		CEIL_DIV(khz, 3200) << 18 |	/* phy_actv_cnt / 16 (5us) */
-		 CEIL_DIV(khz, 256) <<  6 |	/* pllu_stbl_cnt / 256 (1ms) */
+		DIV_ROUND_UP(khz, 3200) << 18 |	/* phy_actv_cnt / 16 (5us) */
+		 DIV_ROUND_UP(khz, 256) <<  6 |	/* pllu_stbl_cnt / 256 (1ms) */
 				  0 <<  4 |	/* SAMP_C/USB3 pwrdn */
 				  0 <<  2 |	/* SAMP_B/XHOST pwrdn */
 				  0 <<  0);	/* SAMP_A/USBD pwrdn */
diff --git a/src/soc/nvidia/tegra124/include/soc/clock.h b/src/soc/nvidia/tegra124/include/soc/clock.h
index d08e26f..bd32f05 100644
--- a/src/soc/nvidia/tegra124/include/soc/clock.h
+++ b/src/soc/nvidia/tegra124/include/soc/clock.h
@@ -200,7 +200,7 @@
  * and voila, upper 7 bits are (ref/freq-1), and lowest bit is h. Since you
  * will assign this to a u8, it gets nicely truncated for you.
  */
-#define CLK_DIVIDER(REF, FREQ)	(div_round_up(((REF) * 2), (FREQ)) - 2)
+#define CLK_DIVIDER(REF, FREQ)	(DIV_ROUND_UP(((REF) * 2), (FREQ)) - 2)
 
 /* Calculate clock frequency value from reference and clock divider value
  * The discussion in the book is pretty lacking.
@@ -253,7 +253,7 @@
  */
 #define clock_configure_i2c_scl_freq(device, src, freq) \
 	_clock_set_div(&clk_rst->clk_src_##device, #device, \
-		div_round_up(TEGRA_##src##_KHZ, (freq) * (0x19 + 1) * 8) - 1, \
+		DIV_ROUND_UP(TEGRA_##src##_KHZ, (freq) * (0x19 + 1) * 8) - 1, \
 		0xffff, src)
 
 enum clock_source {  /* Careful: Not true for all sources, always check TRM! */
diff --git a/src/soc/nvidia/tegra124/verstage.c b/src/soc/nvidia/tegra124/verstage.c
index f8617a8..d99f1a7 100644
--- a/src/soc/nvidia/tegra124/verstage.c
+++ b/src/soc/nvidia/tegra124/verstage.c
@@ -30,7 +30,7 @@
 	/* Whole space is uncached. */
 	mmu_config_range(0, 4096, DCACHE_OFF);
 	/* SRAM is cached. MMU code will round size up to page size. */
-	mmu_config_range((uintptr_t)_sram/MiB, div_round_up(_sram_size, MiB),
+	mmu_config_range((uintptr_t)_sram/MiB, DIV_ROUND_UP(_sram_size, MiB),
 			 DCACHE_WRITEBACK);
 	mmu_disable_range(0, 1);
 	dcache_mmu_enable();
diff --git a/src/soc/nvidia/tegra210/clock.c b/src/soc/nvidia/tegra210/clock.c
index 6ce2ba1..5484d62 100644
--- a/src/soc/nvidia/tegra210/clock.c
+++ b/src/soc/nvidia/tegra210/clock.c
@@ -330,7 +330,7 @@
 
 	/* CFG1 */
 	u32 pllu_enb_ct = 0;
-	u32 phy_stb_ct = div_round_up(khz, 300);  /* phy_stb_ct = 128 */
+	u32 phy_stb_ct = DIV_ROUND_UP(khz, 300);  /* phy_stb_ct = 128 */
 	write32(CLK_RST_REG(utmip_pll_cfg1),
 		pllu_enb_ct << UTMIP_CFG1_PLLU_ENABLE_DLY_COUNT_SHIFT |
 		UTMIP_CFG1_FORCE_PLLU_POWERDOWN_ENABLE |
@@ -341,7 +341,7 @@
 
 	/* CFG2 */
 	u32 pllu_stb_ct = 0;
-	u32 phy_act_ct = div_round_up(khz, 6400); /* phy_act_ct = 6 */
+	u32 phy_act_ct = DIV_ROUND_UP(khz, 6400); /* phy_act_ct = 6 */
 	write32(CLK_RST_REG(utmip_pll_cfg2),
 		phy_act_ct << UTMIP_CFG2_PLL_ACTIVE_DLY_COUNT_SHIFT |
 		pllu_stb_ct << UTMIP_CFG2_PLLU_STABLE_COUNT_SHIFT |
diff --git a/src/soc/nvidia/tegra210/include/soc/clock.h b/src/soc/nvidia/tegra210/include/soc/clock.h
index 50d7260..87d0850 100644
--- a/src/soc/nvidia/tegra210/include/soc/clock.h
+++ b/src/soc/nvidia/tegra210/include/soc/clock.h
@@ -289,7 +289,7 @@
  * and voila, upper 7 bits are (ref/freq-1), and lowest bit is h. Since you
  * will assign this to a u8, it gets nicely truncated for you.
  */
-#define CLK_DIVIDER(REF, FREQ)	(div_round_up(((REF) * 2), (FREQ)) - 2)
+#define CLK_DIVIDER(REF, FREQ)	(DIV_ROUND_UP(((REF) * 2), (FREQ)) - 2)
 
 /* Calculate clock frequency value from reference and clock divider value
  * The discussion in the book is pretty lacking.
@@ -324,7 +324,8 @@
 			src << CLK_SOURCE_SHIFT | div);
 }
 
-#define get_i2c_clk_div(src,freq) (div_round_up(src, (freq) * (0x19 + 1) * 8) - 1)
+#define get_i2c_clk_div(src, freq)	\
+				(DIV_ROUND_UP(src, (freq) * (0x19 + 1) * 8) - 1)
 #define get_clk_div(src,freq)     CLK_DIVIDER(src,freq)
 #define CLK_DIV_MASK              0xff
 #define CLK_DIV_MASK_I2C          0xffff
diff --git a/src/soc/rockchip/common/i2c.c b/src/soc/rockchip/common/i2c.c
index f99b3b5..8629a70 100644
--- a/src/soc/rockchip/common/i2c.c
+++ b/src/soc/rockchip/common/i2c.c
@@ -280,7 +280,7 @@
 
 	/* SCL Divisor = 8*(CLKDIVL + 1 + CLKDIVH + 1)
 	   SCL = PCLK / SCLK Divisor */
-	clk_div = div_round_up(i2c_src_clk, hz * 8);
+	clk_div = DIV_ROUND_UP(i2c_src_clk, hz * 8);
 	divh = clk_div * 3 / 7 - 1;
 	divl = clk_div - divh - 2;
 	i2c_clk = i2c_src_clk / (8 * (divl + 1 + divh + 1));
diff --git a/src/soc/rockchip/common/rk808.c b/src/soc/rockchip/common/rk808.c
index 9b4708c..58d910c 100644
--- a/src/soc/rockchip/common/rk808.c
+++ b/src/soc/rockchip/common/rk808.c
@@ -92,13 +92,13 @@
 	case 4:
 	case 5:
 	case 8:
-		vsel = div_round_up(millivolts, 100) - 18;
+		vsel = DIV_ROUND_UP(millivolts, 100) - 18;
 		assert(vsel <= 0x10);
 		break;
 	case 3:
 	case 6:
 	case 7:
-		vsel = div_round_up(millivolts, 100) - 8;
+		vsel = DIV_ROUND_UP(millivolts, 100) - 8;
 		assert(vsel <= 0x11);
 		break;
 	default:
@@ -118,12 +118,12 @@
 	case 1:
 	case 2:
 		/* 25mV steps. base = 29 * 25mV = 725 */
-		vsel = (div_round_up(millivolts, 25) - 29) * 2 + 1;
+		vsel = (DIV_ROUND_UP(millivolts, 25) - 29) * 2 + 1;
 		assert(vsel <= 0x3f);
 		buck_reg = BUCK1SEL + 4 * (buck - 1);
 		break;
 	case 4:
-		vsel = div_round_up(millivolts, 100) - 18;
+		vsel = DIV_ROUND_UP(millivolts, 100) - 18;
 		assert(vsel <= 0xf);
 		buck_reg = BUCK4SEL;
 		break;
diff --git a/src/soc/rockchip/rk3288/clock.c b/src/soc/rockchip/rk3288/clock.c
index 1b1c135..28c7a426 100644
--- a/src/soc/rockchip/rk3288/clock.c
+++ b/src/soc/rockchip/rk3288/clock.c
@@ -519,15 +519,15 @@
 		return -1;
 	}
 
-	no = div_round_up(VCO_MIN_KHZ, freq_khz);
+	no = DIV_ROUND_UP(VCO_MIN_KHZ, freq_khz);
 	if (ext_div) {
-		*ext_div = div_round_up(no, max_no);
-		no = div_round_up(no, *ext_div);
+		*ext_div = DIV_ROUND_UP(no, max_no);
+		no = DIV_ROUND_UP(no, *ext_div);
 	}
 
 	/* only even divisors (and 1) are supported */
 	if (no > 1)
-		no = div_round_up(no, 2) * 2;
+		no = DIV_ROUND_UP(no, 2) * 2;
 
 	vco_khz = freq_khz * no;
 	if (ext_div)
diff --git a/src/soc/rockchip/rk3288/sdram.c b/src/soc/rockchip/rk3288/sdram.c
index ad5661e..4149a47 100644
--- a/src/soc/rockchip/rk3288/sdram.c
+++ b/src/soc/rockchip/rk3288/sdram.c
@@ -644,7 +644,7 @@
 static void phy_cfg(u32 channel, const struct rk3288_sdram_params *sdram_params)
 {
 	u32 i;
-	u32 dinit2 = div_round_up(sdram_params->ddr_freq/MHz * 200000, 1000);
+	u32 dinit2 = DIV_ROUND_UP(sdram_params->ddr_freq/MHz * 200000, 1000);
 	struct rk3288_ddr_publ_regs *ddr_publ_regs = rk3288_ddr_publ[channel];
 	struct rk3288_msch_regs *msch_regs = rk3288_msch[channel];
 
@@ -658,14 +658,14 @@
 	write32(&msch_regs->devtodev,
 		BUSWRTORD(2) | BUSRDTOWR(2) | BUSRDTORD(1));
 	write32(&ddr_publ_regs->ptr[0],
-	   PRT_DLLLOCK(div_round_up(sdram_params->ddr_freq / MHz * 5120, 1000))
-	 | PRT_DLLSRST(div_round_up(sdram_params->ddr_freq / MHz * 50, 1000))
+	   PRT_DLLLOCK(DIV_ROUND_UP(sdram_params->ddr_freq / MHz * 5120, 1000))
+	 | PRT_DLLSRST(DIV_ROUND_UP(sdram_params->ddr_freq / MHz * 50, 1000))
 	 | PRT_ITMSRST(8));
 	write32(&ddr_publ_regs->ptr[1],
-	   PRT_DINIT0(div_round_up(sdram_params->ddr_freq / MHz * 500000, 1000))
-	 | PRT_DINIT1(div_round_up(sdram_params->ddr_freq / MHz * 400, 1000)));
+	   PRT_DINIT0(DIV_ROUND_UP(sdram_params->ddr_freq / MHz * 500000, 1000))
+	 | PRT_DINIT1(DIV_ROUND_UP(sdram_params->ddr_freq / MHz * 400, 1000)));
 	write32(&ddr_publ_regs->ptr[2], PRT_DINIT2(MIN(dinit2, 0x1ffff))
-	 | PRT_DINIT3(div_round_up(sdram_params->ddr_freq / MHz * 1000, 1000)));
+	 | PRT_DINIT3(DIV_ROUND_UP(sdram_params->ddr_freq / MHz * 1000, 1000)));
 
 	switch (sdram_params->dramtype) {
 	case LPDDR3:
diff --git a/src/soc/rockchip/rk3399/clock.c b/src/soc/rockchip/rk3399/clock.c
index 0b8c83f..cce1d69e 100644
--- a/src/soc/rockchip/rk3399/clock.c
+++ b/src/soc/rockchip/rk3399/clock.c
@@ -429,10 +429,10 @@
 		return -1;
 	}
 
-	postdiv1 = div_round_up(VCO_MIN_KHZ, freq_khz);
+	postdiv1 = DIV_ROUND_UP(VCO_MIN_KHZ, freq_khz);
 	if (postdiv1 > max_postdiv1) {
-		postdiv2 = div_round_up(postdiv1, max_postdiv1);
-		postdiv1 = div_round_up(postdiv1, postdiv2);
+		postdiv2 = DIV_ROUND_UP(postdiv1, max_postdiv1);
+		postdiv1 = DIV_ROUND_UP(postdiv1, postdiv2);
 	}
 
 	vco_khz = freq_khz * postdiv1 * postdiv2;
@@ -605,9 +605,9 @@
 	apll_hz = apll_cfgs[freq]->freq;
 	rkclk_set_pll(pll_con, apll_cfgs[freq]);
 
-	aclkm_div = div_round_up(apll_hz, ACLKM_CORE_HZ) - 1;
-	pclk_dbg_div = div_round_up(apll_hz, PCLK_DBG_HZ) - 1;
-	atclk_div = div_round_up(apll_hz, ATCLK_CORE_HZ) - 1;
+	aclkm_div = DIV_ROUND_UP(apll_hz, ACLKM_CORE_HZ) - 1;
+	pclk_dbg_div = DIV_ROUND_UP(apll_hz, PCLK_DBG_HZ) - 1;
+	atclk_div = DIV_ROUND_UP(apll_hz, ATCLK_CORE_HZ) - 1;
 
 	write32(&cru_ptr->clksel_con[con_base],
 		RK_CLRSETBITS(ACLKM_CORE_DIV_CON_MASK <<
diff --git a/src/soc/rockchip/rk3399/mipi.c b/src/soc/rockchip/rk3399/mipi.c
index 813746e..114b202 100644
--- a/src/soc/rockchip/rk3399/mipi.c
+++ b/src/soc/rockchip/rk3399/mipi.c
@@ -45,7 +45,7 @@
 	int two_frames;
 	unsigned int refresh = edid->mode.refresh;
 
-	two_frames = div_round_up(MSECS_PER_SEC * 2, refresh);
+	two_frames = DIV_ROUND_UP(MSECS_PER_SEC * 2, refresh);
 	mdelay(two_frames);
 }
 
@@ -158,7 +158,7 @@
 static int rk_mipi_dsi_phy_init(struct rk_mipi_dsi *dsi)
 {
 	int i, vco, val;
-	int lane_mbps = div_round_up(dsi->lane_bps, USECS_PER_SEC);
+	int lane_mbps = DIV_ROUND_UP(dsi->lane_bps, USECS_PER_SEC);
 	struct stopwatch sw;
 
 	vco = (lane_mbps < 200) ? 0 : (lane_mbps + 100) / 200;
@@ -318,7 +318,7 @@
 	fref = OSC_HZ;
 
 	/* constraint: 5Mhz <= Fref / N <= 40MHz */
-	min_prediv = div_round_up(fref, 40 * MHz);
+	min_prediv = DIV_ROUND_UP(fref, 40 * MHz);
 	max_prediv = fref / (5 * MHz);
 
 	/* constraint: 80MHz <= Fvco <= 1500Mhz */
@@ -441,7 +441,7 @@
 	u64 lbcc_tmp;
 
 	lbcc_tmp = hcomponent * dsi->lane_bps / (8 * MSECS_PER_SEC);
-	lbcc = div_round_up(lbcc_tmp, edid->mode.pixel_clock);
+	lbcc = DIV_ROUND_UP(lbcc_tmp, edid->mode.pixel_clock);
 
 	return lbcc;
 }
@@ -532,7 +532,7 @@
 	 * which is:
 	 *     (lane_mbps >> 3) / 20 > esc_clk_division
 	 */
-	u32 esc_clk_division = div_round_up(dsi->lane_bps,
+	u32 esc_clk_division = DIV_ROUND_UP(dsi->lane_bps,
 					    8 * 20 * USECS_PER_SEC);
 
 	write32(&dsi->mipi_regs->dsi_pwr_up, RESET);
diff --git a/src/soc/samsung/exynos5250/cpu.c b/src/soc/samsung/exynos5250/cpu.c
index aed1114..4fdb8f8 100644
--- a/src/soc/samsung/exynos5250/cpu.c
+++ b/src/soc/samsung/exynos5250/cpu.c
@@ -112,7 +112,7 @@
 	u32 lcdbase = get_fb_base_kb() * KiB;
 
 	ram_resource(dev, 0, RAM_BASE_KB, RAM_SIZE_KB - FB_SIZE_KB);
-	mmio_resource(dev, 1, lcdbase / KiB, CEIL_DIV(fb_size, KiB));
+	mmio_resource(dev, 1, lcdbase / KiB, DIV_ROUND_UP(fb_size, KiB));
 
 	exynos_displayport_init(dev, lcdbase, fb_size);
 
diff --git a/src/soc/samsung/exynos5420/clock.c b/src/soc/samsung/exynos5420/clock.c
index 04125d9..0da3522 100644
--- a/src/soc/samsung/exynos5420/clock.c
+++ b/src/soc/samsung/exynos5420/clock.c
@@ -345,7 +345,7 @@
 	}
 	printk(BIOS_DEBUG, "%s(%d): sdclkin: %ld\n", __func__, device_index, sdclkin);
 
-	cclkin = CEIL_DIV(sdclkin, freq);
+	cclkin = DIV_ROUND_UP(sdclkin, freq);
 	set_mmc_clk(device_index, cclkin);
 	return 0;
 }
diff --git a/src/soc/samsung/exynos5420/cpu.c b/src/soc/samsung/exynos5420/cpu.c
index ecda54b..fa4cd06 100644
--- a/src/soc/samsung/exynos5420/cpu.c
+++ b/src/soc/samsung/exynos5420/cpu.c
@@ -114,7 +114,7 @@
 	dcache_clean_invalidate_by_mva((void *)lower, upper - lower);
 	mmu_config_range(lower / MiB, (upper - lower) / MiB, DCACHE_OFF);
 
-	mmio_resource(dev, 1, lcdbase/KiB, CEIL_DIV(fb_size, KiB));
+	mmio_resource(dev, 1, lcdbase/KiB, DIV_ROUND_UP(fb_size, KiB));
 }
 
 static void tps65090_thru_ec_fet_disable(int index)
@@ -134,7 +134,7 @@
 	u32 lcdbase = get_fb_base_kb() * KiB;
 
 	ram_resource(dev, 0, RAM_BASE_KB, RAM_SIZE_KB - FB_SIZE_KB);
-	mmio_resource(dev, 1, lcdbase / KiB, CEIL_DIV(fb_size, KiB));
+	mmio_resource(dev, 1, lcdbase / KiB, DIV_ROUND_UP(fb_size, KiB));
 
 	/*
 	 * Disable LCD FETs before we do anything with the display.